From mdehaan at redhat.com Wed Aug 1 21:19:03 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 01 Aug 2007 17:19:03 -0400 Subject: [et-mgmt-tools] Cobbler imports from NAS boxes, etc Message-ID: <46B0F8C7.5000007@redhat.com> I've added some code to Cobbler to make importing things like trees on external filers much easier. Mirroring content to /var/www/cobbler for these configurations is no longer required. This means that for most users, the needs to run "cobbler distro add" vs "cobbler import" will be greatly diminished. This can be invoked (using upstream source) as: cobbler import --mirror=/path/where/filer/is/mounted --name=filer --available-as=nfs://nfsserver.example.org:/is/mounted/here This will be useful for folks who do not have very large /var partitions for mirroring content under the Apache space, or for folks who want to store their distro trees somewhere else than on the Cobbler server. "tree" will be set correctly in the kickstart, relative to the path where everything is discovered. If you have multiple distros available in the location, Cobbler will crawl through the directories and auto-add all of them. This will also be in the 0.6.0 release, and I've alreaded updated the web import examples and the cobbler manpage. --Michael From crobinso at redhat.com Wed Aug 1 21:37:05 2007 From: crobinso at redhat.com (Cole Robinson) Date: Wed, 01 Aug 2007 17:37:05 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components Message-ID: <46B0FD01.4040803@redhat.com> Hello all, Attached is a patch that adds some simple accessibility names for input components of virt-manager. This is to facilitate future gui testing. Thanks, Cole -- Cole Robinson crobinso at redhat.com -------------- next part -------------- A non-text attachment was scrubbed... Name: virt-manager-a11y-name-updates.patch Type: text/x-patch Size: 26733 bytes Desc: not available URL: From hbrock at redhat.com Wed Aug 1 21:40:41 2007 From: hbrock at redhat.com (Hugh Brock) Date: Wed, 01 Aug 2007 17:40:41 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B0FD01.4040803@redhat.com> References: <46B0FD01.4040803@redhat.com> Message-ID: <46B0FDD9.2050705@redhat.com> Cole Robinson wrote: > Hello all, > > Attached is a patch that adds some simple accessibility names for input > components of virt-manager. This is to facilitate future gui testing. > > Thanks, > Cole > +1. I'm going to commit this unless anyone objects... --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From hbrock at redhat.com Wed Aug 1 22:20:39 2007 From: hbrock at redhat.com (Hugh Brock) Date: Wed, 01 Aug 2007 18:20:39 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B0FD01.4040803@redhat.com> References: <46B0FD01.4040803@redhat.com> Message-ID: <46B10737.2070705@redhat.com> Cole Robinson wrote: > Hello all, > > Attached is a patch that adds some simple accessibility names for input > components of virt-manager. This is to facilitate future gui testing. > > Thanks, > Cole > The patch doesn't seem to break anything, but I am getting a warning: libglade-WARNING **: Unexpected element inside . You know anything about this? Thanks, --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From saetaes at gmail.com Thu Aug 2 03:59:13 2007 From: saetaes at gmail.com (Mike M) Date: Wed, 1 Aug 2007 23:59:13 -0400 Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart Message-ID: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> Hello, I'm curious to know how people deal with pre-install steps like configuring hardware RAID arrays, or post-install things like installing server-specific non-Red Hat packages, system setting tweaks, etc, during kickstart. Is kickstart's %pre and %post the best way to do this, or does Cobbler have anything up its sleeve to do this differently? It seems like profiles or even perhaps triggers could be useful, but I'm not quite sure what the best way to accomplish this would be. To illustrate my point, say I have 100 Dell PowerEdge 2950 servers all requiring the same base Red Hat ES 4 installation: 50 need to be configured with hardware RAID 5 and Apache installed in addition to the base OS, 30 need to be configured with hardrware RAID 10 and MySQL, and 20 need to be configured with hardware RAID 1 and Webalizer. Obviously, I don't want to manually configure the RAID arrays prior to provisioning them if I can avoid it. Provided Dell has a tool that runs under linux that can configure the hardware RAID controller *and* I can associate a MAC address to a required RAID config and additional packages, what is the best way to do both the pre-install and the post-install steps? Thanks in advance for any help! Mike From hhoffman at ip-solutions.net Thu Aug 2 12:48:43 2007 From: hhoffman at ip-solutions.net (Harry Hoffman) Date: Thu, 2 Aug 2007 08:48:43 -0400 (EDT) Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart In-Reply-To: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> References: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> Message-ID: <36334.68.80.142.12.1186058923.squirrel@www.ip-solutions.net> Hi, If you find a way to do this outside of the RAID bios I would be very interested. There are utilities to manage the raid once the OS is installed but I believe it's more basic stuff (i.e. check status, list containers, show hot-swap). If these utilities could fully managed the raid (i.e. create/delete containers) then you could potentially build these utilities into the boot images are provision that way ( no idea if it's possible). Cheers, Harry > Hello, > > I'm curious to know how people deal with pre-install steps like > configuring hardware RAID arrays, or post-install things like > installing server-specific non-Red Hat packages, system setting > tweaks, etc, during kickstart. Is kickstart's %pre and %post the best > way to do this, or does Cobbler have anything up its sleeve to do this > differently? It seems like profiles or even perhaps triggers could be > useful, but I'm not quite sure what the best way to accomplish this > would be. > > To illustrate my point, say I have 100 Dell PowerEdge 2950 servers all > requiring the same base Red Hat ES 4 installation: 50 need to be > configured with hardware RAID 5 and Apache installed in addition to > the base OS, 30 need to be configured with hardrware RAID 10 and > MySQL, and 20 need to be configured with hardware RAID 1 and > Webalizer. Obviously, I don't want to manually configure the RAID > arrays prior to provisioning them if I can avoid it. Provided Dell > has a tool that runs under linux that can configure the hardware RAID > controller *and* I can associate a MAC address to a required RAID > config and additional packages, what is the best way to do both the > pre-install and the post-install steps? > > Thanks in advance for any help! > > Mike > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From saetaes at gmail.com Thu Aug 2 13:04:37 2007 From: saetaes at gmail.com (Mike M) Date: Thu, 2 Aug 2007 09:04:37 -0400 Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart In-Reply-To: <36334.68.80.142.12.1186058923.squirrel@www.ip-solutions.net> References: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> <36334.68.80.142.12.1186058923.squirrel@www.ip-solutions.net> Message-ID: <8552c3a30708020604y2f6dfc29g3c64eba637ac894a@mail.gmail.com> On 8/2/07, Harry Hoffman wrote: > If you find a way to do this outside of the RAID bios I would be very > interested. > > There are utilities to manage the raid once the OS is installed but I > believe it's more basic stuff (i.e. check status, list containers, show > hot-swap). Dell actually provides something that *should* work - Dell OpenManage Deployment Toolkit: http://support.dell.com/support/downloads/format.aspx?c=us&l=en&s=gen&deviceid=4450&libid=36&releaseid=R158116&vercnt=4&formatcnt=0&SystemID=PWE_2950&servicetag=&os=LIN4&osl=en&catid=-1&impid=-1 I haven't tried using the RAID tools just yet, but it does seem feasible to shim it into kickstart. I assume other hardware vendors have similar tools. Mike From crobinso at redhat.com Thu Aug 2 14:07:52 2007 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 02 Aug 2007 10:07:52 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B10737.2070705@redhat.com> References: <46B0FD01.4040803@redhat.com> <46B10737.2070705@redhat.com> Message-ID: <46B1E538.8090203@redhat.com> Hugh Brock wrote: > Cole Robinson wrote: >> Hello all, >> >> Attached is a patch that adds some simple accessibility names for input >> components of virt-manager. This is to facilitate future gui testing. >> >> Thanks, >> Cole >> > > The patch doesn't seem to break anything, but I am getting a warning: > libglade-WARNING **: Unexpected element inside . > You know anything about this? > > Thanks, > --Hugh > Hmm, I tried to reproduce it but I'm not seeing it on fully updated F7 at least. I'm guessing this happens when you start up the patched virt-manager? - Cole -- Cole Robinson crobinso at redhat.com From mdehaan at redhat.com Thu Aug 2 14:43:32 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 02 Aug 2007 10:43:32 -0400 Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart In-Reply-To: <8552c3a30708020604y2f6dfc29g3c64eba637ac894a@mail.gmail.com> References: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> <36334.68.80.142.12.1186058923.squirrel@www.ip-solutions.net> <8552c3a30708020604y2f6dfc29g3c64eba637ac894a@mail.gmail.com> Message-ID: <46B1ED94.6060008@redhat.com> Mike M wrote: > On 8/2/07, Harry Hoffman wrote: > >> If you find a way to do this outside of the RAID bios I would be very >> interested. >> >> There are utilities to manage the raid once the OS is installed but I >> believe it's more basic stuff (i.e. check status, list containers, show >> hot-swap). >> > > Dell actually provides something that *should* work - Dell OpenManage > Deployment Toolkit: > > http://support.dell.com/support/downloads/format.aspx?c=us&l=en&s=gen&deviceid=4450&libid=36&releaseid=R158116&vercnt=4&formatcnt=0&SystemID=PWE_2950&servicetag=&os=LIN4&osl=en&catid=-1&impid=-1 > > I haven't tried using the RAID tools just yet, but it does seem > feasible to shim it into kickstart. I assume other hardware vendors > have similar tools. > Ironically, I used to write hardware RAID configuration software in a previous life :) In theory, it should be possible to (a) retrieve the OEM command line tool in %pre (wget?) and configure hardware RAID with it there (in %pre), so you wouldn't have to rebuild the initrd just to add the raid tools. Either way, after running the RAID tool in %pre you will probably have to do something like scsi-add-single to get Anaconda to see the new drives prior to partitioning setup. Linux does not generally discover new logical drives without being told to look. Your driver/RAID tool might vary though. I haven't really tried this to see if it would work, but I don't see why it wouldn't. Obviously you are also going to need your drivers for the particular hardware installed first, so you'll probably want to look into the "driverdisk" option in kickstart if they are not enabled by default. Using the shell (Control-Alt-F2?) is probably a good idea to do this all manually the first time, to see what problems you might encounter. Possibly put a sleep in pre to give you some time to test your commands within the boot environment and then write down what you need to do. http://www.red-hat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-options.html --Michael > Mike > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From hbrock at redhat.com Thu Aug 2 15:01:19 2007 From: hbrock at redhat.com (Hugh Brock) Date: Thu, 02 Aug 2007 11:01:19 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B1E538.8090203@redhat.com> References: <46B0FD01.4040803@redhat.com> <46B10737.2070705@redhat.com> <46B1E538.8090203@redhat.com> Message-ID: <46B1F1BF.9030305@redhat.com> Cole Robinson wrote: > Hugh Brock wrote: >> Cole Robinson wrote: >>> Hello all, >>> >>> Attached is a patch that adds some simple accessibility names for input >>> components of virt-manager. This is to facilitate future gui testing. >>> >>> Thanks, >>> Cole >>> >> The patch doesn't seem to break anything, but I am getting a warning: >> libglade-WARNING **: Unexpected element inside . >> You know anything about this? >> >> Thanks, >> --Hugh >> > > Hmm, I tried to reproduce it but I'm not seeing it on fully updated F7 > at least. I'm guessing this happens when you start up the patched > virt-manager? > Yeah, I get it when I open the create wizard actually. I was testing on, umm, latest F7 I think. Could be a bad merge since your patch to vmm-create.glade merged with offsets. However the message itself is odd; should be legal inside . Note I see the warning when I run virt-manager from the command line a la "python virt-manager.py" -- so I am getting all the gtk spew as well. I'm inclined to ignore the message unless we think it indicates something more serious... --H -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From crobinso at redhat.com Thu Aug 2 15:34:14 2007 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 02 Aug 2007 11:34:14 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B1F1BF.9030305@redhat.com> References: <46B0FD01.4040803@redhat.com> <46B10737.2070705@redhat.com> <46B1E538.8090203@redhat.com> <46B1F1BF.9030305@redhat.com> Message-ID: <46B1F976.3080101@redhat.com> Hugh Brock wrote: > Cole Robinson wrote: >> Hugh Brock wrote: >>> Cole Robinson wrote: >>>> Hello all, >>>> >>>> Attached is a patch that adds some simple accessibility names for input >>>> components of virt-manager. This is to facilitate future gui testing. >>>> >>>> Thanks, >>>> Cole >>>> >>> The patch doesn't seem to break anything, but I am getting a warning: >>> libglade-WARNING **: Unexpected element inside . >>> You know anything about this? >>> >>> Thanks, >>> --Hugh >>> >> Hmm, I tried to reproduce it but I'm not seeing it on fully updated F7 >> at least. I'm guessing this happens when you start up the patched >> virt-manager? >> > > Yeah, I get it when I open the create wizard actually. I was testing on, > umm, latest F7 I think. Could be a bad merge since your patch to > vmm-create.glade merged with offsets. However the message itself is odd; > should be legal inside . > > Note I see the warning when I run virt-manager from the command line a > la "python virt-manager.py" -- so I am getting all the gtk spew as well. > > I'm inclined to ignore the message unless we think it indicates > something more serious... > > --H > When I 'hg import' the patch against upstream tip I don't see any messages about offsets. It seems to go smoothly. Maybe whatever we are doing differently is the culprit. - Cole -- Cole Robinson crobinso at redhat.com From mdehaan at redhat.com Thu Aug 2 15:49:45 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 02 Aug 2007 11:49:45 -0400 Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart In-Reply-To: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> References: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> Message-ID: <46B1FD19.8000500@redhat.com> Mike M wrote: > post-install things like > installing server-specific non-Red Hat packages Depends on what OS you have. RHEL-5 and FC-6 debued the usage of the "repo" command in kickstart, which allows you to attach, in a kickstart, additional yum repositories to install things from. These can be yum repositories you maintain yourself (say you had a directory full of RPMs in /var/www/html/myrepo and ran createrepo from the yum-utils package on that directory to make it a yum repository), or you have yum repositories like Fedora Extras (FC6 and before), Updates, EPEL, Centos Extras, 3rd party repos, etc. If you are using a distro that supports "repo", this is the best way to install non-vendor packages in kickstart, because you can just mention them directly in the %packages section. The cool thing about the "repo" directive is that you can also add an updates repository to the repo list, so that packages can be already up to date right after install, without having to invoke "yum update -y". Cobbler can even help you create a local mirror of these various external yum repositories and automatically associate them with your cobbler profiles: http://cobbler.et.redhat.com/cobbler-repos.php If you are using a version of Anaconda that doesn't support the "repo" command (like RHEL4), one thing you can do is just set up an external install sourcein %post and do things there, just do wgets of the RPMs and install them in %post, or rely on a configuration management system to do things once the system boots. From atodorov at redhat.com Thu Aug 2 16:09:42 2007 From: atodorov at redhat.com (Alexander Todorov) Date: Thu, 02 Aug 2007 18:09:42 +0200 Subject: [et-mgmt-tools] How to switch betwee different ttys in virt-manager? Message-ID: <46B201C6.3080005@redhat.com> Hello, previously I was able to use the combination: press several times quickly Ctrl then Alt+1/+2, etc to switch to tty1, tty2. Now Ctrl+Alt interferes with release of mouse pointer capture. Is there a new key combination or is this a bug? Greetings, Alexander. From crobinso at redhat.com Thu Aug 2 16:32:47 2007 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 02 Aug 2007 12:32:47 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B1F1BF.9030305@redhat.com> References: <46B0FD01.4040803@redhat.com> <46B10737.2070705@redhat.com> <46B1E538.8090203@redhat.com> <46B1F1BF.9030305@redhat.com> Message-ID: <46B2072F.2000208@redhat.com> Hugh Brock wrote: > Cole Robinson wrote: >> Hugh Brock wrote: >>> Cole Robinson wrote: >>>> Hello all, >>>> >>>> Attached is a patch that adds some simple accessibility names for input >>>> components of virt-manager. This is to facilitate future gui testing. >>>> >>>> Thanks, >>>> Cole >>>> >>> The patch doesn't seem to break anything, but I am getting a warning: >>> libglade-WARNING **: Unexpected element inside . >>> You know anything about this? >>> >>> Thanks, >>> --Hugh >>> >> Hmm, I tried to reproduce it but I'm not seeing it on fully updated F7 >> at least. I'm guessing this happens when you start up the patched >> virt-manager? >> > > Yeah, I get it when I open the create wizard actually. I was testing on, > umm, latest F7 I think. Could be a bad merge since your patch to > vmm-create.glade merged with offsets. However the message itself is odd; > should be legal inside . > > Note I see the warning when I run virt-manager from the command line a > la "python virt-manager.py" -- so I am getting all the gtk spew as well. > > I'm inclined to ignore the message unless we think it indicates > something more serious... > > --H > Well it seems there was something funky with the patch. I reworked it and I'm not seeing the error this time (though it may have slipped by since I missed it on the first try!) - Cole -- Cole Robinson crobinso at redhat.com -------------- next part -------------- A non-text attachment was scrubbed... Name: virt-manager-a11y-name-update02.patch Type: text/x-patch Size: 26733 bytes Desc: not available URL: From hbrock at redhat.com Thu Aug 2 16:41:10 2007 From: hbrock at redhat.com (Hugh Brock) Date: Thu, 02 Aug 2007 12:41:10 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B2072F.2000208@redhat.com> References: <46B0FD01.4040803@redhat.com> <46B10737.2070705@redhat.com> <46B1E538.8090203@redhat.com> <46B1F1BF.9030305@redhat.com> <46B2072F.2000208@redhat.com> Message-ID: <46B20926.3030506@redhat.com> >> Note I see the warning when I run virt-manager from the command line a >> la "python virt-manager.py" -- so I am getting all the gtk spew as well. >> >> I'm inclined to ignore the message unless we think it indicates >> something more serious... >> >> --H >> > > Well it seems there was something funky with the patch. I reworked it > and I'm not seeing the error this time (though it may have slipped by > since I missed it on the first try!) > > - Cole > That has eliminated the error for me. Thanks! --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From hbrock at redhat.com Thu Aug 2 18:27:17 2007 From: hbrock at redhat.com (Hugh Brock) Date: Thu, 02 Aug 2007 14:27:17 -0400 Subject: [et-mgmt-tools] [PATCH] Add a11y names for virt-manager components In-Reply-To: <46B2072F.2000208@redhat.com> References: <46B0FD01.4040803@redhat.com> <46B10737.2070705@redhat.com> <46B1E538.8090203@redhat.com> <46B1F1BF.9030305@redhat.com> <46B2072F.2000208@redhat.com> Message-ID: <46B22205.3070004@redhat.com> > > Well it seems there was something funky with the patch. I reworked it > and I'm not seeing the error this time (though it may have slipped by > since I missed it on the first try!) > > - Cole > Committed, thanks! --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From wright at imageworks.com Fri Aug 3 01:15:58 2007 From: wright at imageworks.com (Peter Wright) Date: Thu, 02 Aug 2007 18:15:58 -0700 Subject: [et-mgmt-tools] boot locally after install via cobbler Message-ID: <46B281CE.9080604@imageworks.com> hi all, i'm currently working on moving a cluster over from Xcat to cobbler. so far things have gone quite smoothly, i have imported several distros - created my own distros and gotten my custom kickstart's working quite easilly! my question is i have not figured out how one has an installation dial back to the cobbler master node and tell it that it no longer needs to to re-install itself. with Xcat during the %post phase your node would set it's status on the Xcat master node to boot locally after install - is there something similar for cobbler? hopefully i'm missing something basic here, but have had no luck reading through the man pages or mailing list archives. thanks! -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From hhoffman at ip-solutions.net Fri Aug 3 01:23:23 2007 From: hhoffman at ip-solutions.net (Harry Hoffman) Date: Thu, 2 Aug 2007 21:23:23 -0400 (EDT) Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B281CE.9080604@imageworks.com> References: <46B281CE.9080604@imageworks.com> Message-ID: <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> Hi Peter, I had this same problem... it should be said that with >= cobbler-0.5 there is a option to edit the system: cobbler system add --name=string --profile=string [--mac=macaddress] [--ip=ipaddress] [--hostname=hostname] [--kopts=string] [--ipad- dress=string] [--ksmeta=string] [--netboot-enabled=Y/N The nice thing about cobbler via (git - yeah, it's a messed up name) is that a make in the d/l'd src directory will build you a rpm with proper version so that yum upgrades will overwrite it. It's a pretty trivial process, feel free to ask questions. Cheers, Harry > hi all, > i'm currently working on moving a cluster over from Xcat to cobbler. so > far things have gone quite smoothly, i have imported several distros - > created my own distros and gotten my custom kickstart's working quite > easilly! > > my question is i have not figured out how one has an installation dial > back to the cobbler master node and tell it that it no longer needs to > to re-install itself. with Xcat during the %post phase your node would > set it's status on the Xcat master node to boot locally after install - > is there something similar for cobbler? > > hopefully i'm missing something basic here, but have had no luck reading > through the man pages or mailing list archives. > > thanks! > -pete > > -- > Peter Wright > Systems Administrator > Sony Pictures Imageworks > wright at imageworks.com > www.imageworks.com > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From wright at imageworks.com Fri Aug 3 01:28:11 2007 From: wright at imageworks.com (Peter Wright) Date: Thu, 02 Aug 2007 18:28:11 -0700 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> Message-ID: <46B284AB.8060104@imageworks.com> Harry Hoffman wrote: > > Hi Peter, > > I had this same problem... it should be said that with >= cobbler-0.5 > there is a option to edit the system: > cobbler system add --name=string --profile=string [--mac=macaddress] > [--ip=ipaddress] [--hostname=hostname] [--kopts=string] [--ipad- > dress=string] [--ksmeta=string] [--netboot-enabled=Y/N > > The nice thing about cobbler via (git - yeah, it's a messed up name) is > that a make in the d/l'd src directory will build you a rpm with proper > version so that yum upgrades will overwrite it. > > It's a pretty trivial process, feel free to ask questions. > Awesome, thanks Harry - I'll start diving into this tomorrow then. Just to make sure I understand clearly. With the newer version I should be able to run something like this during %post: cobbler system edit --name=$MAC --netboot-enabled=N -pete > > > > hi all, > > i'm currently working on moving a cluster over from Xcat to cobbler. so > > far things have gone quite smoothly, i have imported several distros - > > created my own distros and gotten my custom kickstart's working quite > > easilly! > > > > my question is i have not figured out how one has an installation dial > > back to the cobbler master node and tell it that it no longer needs to > > to re-install itself. with Xcat during the %post phase your node would > > set it's status on the Xcat master node to boot locally after install - > > is there something similar for cobbler? > > > > hopefully i'm missing something basic here, but have had no luck reading > > through the man pages or mailing list archives. > > > > thanks! > > -pete > > > > -- > > Peter Wright > > Systems Administrator > > Sony Pictures Imageworks > > wright at imageworks.com > > www.imageworks.com > > > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From hhoffman at ip-solutions.net Fri Aug 3 01:31:22 2007 From: hhoffman at ip-solutions.net (Harry Hoffman) Date: Thu, 2 Aug 2007 21:31:22 -0400 (EDT) Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> Message-ID: <41818.68.80.142.12.1186104682.squirrel@www.ip-solutions.net> wow, I typed that way too quickly... if you use git to check out the latest version of cobbler you will get a option do disable netboot via a cmdline cobbler system edit --name= --netboot-enabled=N If you type make in the src directory it will build you a proper rpm with a revision higher then your current version of cobbler so that yum will *not* try to overwrite the currently installed version. > Hi Peter, > > I had this same problem... it should be said that with >= cobbler-0.5 > there is a option to edit the system: > cobbler system add --name=string --profile=string [--mac=macaddress] > [--ip=ipaddress] [--hostname=hostname] [--kopts=string] [--ipad- > dress=string] [--ksmeta=string] [--netboot-enabled=Y/N > > The nice thing about cobbler via (git - yeah, it's a messed up name) is > that a make in the d/l'd src directory will build you a rpm with proper > version so that yum upgrades will overwrite it. > > It's a pretty trivial process, feel free to ask questions. > > Cheers, > Harry > > >> hi all, >> i'm currently working on moving a cluster over from Xcat to cobbler. so >> far things have gone quite smoothly, i have imported several distros - >> created my own distros and gotten my custom kickstart's working quite >> easilly! >> >> my question is i have not figured out how one has an installation dial >> back to the cobbler master node and tell it that it no longer needs to >> to re-install itself. with Xcat during the %post phase your node would >> set it's status on the Xcat master node to boot locally after install - >> is there something similar for cobbler? >> >> hopefully i'm missing something basic here, but have had no luck reading >> through the man pages or mailing list archives. >> >> thanks! >> -pete >> >> -- >> Peter Wright >> Systems Administrator >> Sony Pictures Imageworks >> wright at imageworks.com >> www.imageworks.com >> >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From hhoffman at ip-solutions.net Fri Aug 3 01:34:41 2007 From: hhoffman at ip-solutions.net (Harry Hoffman) Date: Thu, 2 Aug 2007 21:34:41 -0400 (EDT) Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B284AB.8060104@imageworks.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> Message-ID: <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> Hi Peter, No, I don't think that'll work... but if you wrote a quick cgi to accept the args of mac addr and netboot then you could exec the cobbler command... something like %post wget http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n You'd of course want to do the normal sanitization of user data. I've got a script that I could modify to do this, let me know if you need/want it. Cheers, Harry > Harry Hoffman wrote: >> >> Hi Peter, >> >> I had this same problem... it should be said that with >= cobbler-0.5 >> there is a option to edit the system: >> cobbler system add --name=string --profile=string [--mac=macaddress] >> [--ip=ipaddress] [--hostname=hostname] [--kopts=string] [--ipad- >> dress=string] [--ksmeta=string] [--netboot-enabled=Y/N >> >> The nice thing about cobbler via (git - yeah, it's a messed up name) is >> that a make in the d/l'd src directory will build you a rpm with proper >> version so that yum upgrades will overwrite it. >> >> It's a pretty trivial process, feel free to ask questions. >> > > Awesome, thanks Harry - I'll start diving into this tomorrow then. Just > to make sure I understand clearly. With the newer version I should be > able to run something like this during %post: > > cobbler system edit --name=$MAC --netboot-enabled=N > > -pete > >> >> >> > hi all, >> > i'm currently working on moving a cluster over from Xcat to cobbler. >> so >> > far things have gone quite smoothly, i have imported several distros - >> > created my own distros and gotten my custom kickstart's working quite >> > easilly! >> > >> > my question is i have not figured out how one has an installation dial >> > back to the cobbler master node and tell it that it no longer needs to >> > to re-install itself. with Xcat during the %post phase your node >> would >> > set it's status on the Xcat master node to boot locally after install >> - >> > is there something similar for cobbler? >> > >> > hopefully i'm missing something basic here, but have had no luck >> reading >> > through the man pages or mailing list archives. >> > >> > thanks! >> > -pete >> > >> > -- >> > Peter Wright >> > Systems Administrator >> > Sony Pictures Imageworks >> > wright at imageworks.com >> > www.imageworks.com >> > >> > >> > _______________________________________________ >> > et-mgmt-tools mailing list >> > et-mgmt-tools at redhat.com >> > https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> > >> >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> > > > -- > Peter Wright > Systems Administrator > Sony Pictures Imageworks > wright at imageworks.com > www.imageworks.com > > > From wright at imageworks.com Fri Aug 3 01:36:34 2007 From: wright at imageworks.com (Peter Wright) Date: Thu, 02 Aug 2007 18:36:34 -0700 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> Message-ID: <46B286A2.4080001@imageworks.com> Harry Hoffman wrote: > > Hi Peter, > > No, I don't think that'll work... but if you wrote a quick cgi to accept > the args of mac addr and netboot then you could exec the cobbler > command... > > something like > > %post > > wget > http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n > > > You'd of course want to do the normal sanitization of user data. > > I've got a script that I could modify to do this, let me know if you > need/want it. > > Cheers, > Harry > ok - i think i'm getting this thing now. nice - if you don't mind posting that script i'd love to take a look at it. -p > > > Harry Hoffman wrote: > >> > >> Hi Peter, > >> > >> I had this same problem... it should be said that with >= cobbler-0.5 > >> there is a option to edit the system: > >> cobbler system add --name=string --profile=string [--mac=macaddress] > >> [--ip=ipaddress] [--hostname=hostname] [--kopts=string] [--ipad- > >> dress=string] [--ksmeta=string] [--netboot-enabled=Y/N > >> > >> The nice thing about cobbler via (git - yeah, it's a messed up name) is > >> that a make in the d/l'd src directory will build you a rpm with proper > >> version so that yum upgrades will overwrite it. > >> > >> It's a pretty trivial process, feel free to ask questions. > >> > > > > Awesome, thanks Harry - I'll start diving into this tomorrow then. Just > > to make sure I understand clearly. With the newer version I should be > > able to run something like this during %post: > > > > cobbler system edit --name=$MAC --netboot-enabled=N > > > > -pete > > > >> > >> > >> > hi all, > >> > i'm currently working on moving a cluster over from Xcat to cobbler. > >> so > >> > far things have gone quite smoothly, i have imported several > distros - > >> > created my own distros and gotten my custom kickstart's working quite > >> > easilly! > >> > > >> > my question is i have not figured out how one has an installation > dial > >> > back to the cobbler master node and tell it that it no longer > needs to > >> > to re-install itself. with Xcat during the %post phase your node > >> would > >> > set it's status on the Xcat master node to boot locally after install > >> - > >> > is there something similar for cobbler? > >> > > >> > hopefully i'm missing something basic here, but have had no luck > >> reading > >> > through the man pages or mailing list archives. > >> > > >> > thanks! > >> > -pete > >> > > >> > -- > >> > Peter Wright > >> > Systems Administrator > >> > Sony Pictures Imageworks > >> > wright at imageworks.com > >> > www.imageworks.com > >> > > >> > > >> > _______________________________________________ > >> > et-mgmt-tools mailing list > >> > et-mgmt-tools at redhat.com > >> > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > >> > > >> > >> > >> _______________________________________________ > >> et-mgmt-tools mailing list > >> et-mgmt-tools at redhat.com > >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools > >> > > > > > > -- > > Peter Wright > > Systems Administrator > > Sony Pictures Imageworks > > wright at imageworks.com > > www.imageworks.com > > > > > > > > > -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From jim at rossberry.com Fri Aug 3 02:20:38 2007 From: jim at rossberry.com (Jim Wildman) Date: Thu, 2 Aug 2007 22:20:38 -0400 (EDT) Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart In-Reply-To: <46B1ED94.6060008@redhat.com> References: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> <36334.68.80.142.12.1186058923.squirrel@www.ip-solutions.net> <8552c3a30708020604y2f6dfc29g3c64eba637ac894a@mail.gmail.com> <46B1ED94.6060008@redhat.com> Message-ID: On Thu, 2 Aug 2007, Michael DeHaan wrote: >> http://support.dell.com/support/downloads/format.aspx?c=us&l=en&s=gen&deviceid=4450&libid=36&releaseid=R158116&vercnt=4&formatcnt=0&SystemID=PWE_2950&servicetag=&os=LIN4&osl=en&catid=-1&impid=-1 >> >> I haven't tried using the RAID tools just yet, but it does seem >> feasible to shim it into kickstart. I assume other hardware vendors >> have similar tools. >> > > Ironically, I used to write hardware RAID configuration software in a > previous life :) > > In theory, it should be possible to (a) retrieve the OEM command line tool in > %pre (wget?) and configure hardware RAID with it there (in %pre), so you > wouldn't have to rebuild the initrd just to add the raid tools. Either way, > after running the RAID tool in %pre you will probably have to do something > like scsi-add-single to get Anaconda to see the new drives prior to > partitioning setup. Linux does not generally discover new logical drives > without being told to look. Your driver/RAID tool might vary though. > > I haven't really tried this to see if it would work, but I don't see why it > wouldn't. > > Obviously you are also going to need your drivers for the particular hardware > installed first, so you'll probably want to look into the "driverdisk" option > in kickstart if they > are not enabled by default. Using the shell (Control-Alt-F2?) is probably a > good idea to do this all manually the first time, to see what problems you > might encounter. Possibly put a sleep in pre to give you some time to test > your commands within the boot environment and then write down what you need > to do. > > http://www.red-hat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-options.html > I can confirm it is possible to completely configure an HP Proliant (DL) server including the RAID during a kickstart. I've seen it configured and working, but for various reasons (politics, bah) we did not implement it. As I recall, it took 2 boot cycles, one for the box config, then reboot and do the install, so that would get around the drive detection issue. Much like Dell there is a Proliant Deployment pack with all kinds of goodies in it. ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com "Society in every state is a blessing, but Government, even in its best state, is a necessary evil; in its worst state, an intolerable one." Thomas Paine From saetaes at gmail.com Fri Aug 3 03:44:15 2007 From: saetaes at gmail.com (Mike M) Date: Thu, 2 Aug 2007 23:44:15 -0400 Subject: [et-mgmt-tools] Cobbler + %pre and %post in kickstart In-Reply-To: <46B1ED94.6060008@redhat.com> References: <8552c3a30708012059g1a173b4awdb7fce29680b6659@mail.gmail.com> <36334.68.80.142.12.1186058923.squirrel@www.ip-solutions.net> <8552c3a30708020604y2f6dfc29g3c64eba637ac894a@mail.gmail.com> <46B1ED94.6060008@redhat.com> Message-ID: <8552c3a30708022044nd55fdc2gb96ce2d874d0024c@mail.gmail.com> On 8/2/07, Michael DeHaan wrote: > In theory, it should be possible to (a) retrieve the OEM command line > tool in %pre (wget?) and configure hardware RAID with it there (in > %pre), so you wouldn't have to rebuild the initrd just to add the raid > tools. Either way, after running the RAID tool in %pre you will > probably have to do something like scsi-add-single to get Anaconda to > see the new drives prior to partitioning setup. Unfortunately, this is exactly how we do it now, and it's a mess. In almost all cases, it's just easier to manually configure the RAID array than to deal with extracting drivers from cpio archives, making linux re-discover the disk changes, etc. As you mention, there's a lot of debugging that has to happen, and not just the first time. Basically, every time Dell changes it's RAID controller (SCSI to SAS was especially painful), or we decide to kickstart a different Red Hat Version or Update level, it causes problems. I was hoping for some trick up Cobbler's sleeve that would make it easier, but it seems like that's just wishful thinking on my part... Thanks anyway for the responses! Mike From fj0588di at aa.jp.fujitsu.com Fri Aug 3 08:05:13 2007 From: fj0588di at aa.jp.fujitsu.com (S.Sakamoto) Date: Fri, 3 Aug 2007 17:05:13 +0900 Subject: [et-mgmt-tools] [PATCH] check if the specified disk usingvirt-clone is already used in another guests. In-Reply-To: <200707300954.AFB05251.9EKJ60G9@aa.jp.fujitsu.com> References: <200707300954.AFB05251.9EKJ60G9@aa.jp.fujitsu.com> Message-ID: <200708031705.GAI26568.JGE9K960@aa.jp.fujitsu.com> Hi Would you give me a comment on this patch? If not, please apply it. Thanks, Shigeki Sakamoto. > Hi > > Current virt-clone does not confirm if the specified disk is already > used in another guests or not. So, I create the patch to check if the > specified disk using virt-clone is already used in another guests. > > > Signed-off-by: Shigeki Sakamoto > > Thanks, > Shigeki Sakamoto. > > ---------------------------------------------------------------------- > diff -r 645217bda13b virtinst/CloneManager.py > --- a/virtinst/CloneManager.py Sat Jul 21 13:03:07 2007 -0400 > +++ b/virtinst/CloneManager.py Fri Jul 27 17:38:47 2007 +0900 > @@ -29,6 +29,7 @@ import commands > import commands > import libvirt > import Guest > +import cli > from virtinst import _virtinst as _ > > # > @@ -106,6 +107,13 @@ class CloneDesign(object): > def set_clone_devices(self, devices): > if len(devices) == 0: > raise ValueError, _("New file to use for disk image is required") > + cdev = [] > + cdev_size = [] > + cdev_type = [] > + cdev.append(devices) > + cdev_size,\ > + cdev_type = self._get_clone_devices_info(cdev) > + devices = self._check_file(self._hyper_conn, devices, cdev_size) > self._clone_devices.append(devices) > def get_clone_devices(self): > return self._clone_devices > @@ -320,6 +328,45 @@ class CloneDesign(object): > except libvirt.libvirtError, e: > pass > return check > + > + # > + # check used file func > + # ret : Use File Path > + # > + def _check_file(self, conn, disk, size): > + retryFlg = False > + while 1: > + if disk == None: > + msg = _("What would you like to use as the disk (path)?") > + disk = cli.prompt_for_input(msg, disk) > + > + try: > + d = Guest.VirtualDisk(disk, size) > + if d.is_conflict_disk(conn) is True: > + while 1: > + retryFlg = False > + warnmsg = _("Disk %s is already in use by another guest!\n") % disk > + res = cli.prompt_for_input(warnmsg + _("Do you really want to use the disk (yes or no)? ")) > + try: > + if cli.yes_or_no(res) is True: > + break > + else: > + retryFlg = True > + break > + except ValueError, e: > + print _("ERROR: "), e > + continue > + if retryFlg is True: > + disk = None > + continue > + except ValueError, e: > + print _("ERROR: "), e > + disk = None > + continue > + > + break > + > + return disk > > # > # check used mac func > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Fri Aug 3 14:41:11 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 03 Aug 2007 10:41:11 -0400 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B286A2.4080001@imageworks.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> <46B286A2.4080001@imageworks.com> Message-ID: <46B33E87.50701@redhat.com> Peter Wright wrote: > Harry Hoffman wrote: >> >> Hi Peter, >> >> No, I don't think that'll work... but if you wrote a quick cgi to accept >> the args of mac addr and netboot then you could exec the cobbler >> command... >> >> something like >> >> %post >> >> wget >> http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n >> >> >> You'd of course want to do the normal sanitization of user data. >> >> I've got a script that I could modify to do this, let me know if you >> need/want it. >> >> Cheers, >> Harry >> > > > > ok - i think i'm getting this thing now. nice - if you don't mind > posting that script i'd love to take a look at it. > > -p > Harry has the right idea -- still though, CGI scripts should be running as the apache user and not root. This means they won't (by design) have access to modify the cobbler configuration. What you would really want to do is write a simple script that can /only/ disable the netboot field and then grant SSH access for only that one command. There is some example of that technique posted here, which I personally haven't used, but I have it on good authority that it works well :) http://www.mythic-beasts.com/support/dyndns_howto.html This way (writing a script that calls "cobbler system edit --name=name --netboot-enabled=0") you make sure you've allowed remote access to changing only that one specific flag. (This particular flag has the result of removing the per-system configuration file in /tftpboot that enables the system to boot to a specific PXE target) Incidentally, Matt Hyclak wrote a script to do this before you could do this in the cobbler command line. That script is mentioned on this page: https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerApi The alternative is to SSH is to make the cgi to do this setuid root, which has security implications. Another (perhaps simpler) option is set network boot lower in the BIOS order (so hard drives first), and then when it comes time to reinstall them, you can use "koan --replace-self --server=bootserver.example.com --profile=name" to do the reinstall rather than needing to PXE. If the Linux box is already running, you can invoke that koan call over SSH followed by a call to /sbin/reboot. That will essentially do the same thing, and is what I do and generally recommend. > > > >> >> > Harry Hoffman wrote: >> >> >> >> Hi Peter, >> >> >> >> I had this same problem... it should be said that with >= cobbler-0.5 >> >> there is a option to edit the system: >> >> cobbler system add --name=string --profile=string [--mac=macaddress] >> >> [--ip=ipaddress] [--hostname=hostname] [--kopts=string] >> [--ipad- >> >> dress=string] [--ksmeta=string] [--netboot-enabled=Y/N >> >> >> >> The nice thing about cobbler via (git - yeah, it's a messed up >> name) is >> >> that a make in the d/l'd src directory will build you a rpm with >> proper >> >> version so that yum upgrades will overwrite it. >> >> >> >> It's a pretty trivial process, feel free to ask questions. >> >> >> > >> > Awesome, thanks Harry - I'll start diving into this tomorrow then. >> Just >> > to make sure I understand clearly. With the newer version I should be >> > able to run something like this during %post: >> > >> > cobbler system edit --name=$MAC --netboot-enabled=N >> > >> > -pete >> > >> >> >> >> >> >> > hi all, >> >> > i'm currently working on moving a cluster over from Xcat to >> cobbler. >> >> so >> >> > far things have gone quite smoothly, i have imported several >> distros - >> >> > created my own distros and gotten my custom kickstart's working >> quite >> >> > easilly! >> >> > >> >> > my question is i have not figured out how one has an >> installation dial >> >> > back to the cobbler master node and tell it that it no longer >> needs to >> >> > to re-install itself. with Xcat during the %post phase your node >> >> would >> >> > set it's status on the Xcat master node to boot locally after >> install >> >> - >> >> > is there something similar for cobbler? >> >> > >> >> > hopefully i'm missing something basic here, but have had no luck >> >> reading >> >> > through the man pages or mailing list archives. >> >> > >> >> > thanks! >> >> > -pete >> >> > >> >> > -- >> >> > Peter Wright >> >> > Systems Administrator >> >> > Sony Pictures Imageworks >> >> > wright at imageworks.com >> >> > www.imageworks.com >> >> > >> >> > >> >> > _______________________________________________ >> >> > et-mgmt-tools mailing list >> >> > et-mgmt-tools at redhat.com >> >> > https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> >> > >> >> >> >> >> >> _______________________________________________ >> >> et-mgmt-tools mailing list >> >> et-mgmt-tools at redhat.com >> >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> >> >> > >> > >> > -- >> > Peter Wright >> > Systems Administrator >> > Sony Pictures Imageworks >> > wright at imageworks.com >> > www.imageworks.com >> > >> > >> > >> >> >> > > From mdehaan at redhat.com Fri Aug 3 14:57:17 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 03 Aug 2007 10:57:17 -0400 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B33E87.50701@redhat.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> <46B286A2.4080001@imageworks.com> <46B33E87.50701@redhat.com> Message-ID: <46B3424D.9050907@redhat.com> Michael DeHaan wrote: > Peter Wright wrote: >> Harry Hoffman wrote: >>> >>> Hi Peter, >>> >>> No, I don't think that'll work... but if you wrote a quick cgi to >>> accept >>> the args of mac addr and netboot then you could exec the cobbler >>> command... >>> >>> something like >>> >>> %post >>> >>> wget >>> http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n >>> >>> >>> You'd of course want to do the normal sanitization of user data. >>> >>> I've got a script that I could modify to do this, let me know if you >>> need/want it. >>> >>> Cheers, >>> Harry >>> >> >> >> >> ok - i think i'm getting this thing now. nice - if you don't mind >> posting that script i'd love to take a look at it. >> >> -p >> > Harry has the right idea -- still though, CGI scripts should be > running as the apache user and not root. This means they won't (by > design) have access to modify the cobbler configuration. What you > would really want to do is write a simple script that can /only/ > disable the netboot field and then grant SSH access for only that one > command. There is some example of that technique posted here, which > I personally haven't used, but I have it on good authority that it > works well :) > > http://www.mythic-beasts.com/support/dyndns_howto.html > > This way (writing a script that calls "cobbler system edit --name=name > --netboot-enabled=0") you make sure you've allowed remote access to > changing only that one > specific flag. (This particular flag has the result of removing the > per-system configuration file in /tftpboot that enables the system to > boot to a specific PXE target) > > Incidentally, Matt Hyclak wrote a script to do this before you could > do this in the cobbler command line. That script is mentioned on this > page: > > https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerApi > > The alternative is to SSH is to make the cgi to do this setuid root, > which has security implications. > > Another (perhaps simpler) option is set network boot lower in the BIOS > order (so hard drives first), and then when it comes time to reinstall > them, you can use > "koan --replace-self --server=bootserver.example.com --profile=name" > to do the reinstall rather than needing to PXE. If the Linux box is > already running, you can invoke that koan call over SSH followed by a > call to /sbin/reboot. That will essentially do the same thing, and > is what I do and generally recommend. > > Talking on IRC, We've decided we're going to implement the following: When /var/lib/cobbler/settings parameter "pxe_just_once" is set to 1, we're going to add a line to the bottom of the kickstart to call a CGI script. (You'll also have to add a sudoers entry that we can define in the manpage) There's going to be a wget to a CGI script that takes a cobbler system name as a parameter. This script invokes cobbler_set_netboot via sudo, which can do nothing else but toggle the netboot-enabled parameter. cobbler_set_netboot has permissions to only toggle the netboot flag So, in summary, all a user will need to do is: -- flip the pxe_just_once switch in the settings file -- add a sudoers entry (cobbler check can even show the user what this entry must look like) We can do this :) > >> >> >> >>> >>> > Harry Hoffman wrote: >>> >> >>> >> Hi Peter, >>> >> >>> >> I had this same problem... it should be said that with >= >>> cobbler-0.5 >>> >> there is a option to edit the system: >>> >> cobbler system add --name=string --profile=string [--mac=macaddress] >>> >> [--ip=ipaddress] [--hostname=hostname] [--kopts=string] >>> [--ipad- >>> >> dress=string] [--ksmeta=string] [--netboot-enabled=Y/N >>> >> >>> >> The nice thing about cobbler via (git - yeah, it's a messed up >>> name) is >>> >> that a make in the d/l'd src directory will build you a rpm with >>> proper >>> >> version so that yum upgrades will overwrite it. >>> >> >>> >> It's a pretty trivial process, feel free to ask questions. >>> >> >>> > >>> > Awesome, thanks Harry - I'll start diving into this tomorrow >>> then. Just >>> > to make sure I understand clearly. With the newer version I >>> should be >>> > able to run something like this during %post: >>> > >>> > cobbler system edit --name=$MAC --netboot-enabled=N >>> > >>> > -pete >>> > >>> >> >>> >> >>> >> > hi all, >>> >> > i'm currently working on moving a cluster over from Xcat to >>> cobbler. >>> >> so >>> >> > far things have gone quite smoothly, i have imported several >>> distros - >>> >> > created my own distros and gotten my custom kickstart's working >>> quite >>> >> > easilly! >>> >> > >>> >> > my question is i have not figured out how one has an >>> installation dial >>> >> > back to the cobbler master node and tell it that it no longer >>> needs to >>> >> > to re-install itself. with Xcat during the %post phase your node >>> >> would >>> >> > set it's status on the Xcat master node to boot locally after >>> install >>> >> - >>> >> > is there something similar for cobbler? >>> >> > >>> >> > hopefully i'm missing something basic here, but have had no luck >>> >> reading >>> >> > through the man pages or mailing list archives. >>> >> > >>> >> > thanks! >>> >> > -pete >>> >> > >>> >> > -- >>> >> > Peter Wright >>> >> > Systems Administrator >>> >> > Sony Pictures Imageworks >>> >> > wright at imageworks.com >>> >> > www.imageworks.com >>> >> > >>> >> > >>> >> > _______________________________________________ >>> >> > et-mgmt-tools mailing list >>> >> > et-mgmt-tools at redhat.com >>> >> > https://www.redhat.com/mailman/listinfo/et-mgmt-tools >>> >> > >>> >> >>> >> >>> >> _______________________________________________ >>> >> et-mgmt-tools mailing list >>> >> et-mgmt-tools at redhat.com >>> >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >>> >> >>> > >>> > >>> > -- >>> > Peter Wright >>> > Systems Administrator >>> > Sony Pictures Imageworks >>> > wright at imageworks.com >>> > www.imageworks.com >>> > >>> > >>> > >>> >>> >>> >> >> > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From mdehaan at redhat.com Fri Aug 3 15:40:23 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 03 Aug 2007 11:40:23 -0400 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B3424D.9050907@redhat.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> <46B286A2.4080001@imageworks.com> <46B33E87.50701@redhat.com> <46B3424D.9050907@redhat.com> Message-ID: <46B34C67.8070901@redhat.com> Michael DeHaan wrote: > Michael DeHaan wrote: >> Peter Wright wrote: >>> Harry Hoffman wrote: >>>> >>>> Hi Peter, >>>> >>>> No, I don't think that'll work... but if you wrote a quick cgi to >>>> accept >>>> the args of mac addr and netboot then you could exec the cobbler >>>> command... >>>> >>>> something like >>>> >>>> %post >>>> >>>> wget >>>> http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n >>>> >>>> >>>> You'd of course want to do the normal sanitization of user data. >>>> >>>> I've got a script that I could modify to do this, let me know if you >>>> need/want it. >>>> >>>> Cheers, >>>> Harry >>>> >>> >>> >>> >>> ok - i think i'm getting this thing now. nice - if you don't mind >>> posting that script i'd love to take a look at it. >>> >>> -p >>> >> Harry has the right idea -- still though, CGI scripts should be >> running as the apache user and not root. This means they won't (by >> design) have access to modify the cobbler configuration. What you >> would really want to do is write a simple script that can /only/ >> disable the netboot field and then grant SSH access for only that one >> command. There is some example of that technique posted here, which >> I personally haven't used, but I have it on good authority that it >> works well :) >> >> http://www.mythic-beasts.com/support/dyndns_howto.html >> >> This way (writing a script that calls "cobbler system edit >> --name=name --netboot-enabled=0") you make sure you've allowed remote >> access to changing only that one >> specific flag. (This particular flag has the result of removing the >> per-system configuration file in /tftpboot that enables the system to >> boot to a specific PXE target) >> >> Incidentally, Matt Hyclak wrote a script to do this before you could >> do this in the cobbler command line. That script is mentioned on >> this page: >> >> https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerApi >> >> The alternative is to SSH is to make the cgi to do this setuid root, >> which has security implications. >> >> Another (perhaps simpler) option is set network boot lower in the >> BIOS order (so hard drives first), and then when it comes time to >> reinstall them, you can use >> "koan --replace-self --server=bootserver.example.com --profile=name" >> to do the reinstall rather than needing to PXE. If the Linux box is >> already running, you can invoke that koan call over SSH followed by a >> call to /sbin/reboot. That will essentially do the same thing, and >> is what I do and generally recommend. >> >> > > Talking on IRC, > > We've decided we're going to implement the following: > > When /var/lib/cobbler/settings parameter "pxe_just_once" is set to 1, > we're going to add a line to the bottom of the kickstart to call > a CGI script. (You'll also have to add a sudoers entry that we can > define in the manpage) > > There's going to be a wget to a CGI script that takes a cobbler system > name as a parameter. > This script invokes cobbler_set_netboot via sudo, which can do nothing > else but toggle the netboot-enabled parameter. > cobbler_set_netboot has permissions to only toggle the netboot flag > > So, in summary, all a user will need to do is: > -- flip the pxe_just_once switch in the settings file > -- add a sudoers entry (cobbler check can even show the user what this > entry must look like) > > We can do this :) > > I was overthinking this. We can use the XMLRPC interface to help out the CGI script and eliminate the sudo problem altogether. Even better... > > >> >>> >>> >>> >>>> >>>> > Harry Hoffman wrote: >>>> >> >>>> >> Hi Peter, >>>> >> >>>> >> I had this same problem... it should be said that with >= >>>> cobbler-0.5 >>>> >> there is a option to edit the system: >>>> >> cobbler system add --name=string --profile=string >>>> [--mac=macaddress] >>>> >> [--ip=ipaddress] [--hostname=hostname] [--kopts=string] >>>> [--ipad- >>>> >> dress=string] [--ksmeta=string] [--netboot-enabled=Y/N >>>> >> >>>> >> The nice thing about cobbler via (git - yeah, it's a messed up >>>> name) is >>>> >> that a make in the d/l'd src directory will build you a rpm with >>>> proper >>>> >> version so that yum upgrades will overwrite it. >>>> >> >>>> >> It's a pretty trivial process, feel free to ask questions. >>>> >> >>>> > >>>> > Awesome, thanks Harry - I'll start diving into this tomorrow >>>> then. Just >>>> > to make sure I understand clearly. With the newer version I >>>> should be >>>> > able to run something like this during %post: >>>> > >>>> > cobbler system edit --name=$MAC --netboot-enabled=N >>>> > >>>> > -pete >>>> > >>>> >> >>>> >> >>>> >> > hi all, >>>> >> > i'm currently working on moving a cluster over from Xcat to >>>> cobbler. >>>> >> so >>>> >> > far things have gone quite smoothly, i have imported several >>>> distros - >>>> >> > created my own distros and gotten my custom kickstart's >>>> working quite >>>> >> > easilly! >>>> >> > >>>> >> > my question is i have not figured out how one has an >>>> installation dial >>>> >> > back to the cobbler master node and tell it that it no longer >>>> needs to >>>> >> > to re-install itself. with Xcat during the %post phase your node >>>> >> would >>>> >> > set it's status on the Xcat master node to boot locally after >>>> install >>>> >> - >>>> >> > is there something similar for cobbler? >>>> >> > >>>> >> > hopefully i'm missing something basic here, but have had no luck >>>> >> reading >>>> >> > through the man pages or mailing list archives. >>>> >> > >>>> >> > thanks! >>>> >> > -pete >>>> >> > >>>> >> > -- >>>> >> > Peter Wright >>>> >> > Systems Administrator >>>> >> > Sony Pictures Imageworks >>>> >> > wright at imageworks.com >>>> >> > www.imageworks.com >>>> >> > >>>> >> > >>>> >> > _______________________________________________ >>>> >> > et-mgmt-tools mailing list >>>> >> > et-mgmt-tools at redhat.com >>>> >> > https://www.redhat.com/mailman/listinfo/et-mgmt-tools >>>> >> > >>>> >> >>>> >> >>>> >> _______________________________________________ >>>> >> et-mgmt-tools mailing list >>>> >> et-mgmt-tools at redhat.com >>>> >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >>>> >> >>>> > >>>> > >>>> > -- >>>> > Peter Wright >>>> > Systems Administrator >>>> > Sony Pictures Imageworks >>>> > wright at imageworks.com >>>> > www.imageworks.com >>>> > >>>> > >>>> > >>>> >>>> >>>> >>> >>> >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From wright at imageworks.com Fri Aug 3 15:46:27 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 03 Aug 2007 08:46:27 -0700 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B3424D.9050907@redhat.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> <46B286A2.4080001@imageworks.com><46B33E87.50701@redhat.com> <46B3424D.9050907@redhat.com> Message-ID: <46B34DD3.3020505@imageworks.com> Michael DeHaan wrote: > > Michael DeHaan wrote: > > Peter Wright wrote: > >> Harry Hoffman wrote: > >>> > >>> Hi Peter, > >>> > >>> No, I don't think that'll work... but if you wrote a quick cgi to > >>> accept > >>> the args of mac addr and netboot then you could exec the cobbler > >>> command... > >>> > >>> something like > >>> > >>> %post > >>> > >>> wget > >>> > http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n > > >>> > > > >>> > >>> You'd of course want to do the normal sanitization of user data. > >>> > >>> I've got a script that I could modify to do this, let me know if you > >>> need/want it. > >>> > >>> Cheers, > >>> Harry > >>> > >> > >> > >> > >> ok - i think i'm getting this thing now. nice - if you don't mind > >> posting that script i'd love to take a look at it. > >> > >> -p > >> > > Harry has the right idea -- still though, CGI scripts should be > > running as the apache user and not root. This means they won't (by > > design) have access to modify the cobbler configuration. What you > > would really want to do is write a simple script that can /only/ > > disable the netboot field and then grant SSH access for only that one > > command. There is some example of that technique posted here, which > > I personally haven't used, but I have it on good authority that it > > works well :) > > > > http://www.mythic-beasts.com/support/dyndns_howto.html > > > > This way (writing a script that calls "cobbler system edit --name=name > > --netboot-enabled=0") you make sure you've allowed remote access to > > changing only that one > > specific flag. (This particular flag has the result of removing the > > per-system configuration file in /tftpboot that enables the system to > > boot to a specific PXE target) > > > > Incidentally, Matt Hyclak wrote a script to do this before you could > > do this in the cobbler command line. That script is mentioned on this > > page: > > > > https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerApi > > > > The alternative is to SSH is to make the cgi to do this setuid root, > > which has security implications. > > > > Another (perhaps simpler) option is set network boot lower in the BIOS > > order (so hard drives first), and then when it comes time to reinstall > > them, you can use > > "koan --replace-self --server=bootserver.example.com --profile=name" > > to do the reinstall rather than needing to PXE. If the Linux box is > > already running, you can invoke that koan call over SSH followed by a > > call to /sbin/reboot. That will essentially do the same thing, and > > is what I do and generally recommend. > > > > > > Talking on IRC, > > We've decided we're going to implement the following: > > When /var/lib/cobbler/settings parameter "pxe_just_once" is set to 1, > we're going to add a line to the bottom of the kickstart to call > a CGI script. (You'll also have to add a sudoers entry that we can > define in the manpage) > > There's going to be a wget to a CGI script that takes a cobbler system > name as a parameter. > This script invokes cobbler_set_netboot via sudo, which can do nothing > else but toggle the netboot-enabled parameter. > cobbler_set_netboot has permissions to only toggle the netboot flag > > So, in summary, all a user will need to do is: > -- flip the pxe_just_once switch in the settings file > -- add a sudoers entry (cobbler check can even show the user what this > entry must look like) > > We can do this :) > OK, this sounds like a reasonable solution for me, thanks all! -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From mdehaan at redhat.com Fri Aug 3 16:04:37 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 03 Aug 2007 12:04:37 -0400 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B34DD3.3020505@imageworks.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> <46B286A2.4080001@imageworks.com><46B33E87.50701@redhat.com> <46B3424D.9050907@redhat.com> <46B34DD3.3020505@imageworks.com> Message-ID: <46B35215.5020306@redhat.com> Peter Wright wrote: > Michael DeHaan wrote: >> >> Michael DeHaan wrote: >> > Peter Wright wrote: >> >> Harry Hoffman wrote: >> >>> >> >>> Hi Peter, >> >>> >> >>> No, I don't think that'll work... but if you wrote a quick cgi to >> >>> accept >> >>> the args of mac addr and netboot then you could exec the cobbler >> >>> command... >> >>> >> >>> something like >> >>> >> >>> %post >> >>> >> >>> wget >> >>> >> http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n >> >> >>> >> > > >> >>> >> >>> You'd of course want to do the normal sanitization of user data. >> >>> >> >>> I've got a script that I could modify to do this, let me know if you >> >>> need/want it. >> >>> >> >>> Cheers, >> >>> Harry >> >>> >> >> >> >> >> >> >> >> ok - i think i'm getting this thing now. nice - if you don't mind >> >> posting that script i'd love to take a look at it. >> >> >> >> -p >> >> >> > Harry has the right idea -- still though, CGI scripts should be >> > running as the apache user and not root. This means they won't (by >> > design) have access to modify the cobbler configuration. What you >> > would really want to do is write a simple script that can /only/ >> > disable the netboot field and then grant SSH access for only that one >> > command. There is some example of that technique posted here, which >> > I personally haven't used, but I have it on good authority that it >> > works well :) >> > >> > http://www.mythic-beasts.com/support/dyndns_howto.html >> > >> > This way (writing a script that calls "cobbler system edit --name=name >> > --netboot-enabled=0") you make sure you've allowed remote access to >> > changing only that one >> > specific flag. (This particular flag has the result of removing the >> > per-system configuration file in /tftpboot that enables the system to >> > boot to a specific PXE target) >> > >> > Incidentally, Matt Hyclak wrote a script to do this before you could >> > do this in the cobbler command line. That script is mentioned on this >> > page: >> > >> > https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerApi >> > >> > The alternative is to SSH is to make the cgi to do this setuid root, >> > which has security implications. >> > >> > Another (perhaps simpler) option is set network boot lower in the BIOS >> > order (so hard drives first), and then when it comes time to reinstall >> > them, you can use >> > "koan --replace-self --server=bootserver.example.com --profile=name" >> > to do the reinstall rather than needing to PXE. If the Linux box is >> > already running, you can invoke that koan call over SSH followed by a >> > call to /sbin/reboot. That will essentially do the same thing, and >> > is what I do and generally recommend. >> > >> > >> >> Talking on IRC, >> >> We've decided we're going to implement the following: >> >> When /var/lib/cobbler/settings parameter "pxe_just_once" is set to 1, >> we're going to add a line to the bottom of the kickstart to call >> a CGI script. (You'll also have to add a sudoers entry that we can >> define in the manpage) >> >> There's going to be a wget to a CGI script that takes a cobbler system >> name as a parameter. >> This script invokes cobbler_set_netboot via sudo, which can do nothing >> else but toggle the netboot-enabled parameter. >> cobbler_set_netboot has permissions to only toggle the netboot flag >> >> So, in summary, all a user will need to do is: >> -- flip the pxe_just_once switch in the settings file >> -- add a sudoers entry (cobbler check can even show the user what this >> entry must look like) >> >> We can do this :) >> > > > OK, this sounds like a reasonable solution for me, thanks all! > > -pete > > I just committed this to git, if anyone wants to try it out... All you have to do is change pxe_just_once to 1 in /var/lib/cobbler/settings and it works out of the box. The $kickstart_done symbol in the kickstart file is replaced automatically with the correct wget to make this happen. To make a system PXE again after it trips off, just run: cobbler system edit --name=foo --netboot-enabled=1 This behavior is off by default. --Michael From wright at imageworks.com Fri Aug 3 17:10:50 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 03 Aug 2007 10:10:50 -0700 Subject: [et-mgmt-tools] boot locally after install via cobbler In-Reply-To: <46B35215.5020306@redhat.com> References: <46B281CE.9080604@imageworks.com> <42399.68.80.142.12.1186104203.squirrel@www.ip-solutions.net> <46B284AB.8060104@imageworks.com> <41834.68.80.142.12.1186104881.squirrel@www.ip-solutions.net> <46B286A2.4080001@imageworks.com><46B33E87.50701@redhat.com> <46B3424D.9050907@redhat.com><46B34DD3.3020505@imageworks.com> <46B35215.5020306@redhat.com> Message-ID: <46B3619A.5070800@imageworks.com> Michael DeHaan wrote: > > Peter Wright wrote: > > Michael DeHaan wrote: > >> > >> Michael DeHaan wrote: > >> > Peter Wright wrote: > >> >> Harry Hoffman wrote: > >> >>> > >> >>> Hi Peter, > >> >>> > >> >>> No, I don't think that'll work... but if you wrote a quick cgi to > >> >>> accept > >> >>> the args of mac addr and netboot then you could exec the cobbler > >> >>> command... > >> >>> > >> >>> something like > >> >>> > >> >>> %post > >> >>> > >> >>> wget > >> >>> > >> http://cobbler/cgi-bin/done_install?mac=00:11:22:33:44:55&netboot=n > > >> > > > >> >>> > >> >> > >> > >> >>> > >> >>> You'd of course want to do the normal sanitization of user data. > >> >>> > >> >>> I've got a script that I could modify to do this, let me know > if you > >> >>> need/want it. > >> >>> > >> >>> Cheers, > >> >>> Harry > >> >>> > >> >> > >> >> > >> >> > >> >> ok - i think i'm getting this thing now. nice - if you don't mind > >> >> posting that script i'd love to take a look at it. > >> >> > >> >> -p > >> >> > >> > Harry has the right idea -- still though, CGI scripts should be > >> > running as the apache user and not root. This means they won't (by > >> > design) have access to modify the cobbler configuration. What you > >> > would really want to do is write a simple script that can /only/ > >> > disable the netboot field and then grant SSH access for only that one > >> > command. There is some example of that technique posted here, which > >> > I personally haven't used, but I have it on good authority that it > >> > works well :) > >> > > >> > http://www.mythic-beasts.com/support/dyndns_howto.html > >> > > >> > This way (writing a script that calls "cobbler system edit > --name=name > >> > --netboot-enabled=0") you make sure you've allowed remote access to > >> > changing only that one > >> > specific flag. (This particular flag has the result of removing the > >> > per-system configuration file in /tftpboot that enables the system to > >> > boot to a specific PXE target) > >> > > >> > Incidentally, Matt Hyclak wrote a script to do this before you could > >> > do this in the cobbler command line. That script is mentioned on > this > >> > page: > >> > > >> > https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerApi > >> > > >> > The alternative is to SSH is to make the cgi to do this setuid root, > >> > which has security implications. > >> > > >> > Another (perhaps simpler) option is set network boot lower in the > BIOS > >> > order (so hard drives first), and then when it comes time to > reinstall > >> > them, you can use > >> > "koan --replace-self --server=bootserver.example.com --profile=name" > >> > to do the reinstall rather than needing to PXE. If the Linux box is > >> > already running, you can invoke that koan call over SSH followed by a > >> > call to /sbin/reboot. That will essentially do the same thing, and > >> > is what I do and generally recommend. > >> > > >> > > >> > >> Talking on IRC, > >> > >> We've decided we're going to implement the following: > >> > >> When /var/lib/cobbler/settings parameter "pxe_just_once" is set to 1, > >> we're going to add a line to the bottom of the kickstart to call > >> a CGI script. (You'll also have to add a sudoers entry that we can > >> define in the manpage) > >> > >> There's going to be a wget to a CGI script that takes a cobbler system > >> name as a parameter. > >> This script invokes cobbler_set_netboot via sudo, which can do nothing > >> else but toggle the netboot-enabled parameter. > >> cobbler_set_netboot has permissions to only toggle the netboot flag > >> > >> So, in summary, all a user will need to do is: > >> -- flip the pxe_just_once switch in the settings file > >> -- add a sudoers entry (cobbler check can even show the user what this > >> entry must look like) > >> > >> We can do this :) > >> > > > > > > OK, this sounds like a reasonable solution for me, thanks all! > > > > -pete > > > > > > I just committed this to git, if anyone wants to try it out... > > All you have to do is change pxe_just_once to 1 in > /var/lib/cobbler/settings and it works out of the box. The > $kickstart_done > symbol in the kickstart file is replaced automatically with the correct > wget to make this happen. > > To make a system PXE again after it trips off, just run: > > cobbler system edit --name=foo --netboot-enabled=1 > > This behavior is off by default. > it works! i had to set my SELinux policy to permissive as i have not setup any custom rule-sets yet. i'll get those setup shortly. all i had to do was add the $kickstart_done line to my custom kickstart %post section and we are off to the races. thanks all! -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From hyclak at math.ohiou.edu Fri Aug 3 17:57:33 2007 From: hyclak at math.ohiou.edu (Matt Hyclak) Date: Fri, 3 Aug 2007 13:57:33 -0400 Subject: [et-mgmt-tools] Cobbler Makefile patch Message-ID: <20070803175733.GL5703@math.ohiou.edu> Just a suggested addition to the cobbler Makefile to create only an SRPM - that allows for building in chroots (e.g. mock) and not having to have everything installed on the build box. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263 -------------- next part -------------- --- a/Makefile +++ b/Makefile @@ -42,3 +42,15 @@ rpms: manpage sdist --define "_specdir %{_topdir}" \ --define "_sourcedir %{_topdir}" \ -ba cobbler.spec + +srpm: manpage sdist + mkdir -p rpm-build + cp dist/*.gz rpm-build/ + rpmbuild --define "_topdir %(pwd)/rpm-build" \ + --define "_builddir %{_topdir}" \ + --define "_rpmdir %{_topdir}" \ + --define "_srcrpmdir %{_topdir}" \ + --define '_rpmfilename %%{NAME}-%%{VERSION}-%%{RELEASE}.%%{ARCH}.rpm' \ + --define "_specdir %{_topdir}" \ + --define "_sourcedir %{_topdir}" \ + -bs --nodeps cobbler.spec From mdehaan at redhat.com Fri Aug 3 18:31:22 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 03 Aug 2007 14:31:22 -0400 Subject: [et-mgmt-tools] Cobbler Makefile patch In-Reply-To: <20070803175733.GL5703@math.ohiou.edu> References: <20070803175733.GL5703@math.ohiou.edu> Message-ID: <46B3747A.1050606@redhat.com> Matt Hyclak wrote: > Just a suggested addition to the cobbler Makefile to create only an SRPM - > that allows for building in chroots (e.g. mock) and not having to have > everything installed on the build box. > > Matt > > > ------------------------------------------------------------------------ > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools applied, thanks! From drew.einhorn at gmail.com Sun Aug 5 17:04:34 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Sun, 5 Aug 2007 11:04:34 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46AE51C9.3000706@redhat.com> References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> Message-ID: On 7/30/07, Michael DeHaan wrote: > > drew einhorn wrote: > I think it won't be too difficult to use VMperl to create scripts to > > make an appropriate dhcp reservations, or create a VM to order. > > Will let you know when I have something more concrete than plans. > Excellent. > Been making progress. What I need now is a python incantation to retrieve ks metadata, and virt info from a cobbler profile, so I can fill in line arguments on shell commands my script is issuing. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From mdehaan at redhat.com Mon Aug 6 14:45:30 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 06 Aug 2007 10:45:30 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> Message-ID: <46B7340A.30901@redhat.com> drew einhorn wrote: > > > On 7/30/07, *Michael DeHaan* > wrote: > > drew einhorn wrote: > > > I think it won't be too difficult to use VMperl to create > scripts to > > make an appropriate dhcp reservations, or create a VM to order. > > Will let you know when I have something more concrete than plans. > Excellent. > > > Been making progress. > > What I need now is a python incantation to retrieve ks metadata, > and virt info from a cobbler profile, so I can fill in line arguments > on shell commands my script is issuing. > > This might help get you started... what does the script do? import cobbler.api as cobbler_api api = cobbler_api.BootAPI() for x in api.profiles(): print x.to_datastruct() > > -- > Drew Einhorn > ------------------------------------------------------------------------ > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From jonathan.underwood at gmail.com Mon Aug 6 15:38:46 2007 From: jonathan.underwood at gmail.com (Jonathan Underwood) Date: Mon, 6 Aug 2007 16:38:46 +0100 Subject: [et-mgmt-tools] Adding -usbdevice=tablet for qemu domains Message-ID: <645d17210708060838u92e12a1j1060aeb95f1318e9@mail.gmail.com> Hi, I am using virt-manager to create a guest Windows XP installation using qemu-kvm. I am finding that the mouse doesn't work well in virt-manager sessions, but it works fine when starting a qemu SDL window from the command line. My reading leads me to understand this is because virt-manager sessions use VNC to communicate with the guest domain. Further reading about qemu-dm and VNC and mice suggests that the way around this ubiquitous problem is gotten around by passing usbdevice=tablet to qemu. So my question is, how to do this within virt-manager/virsh? Also, given that this problem won't go away until the VNC client is fixed, and that this same issue affects Xen domains, wouldn't this be a useful defaul anyway? Thanks, Jonathan From berrange at redhat.com Mon Aug 6 15:46:42 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Mon, 6 Aug 2007 16:46:42 +0100 Subject: [et-mgmt-tools] Adding -usbdevice=tablet for qemu domains In-Reply-To: <645d17210708060838u92e12a1j1060aeb95f1318e9@mail.gmail.com> References: <645d17210708060838u92e12a1j1060aeb95f1318e9@mail.gmail.com> Message-ID: <20070806154642.GE14393@redhat.com> On Mon, Aug 06, 2007 at 04:38:46PM +0100, Jonathan Underwood wrote: > Hi, > > I am using virt-manager to create a guest Windows XP installation > using qemu-kvm. I am finding that the mouse doesn't work well in > virt-manager sessions, but it works fine when starting a qemu SDL > window from the command line. My reading leads me to understand this > is because virt-manager sessions use VNC to communicate with the guest > domain. > > Further reading about qemu-dm and VNC and mice suggests that the way > around this ubiquitous problem is gotten around by passing > usbdevice=tablet to qemu. So my question is, how to do this within > virt-manager/virsh? You can't currently. We're working on this support for the next release though, at which point it'll be done automatically without need for any special action by the user. Dan -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From khpark at linuxone.co.kr Mon Aug 6 16:00:28 2007 From: khpark at linuxone.co.kr (khpark at linuxone.co.kr) Date: Tue, 7 Aug 2007 01:00:28 +0900 (KST) Subject: [et-mgmt-tools] RE: et-mgmt-tools Digest, Vol 12, Issue 9 Message-ID: <20070806160028.AD19C1317A@aspspam.linuxone.co.kr> An HTML attachment was scrubbed... URL: From drew.einhorn at gmail.com Mon Aug 6 16:46:08 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Mon, 6 Aug 2007 10:46:08 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46B7340A.30901@redhat.com> References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> Message-ID: I hoping that Real Soon Now it will create a VM to order with lots of custom vmware settings via ks_meta variables. At the moment I'm hung up on VMware API problems, so it doesn't do anything useful yet. The APIs either won't install or puke after they are installed. I can generate .vmx files to order, but scp and ssh commands to install them on the VMware server, either require permanently granting too much access to the server or you have to authenticate yourself several times. Pretty sure I understand most of the values you python script prints out. But what are: parent, depth, virt_type, and virt_path? On 8/6/07, Michael DeHaan wrote: > > drew einhorn wrote: > > > > > > On 7/30/07, *Michael DeHaan* > > wrote: > > > > drew einhorn wrote: > > > > > I think it won't be too difficult to use VMperl to create > > scripts to > > > make an appropriate dhcp reservations, or create a VM to order. > > > Will let you know when I have something more concrete than plans. > > Excellent. > > > > > > Been making progress. > > > > What I need now is a python incantation to retrieve ks metadata, > > and virt info from a cobbler profile, so I can fill in line arguments > > on shell commands my script is issuing. > > > > > > This might help get you started... what does the script do? > > > import cobbler.api as cobbler_api > > api = cobbler_api.BootAPI() > > for x in api.profiles(): > print x.to_datastruct() > > > > > > > > > > -- > > Drew Einhorn > > ------------------------------------------------------------------------ > > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From mdehaan at redhat.com Mon Aug 6 17:06:06 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 06 Aug 2007 13:06:06 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> Message-ID: <46B754FE.40801@redhat.com> drew einhorn wrote: > I hoping that Real Soon Now it will create a VM to order with lots of > custom vmware settings via ks_meta variables. Very cool. It seems like you'd really benefit from using the XMLRPC API (see cobblerd.py for details) instead then. This way you could install VMware on any machine, not just the machine that ran the Cobbler server. https://hosted.fedoraproject.org/projects/cobbler/wiki/CobblerXmlrpc Particularly, you'd want to use functions like get_system_for_koan(system_name) -- which still returns a hash/dictionary like the local API does. Not only does the XMLRPC API work remotely, but it also makes sure that all of the cobbler objects are blended together just like they would be if used from koan or PXE. Calling the local API does not ensure this, which means you'd just get the metadata about the system, but not the profile or distro object. koan is an example of an application using the Cobbler XMLRPC API and the source of that app could be useful. > > At the moment I'm hung up on VMware API problems, so it doesn't do > anything useful yet. The APIs either won't install or puke > after they are installed. I can generate .vmx files to order, but > scp and ssh commands to install them on the VMware server, > either require permanently granting too much access to the server or > you have to authenticate yourself several times. > > Pretty sure I understand most of the values you python script > prints out. > > But what are: parent, depth, virt_type, and virt_path? parent/depth are related to the ability to inherit profiles. 'parent' can be set on a profile to mean that the profile should take any unspecified settings from another profile. For instance, "rhel5webserver" might inherit from "rhel5server". This is documented as "--parent" in the manpage. Depth is just an internal variable used for being able to make "cobbler list" formatting of the cobbler object graph look correct. It is not really useful in any context other than trying to render/sort the object graph -- something most apps don't need to do. Again, if you're writing a client of cobbler, usage of the XMLRPC API and using the *_for_koan functions is the best way to go. It's simpler -- and it's also usable remotely (and without the need for root access). > > > > On 8/6/07, *Michael DeHaan* < mdehaan at redhat.com > > wrote: > > drew einhorn wrote: > > > > > > On 7/30/07, *Michael DeHaan* > > >> wrote: > > > > drew einhorn wrote: > > > > > I think it won't be too difficult to use VMperl to create > > scripts to > > > make an appropriate dhcp reservations, or create a VM to > order. > > > Will let you know when I have something more concrete than > plans. > > Excellent. > > > > > > Been making progress. > > > > What I need now is a python incantation to retrieve ks metadata, > > and virt info from a cobbler profile, so I can fill in line > arguments > > on shell commands my script is issuing. > > > > > > This might help get you started... what does the script do? > > > import cobbler.api as cobbler_api > > api = cobbler_api.BootAPI() > > for x in api.profiles(): > print x.to_datastruct() > > > > > > > > > > -- > > Drew Einhorn > > > ------------------------------------------------------------------------ > > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > > > -- > Drew Einhorn > ------------------------------------------------------------------------ > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From mdehaan at redhat.com Mon Aug 6 17:17:19 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 06 Aug 2007 13:17:19 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> Message-ID: <46B7579F.8040108@redhat.com> > > But what are: parent, depth, virt_type, and virt_path? > So I forgot to answer the last two. --virt-type (see manpage). This specifies whether (if being installed virtually) the guest should be Xen paravirt ("xenpv") or qemu/KVM ("qemu"). There is also a new notion of "auto" which means, "see what I have installed and if the kernels are appropriate". If your script gets rolled into koan, we'd also add a "vmware" here, though you can ignore this for now. If --virt-type is not specified in Cobbler, it can also be overriden on the koan command line. --virt-path (see manpage). This is related to the patch Adam Rosenwald sent over. Basically it allows for specifying where the disk image file (or paritition) needs to go. This also might specify a LVM volume group. Again, syntax on valid options is documented in the manpage. If left blank, koan decides the defaults. If --virt-path doesn't make sense for VMware, you can ignore it for now. It might be something we would want to do later. From rjones at redhat.com Mon Aug 6 21:15:48 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Mon, 06 Aug 2007 22:15:48 +0100 Subject: [et-mgmt-tools] ANNOUNCE: Virt-top - a top-like utility for displaying virtualization stats Message-ID: <46B78F84.10309@redhat.com> I'm pleased to announce the first release of virt-top, which is a top-like utility for displaying virtualization stats. It aims to look and feel very much like regular 'top', so as to be as familiar as possible for systems administrators. You can also use it as a pleasant replacement for xentop. It uses libvirt, so can display stats across a variety of different hypervisors and virtualization systems (not just Xen, although that is where the testing has gone so far). http://et.redhat.com/~rjones/virt-top/ The license is a combination of LGPL (for the library) and GPL (for the virt-top program). Current status -------------- You can view domains and use familiar keys like 'P'/'M'/... to sort by processor/memory/..., and 'd'/'s' to set the delay between updates. Also some common top command-line options are implemented. The man page is here: http://et.redhat.com/~rjones/virt-top/virt-top.txt There are a variety of source and binary RPMs available for Fedora users. I don't yet have a working Debian/Ubuntu package, but will have a go at making one tomorrow. The next thing I'll be working on is showing virtual and physical CPU usage of guests. After that I'm hoping to discuss extensions to libvirt to make other interesting statistics available to virt-top, in particular disk and network I/O stats. There are a few data collection artifacts which need to be investigated. In particular, %CPU sometimes goes over 100%. Obviously accurate data collection is an important goal for this tool. Memory usage is good: typical 'RES' (in regular top) for virt-top is just under 2.5 MB, and I've had it running for hours at a time without memory usage increasing, which seems to indicate that there aren't any major memory leaks. Development ----------- The program is currently very small: exactly 500 lines of code! If you want to dive in and send me patches they are most welcome, but remember that I'm trying to make this utility act as much like 'top' as possible, so if 'top' does it in a particular way, then I'd prefer virt-top to do the same thing. (May not apply to obscure top functionality, if top does something silly). If you're a systems administrator, not a programmer, and you'd like virt-top to have some particular feature, then please let me know, and I will be happy to review it and code it for you. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: virt-top.png Type: image/png Size: 38338 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From berrange at redhat.com Mon Aug 6 21:31:24 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Mon, 6 Aug 2007 22:31:24 +0100 Subject: [et-mgmt-tools] Re: [Libvir] ANNOUNCE: Virt-top - a top-like utility for displaying virtualization stats In-Reply-To: <46B78F84.10309@redhat.com> References: <46B78F84.10309@redhat.com> Message-ID: <20070806213124.GG30665@redhat.com> On Mon, Aug 06, 2007 at 10:15:48PM +0100, Richard W.M. Jones wrote: > There are a few data collection artifacts which need to be investigated. > In particular, %CPU sometimes goes over 100%. Obviously accurate data > collection is an important goal for this tool. AFAICT it is impossible to stop it going over 100% - we have same issue in virt-manager when calculating CPU usage. To caculate the % usage you are looking at the differential between CPU time from two calls to virDomainGetInfo vs the time period between two gettimeofday() calls. If the time delta between your gettimeofday() calls does not exactly match the time delta between your virDomainGetInfo() calls, and the domain you're measuring is very active then you can end up calcuating > 100% occassionally. Not by much - typically < 1% over, but I don't really see any way around it. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From tdiehl at rogueind.com Wed Aug 8 05:09:19 2007 From: tdiehl at rogueind.com (Tom Diehl) Date: Wed, 8 Aug 2007 01:09:19 -0400 (EDT) Subject: [et-mgmt-tools] Cobbler check backtrace Message-ID: Hi, I just upgraded to cobbler-0.5.1-1. When I run cobbler check I get the following backtrace: (rocky pts13) # cobbler check Traceback (most recent call last): File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 724, in main BootCLI(sys.argv).run() File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 115, in run self.relay_args(self.args[1:], self.commands['toplevel']) File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 550, in relay_args commands[args[0]](args[1:]) File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 587, in check status = self.api.check() File "/usr/lib/python2.3/site-packages/cobbler/api.py", line 118, in check return check.run() File "/usr/lib/python2.3/site-packages/cobbler/action_check.py", line 51, in run self.check_tftpd_conf(status) File "/usr/lib/python2.3/site-packages/cobbler/action_check.py", line 129, in check_tftpd_conf status.append(_("change 'server_args' to '-s %(args)' in %(file)s") % { "file" : self.settings.tftpboot, "args" : self.settings.tftpboot }) ValueError: unsupported format character ''' (0x27) at index 35 (rocky pts13) # It would appear that this is caused by the following line in my settings file: ftpboot: /tftpboot/pxe I know that adding the pxe dir in the above path is wrong as far as cobbler check is concerned but other than complaining about it cobbler never backtraced with the previous version[1]. For my particular setup the ftpboot path is correct. FWIW This machine is running EL4. Should I put this in bugzilla? Regards, Tom 1. "previous version" in this case == cobbler-0.3.3-1 From mdehaan at redhat.com Wed Aug 8 14:17:43 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 08 Aug 2007 10:17:43 -0400 Subject: [et-mgmt-tools] Cobbler check backtrace In-Reply-To: References: Message-ID: <46B9D087.6010203@redhat.com> Tom Diehl wrote: > > I just upgraded to cobbler-0.5.1-1. When I run cobbler check I get the > following > backtrace: > > Should I put this in bugzilla? No need to, I've fixed it already... Thanks for the report! http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h=4f9df1448b0a61a2d431e5de35ff03f9df48b9f7 For what it's worth, 0.5.1-1 is/was a testing release. I'm planning to release 0.6.0 (stable) towards the end of this week. If you checkout and build from the latest source you'll essentially get the above fix and 99.9% (maybe 100%) of what will be in 0.6.0. --Michael From mdehaan at redhat.com Wed Aug 8 16:06:54 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 08 Aug 2007 12:06:54 -0400 Subject: [et-mgmt-tools] Using Cobbler with KVM Message-ID: <46B9EA1E.8080400@redhat.com> Hi folks, I ran across a problem where virtualized installs with koan using KVM on F-7 result in a guest that can't be restarted post-install. Here's a patch to the virtinst library from Dan Berrange to fix that. --- /usr/lib/python2.5/site-packages/virtinst/DistroManager.py 2007-08-08 11:47:52.000000000 -0400 +++ /tmp/fix.py 2007-08-08 11:48:18.000000000 -0400 @@ -685,21 +685,21 @@ if hvm: type = "hvm" else: type = "linux" if arch: osblob += " %s\n" % (arch, type) else: osblob += " %s\n" % type - if self.install["kernel"]: + if install and self.install["kernel"]: osblob += " %s\n" % self.install["kernel"] osblob += " %s\n" % self.install["initrd"] osblob += " %s\n" % self.install["extraargs"] else: if loader: osblob += " %s\n" % loader if install: osblob += " \n" else: ~ ~ To review, the syntax to install via KVM with a newer cobbler/koan is: koan --virt [--virt-type=qemu] --profile=profile_name --server=bootserver.example.com You can use the same profiles that you use to boot baremetal to install via qemu/KVM, don't use the ones with "xen" kernels. Note that you can also set --virt-type in cobbler, or even change the default virt type in /var/lib/cobbler/settings to default all new installs with koan to use KVM. --Michael From drew.einhorn at gmail.com Wed Aug 8 16:32:35 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Wed, 8 Aug 2007 10:32:35 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46B754FE.40801@redhat.com> References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> Message-ID: On 8/6/07, Michael DeHaan wrote: > Again, if you're writing a client of cobbler, usage of the XMLRPC API > and using the *_for_koan > functions is the best way to go. It's simpler -- and it's also usable > remotely (and without the need for root access). The cobbler XMLRPC API is much better. What I have so far is not a cobbler client. It's a python program that sets it's defaults, then add/overirde settings from cobbler, cobbler's blended settings via XML-RPC make this even cooler, then add/override settings from command line arguments, then a little bit of code fine tunes the settings. then call the validation code to check the setting. Then we generat a .vmx file for the VM. scp the .vmx to the vmware server, ssh some commands to create the virtual disk and register the VM. then we run a cobbler system add on the local host gotta add the cobbler sync then we power on the VM. And off we go. Maybe not looks like I've got a problem with dnsmasq. Been working on the vmware side an have tried it since I switched to dnsmasq The VM is PXE booting with the right mac address With tcpdump inside the cobbler host I see a bunch of these 09:16:31.293655 rarp who-is 00:50:56:00:00:ff (oui Unknown) tell 00:50:56:00:00:ff (oui Unknown) 09:16:31.294312 IP 0.0.0.0.bootpc > 255.255.255.255.bootps: BOOTP/DHCP, Request from 00:50:56:00:00:ff (oui Unknown), length: 548 they got through the firewall, it was willing to be it was a firewall problem. and we are not aswering. Here's an excerpt from cobbler report system : ctl01 profile : ctl kernel options : {} ks metadata : {} ip address : 10.0.0.66 mac address : 00:50:56:00:00:ff hostname : ctl01.sms pxe info set? : True config id : 01-00-50-56-00-00-ff netboot enabled? : 1 kickstart : <> virt type : <> virt path : <> Ant the output from #cobbler sync sync distro: CentOS5-i386 sync profile: ctl sync profile: smstrac sync profile: CentOS5-i386 sync profile: o3d sync profile: gtm sync system: 00:50:56:00:00:01 sync system: ctl01 Shutting down Lightweight caching nameserver (dnsmasq): [ OK ] Starting Lightweight caching nameserver (dnsmasq): [ OK ] /tmp/tftpboot/pxelinux.cfg/01-00-50-56-00-00-ff has default linux prompt 0 timeout 1 label linux kernel /images/CentOS5-i386/vmlinuz append ksdevice=eth0 lang= text syslog=10.0.0.21:25150 initrd=/images/CentOS5-i386/initrd.img ks= http://10.0.0.21/cblr/kickstarts_sys/ctl01/ks.cfg looks good let's cycle power on the VM still no dhcp offers. let's do a ps dnsmasq is still running let's look in the log files don't see anything at all about dnsmasq maybe I'm not looking in the right place. Not sure when I'll get back to this. Got a lot of stuff to do to get ready to go on vacation next week. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From mdehaan at redhat.com Wed Aug 8 18:34:54 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 08 Aug 2007 14:34:54 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> Message-ID: <46BA0CCE.7050406@redhat.com> drew einhorn wrote: > > > The cobbler XMLRPC API is much better. > > What I have so far is not a cobbler client. Understood -- you're doing the parts to set up the config + the client side bits to create the image. The later half seems like the part I'd want (in theory) to add to koan. (just the client side piece, which we could possibly invoke SSH if needed to make that controllable server side). Either way, yes, I'd love to see this when you've got it all working. > > Maybe not looks like I've got a problem with dnsmasq. > Been working on the vmware side an have tried it since > I switched to dnsmasq It's hard to tell for sure from what you've posted (lots of networking things could be afoot), but what does your /etc/cobbler/dnsmasq.template look like? How about the generated /etc/dnsmasq.conf? Those are a few things to start with. Since you didn't mention editing /etc/cobbler/dnsmasq.conf it might not be correct for your environment. There are some setup instructions including information at logging here: http://www.thekelleys.org.uk/dnsmasq/docs/setup.html -- keep in mind that cobbler will generate it's own dnsmasq.conf based on the template, and the same for /etc/ethers and /var/lib/cobbler/cobbler_hosts -- but that should give you some idea of how it all works. --Michael From tdsivakumar at gmail.com Thu Aug 9 11:42:31 2007 From: tdsivakumar at gmail.com (tdsivakumar amrita) Date: Thu, 9 Aug 2007 17:12:31 +0530 Subject: [et-mgmt-tools] regarding virtual machine manager Message-ID: <22b24e940708090442i25ca3a28h8d1b01ab0806c15b@mail.gmail.com> sir, I try to install windows 2003 using virtual machine manger.I installed fedroa7 in my machine.Using the virtual machine manager i tried to install.Igot all the menu form the cd 's but after the formating and copping over it shows like normal windows installations that i t is going to reboot.Afterthat the windows 2003 starts to install were it is telling that no media found.The media is in that right path .I am stuck in that area .Any help for any body could be good. -------------- next part -------------- An HTML attachment was scrubbed... URL: From rjones at redhat.com Thu Aug 9 14:31:24 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 09 Aug 2007 15:31:24 +0100 Subject: [et-mgmt-tools] ANNOUNCE: Virt-top 0.3.1.5 - a top-like utility for displaying virtualization stats In-Reply-To: <46B78F84.10309@redhat.com> References: <46B78F84.10309@redhat.com> Message-ID: <46BB253C.5010204@redhat.com> There is a significantly updated version of virt-top now available. On systems which support it, you can show the mapping of domains to physical CPUs and the amount of each physical CPU being used. Screenshots: http://et.redhat.com/~rjones/virt-top/screenshots.html There are also binaries available to download for i386 and x86-64. (The binaries are just self-contained Linux executables, no extra files or dependencies are needed, except for libvirt.so). Downloads: http://et.redhat.com/~rjones/virt-top/download.html If you are a systems administrator, what do you think could be improved about virt-top? Try it out and let me have your suggestions ... Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From drew.einhorn at gmail.com Thu Aug 9 15:58:47 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Thu, 9 Aug 2007 09:58:47 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BA0CCE.7050406@redhat.com> References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> Message-ID: On 8/8/07, Michael DeHaan wrote: > It's hard to tell for sure from what you've posted (lots of networking > things could be afoot), but what does your /etc/cobbler/dnsmasq.template > look like? How about the generated /etc/dnsmasq.conf? Those are a > few things to start with. Since you didn't mention editing > /etc/cobbler/dnsmasq.conf it might not be correct for your environment. > > There are some setup instructions including information at logging here: > http://www.thekelleys.org.uk/dnsmasq/docs/setup.html -- keep in mind > that cobbler will generate > it's own dnsmasq.conf based on the template, and the same for > /etc/ethers and /var/lib/cobbler/cobbler_hosts -- but that should give > you some idea of how it all works. > > > > --Michael > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > I've tried a few variations on /etc/cobbler/dnsmasq.template Here's my latest: $cat /etc/cobbler/dnsmasq.template # Cobbler generated configuration file for dnsmasq # $date # # resolve.conf .. ? #no-poll #enable-dbus read-ethers addn-hosts = /var/lib/cobbler/cobbler_hosts dhcp-range=10.0.0.128, 10.0.0.254 dhcp-option=3,$next_server dhcp-lease-max=1000 dhcp-authoritative dhcp-boot=pxelinux.0 dhcp-boot=net:normalarch,pxelinux.0 dhcp-boot=net:ia64,$elilo $insert_cobbler_system_definitions And the resulting $cat /etc/dnsmasq.conf # Cobbler generated configuration file for dnsmasq # Thu Aug 9 15:42:24 2007 # resolve.conf .. ? #no-poll #enable-dbus read-ethers addn-hosts = /var/lib/cobbler/cobbler_hosts dhcp-range=10.0.0.128, 10.0.0.254 dhcp-option=3,10.0.0.21 dhcp-lease-max=1000 dhcp-authoritative dhcp-boot=pxelinux.0 dhcp-boot=net:normalarch,pxelinux.0 dhcp-boot=net:ia64,elilo-3.6-ia64.efi and $ cat /etc/ethers 00:50:56:00:00:FF 10.0.0.66 and $ cat /var/lib/cobbler/cobbler_hosts 10.0.0.66 ctl01.sms Still pxeboot complains: No DHCP or proxyDHCP offers were received. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From mdehaan at redhat.com Thu Aug 9 16:31:44 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 09 Aug 2007 12:31:44 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> Message-ID: <46BB4170.60909@redhat.com> drew einhorn wrote: > > > On 8/8/07, *Michael DeHaan* > wrote: > > It's hard to tell for sure from what you've posted (lots of networking > things could be afoot), but what does your > /etc/cobbler/dnsmasq.template > look like? How about the generated /etc/dnsmasq.conf? Those are a > few things to start with. Since you didn't mention editing > /etc/cobbler/dnsmasq.conf it might not be correct for your > environment. > > There are some setup instructions including information at logging > here: > http://www.thekelleys.org.uk/dnsmasq/docs/setup.html -- keep in mind > that cobbler will generate > it's own dnsmasq.conf based on the template, and the same for > /etc/ethers and /var/lib/cobbler/cobbler_hosts -- but that should give > you some idea of how it all works. > > > > --Michael > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > I've tried a few variations on /etc/cobbler/dnsmasq.template > Here's my latest: > > $cat /etc/cobbler/dnsmasq.template > # Cobbler generated configuration file for dnsmasq > # $date > # > > # resolve.conf .. ? > #no-poll > #enable-dbus > read-ethers > addn-hosts = /var/lib/cobbler/cobbler_hosts > > dhcp-range=10.0.0.128 , 10.0.0.254 > dhcp-option=3,$next_server > dhcp-lease-max=1000 > dhcp-authoritative > dhcp-boot=pxelinux.0 > dhcp-boot=net:normalarch,pxelinux.0 > dhcp-boot=net:ia64,$elilo > > $insert_cobbler_system_definitions > > And the resulting > > $cat /etc/dnsmasq.conf > > # Cobbler generated configuration file for dnsmasq > # Thu Aug 9 15:42:24 2007 > > # resolve.conf .. ? > #no-poll > #enable-dbus > read-ethers > addn-hosts = /var/lib/cobbler/cobbler_hosts > > dhcp-range= 10.0.0.128 , 10.0.0.254 > dhcp-option=3,10.0.0.21 > dhcp-lease-max=1000 > dhcp-authoritative > dhcp-boot=pxelinux.0 > dhcp-boot=net:normalarch, pxelinux.0 > dhcp-boot=net:ia64,elilo-3.6-ia64.efi > > and > $ cat /etc/ethers > 00:50:56:00:00:FF 10.0.0.66 > > and > $ cat /var/lib/cobbler/cobbler_hosts > 10.0.0.66 ctl01.sms > > Still pxeboot complains: No DHCP or proxyDHCP offers were received. > > -- > Drew Einhorn > ------------------------------------------------------------------------ > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools Individual networking issues are beyond my ability to debug over email, and probably the scope of this list. You might want to see if you get the same problem with ISC and whether you get DHCP offers for your VMware guests when using your existing DHCP config. If any of those remove the problem, you have a dnsmasq config problem -- otherwise, it might be something else. I would think VMware would PXE just like anything else, but I have no idea what all the pieces to the puzzle are. Sorry I can't be of any more help. --Michael From mdehaan at redhat.com Thu Aug 9 16:32:44 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 09 Aug 2007 12:32:44 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BB4170.60909@redhat.com> References: <46AE4049.8020503@redhat.com> <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> Message-ID: <46BB41AC.1050009@redhat.com> Michael DeHaan wrote: > drew einhorn wrote: >> >> >> On 8/8/07, *Michael DeHaan* > > wrote: >> >> It's hard to tell for sure from what you've posted (lots of >> networking >> things could be afoot), but what does your >> /etc/cobbler/dnsmasq.template >> look like? How about the generated /etc/dnsmasq.conf? Those >> are a >> few things to start with. Since you didn't mention editing >> /etc/cobbler/dnsmasq.conf it might not be correct for your >> environment. >> >> There are some setup instructions including information at logging >> here: >> http://www.thekelleys.org.uk/dnsmasq/docs/setup.html -- keep in mind >> that cobbler will generate >> it's own dnsmasq.conf based on the template, and the same for >> /etc/ethers and /var/lib/cobbler/cobbler_hosts -- but that should >> give >> you some idea of how it all works. >> >> >> >> --Michael >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> >> >> I've tried a few variations on /etc/cobbler/dnsmasq.template >> Here's my latest: >> >> $cat /etc/cobbler/dnsmasq.template >> # Cobbler generated configuration file for dnsmasq >> # $date >> # >> >> # resolve.conf .. ? >> #no-poll >> #enable-dbus >> read-ethers >> addn-hosts = /var/lib/cobbler/cobbler_hosts >> >> dhcp-range=10.0.0.128 , 10.0.0.254 >> >> dhcp-option=3,$next_server >> dhcp-lease-max=1000 >> dhcp-authoritative >> dhcp-boot=pxelinux.0 >> dhcp-boot=net:normalarch,pxelinux.0 >> dhcp-boot=net:ia64,$elilo >> >> $insert_cobbler_system_definitions >> >> And the resulting >> >> $cat /etc/dnsmasq.conf >> >> # Cobbler generated configuration file for dnsmasq >> # Thu Aug 9 15:42:24 2007 >> >> # resolve.conf .. ? >> #no-poll >> #enable-dbus >> read-ethers >> addn-hosts = /var/lib/cobbler/cobbler_hosts >> >> dhcp-range= 10.0.0.128 , 10.0.0.254 >> >> dhcp-option=3,10.0.0.21 >> dhcp-lease-max=1000 >> dhcp-authoritative >> dhcp-boot=pxelinux.0 >> dhcp-boot=net:normalarch, pxelinux.0 >> dhcp-boot=net:ia64,elilo-3.6-ia64.efi >> >> and >> $ cat /etc/ethers >> 00:50:56:00:00:FF 10.0.0.66 >> >> and >> $ cat /var/lib/cobbler/cobbler_hosts >> 10.0.0.66 ctl01.sms >> >> Still pxeboot complains: No DHCP or proxyDHCP offers were received. >> >> -- >> Drew Einhorn >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > Individual networking issues are beyond my ability to debug over > email, and probably the scope of this list. > > You might want to see if you get the same problem with ISC and whether > you get DHCP offers for your VMware > guests when using your existing DHCP config. > > If any of those remove the problem, you have a dnsmasq config problem > -- otherwise, it might be something > else. I would think VMware would PXE just like anything else, but I > have no idea what all the pieces > to the puzzle are. > > Sorry I can't be of any more help. > > --Michael > > > I forgot to add, you might find some further info about debugging this on dnsmasq-discuss: http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss --Michael From mdehaan at redhat.com Thu Aug 9 16:57:17 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 09 Aug 2007 12:57:17 -0400 Subject: [et-mgmt-tools] Cobbler and Koan 0.6.0 released Message-ID: <46BB476D.3070402@redhat.com> Hi folks, 0.6.0 for both Cobbler and Koan should arrive on the mirrors for FC-6/F-7 and EPEL EL-4/EL-5 shortly. (As I understand it, Cobbler is still in EPEL-4 testing until yum can be included in EPEL-4). This release contains a lot of features that we've talked about on the list over the last several months, however 0.5.X was a testing release, so these features (while many of you have played with them) were never pushed out for mass consumption. Now they are! If you were sticking with the stable branch, I've copied and pasted the changelogs below so you can read what is new. Thanks to everyone who helped with testing, bug reports, and patches. You are all very much appreciated. --Michael *** Cobbler changes since last stable release *** * Thu Aug 09 2007 - 0.6.0 - bugfix in error path in "cobbler check" - stable release for 0.5.x * Thr Jul 26 2007 - 0.5.2 (RC) - Have cobbler check ensure services are started - Add cobbler validateks command to look for broken rendered kickstarts - Added -v/--version - Added SNIPPET::foo capability to pull /var/lib/cobbler/snippets/foo into templates (anywhere) - Import can now take an --available-as=nfs://server:/mount/point to do cobbler import without mirroring - Feature to enable "pxe_just_once" for boot loop prevention * Fri Jul 20 2007 - 0.5.1 - Added logging for cobblerd -- /var/log/cobbler/cobblerd.log - Cobblerd now ignores XMLRPC IOError - Added findks.cgi - Misc bugfixing - Added --virt-path, --virt-type * Wed Jun 24 2007 - 0.5.0 - Remove hardcode of /var/www/cobbler in cobblerd - Improve various warning warning messages - cobbler (objecttype) (objectname) now gives info about the object instead of just all objects - Added --hostname to "cobbler system add", --ip-address (or --ip) is also a better alias for the misnamed --pxe-address - Optionally use dnsmasq for DHCP (and DNS!) instead of ISC dhcpd. - Add --mac and remove requirement for --name to be an ip, mac, or hostname. - Manpage cleanup - Patch to allow pre and post triggers - Patch to allow --createrepo-flags and to cache on import, fix multiple calls to createrepo - Various modifications to allow for profile inheritance - All variables in object tree now available for use in templating, nicer blending algorithms - Optional override of --kickstart in system object ** koan changes since last stable release ** * Thu Aug 09 2007 Michael DeHaan - 0.6.0-1 - stable release - don't check for libvirtd running if doesn't exist * Thu Jul 26 2007 Michael DeHaan - 0.5.2-1 - added XMLRPC version checking with cobbler server * Fri Jul 20 2007 Michael DeHaan - 0.5.1-1 - added qemu/kvm with --virt-type - --virt-path support - added --display to dump cobbler objects without taking action - misc code & manpage cleanup * Wed Jun 27 2007 Michael DeHaan - 0.5.0-1 - Upgrades to take advantage of (and work with) Cobbler 0.5.0 From drew.einhorn at gmail.com Thu Aug 9 17:53:13 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Thu, 9 Aug 2007 11:53:13 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BB41AC.1050009@redhat.com> References: <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BB41AC.1050009@redhat.com> Message-ID: On 8/9/07, Michael DeHaan wrote: > > Individual networking issues are beyond my ability to debug over > > email, and probably the scope of this list. > > > > You might want to see if you get the same problem with ISC and whether > > you get DHCP offers for your VMware > > guests when using your existing DHCP config. It was working before I switched to dnsmaq from whatever the default was, I assume it was ISC, but I'll have to go look. > If any of those remove the problem, you have a dnsmasq config problem > > -- otherwise, it might be something > > else. I would think VMware would PXE just like anything else, but I > > have no idea what all the pieces > > to the puzzle are. > > > > Sorry I can't be of any more help. > > You've already been a great help, I forgot to add, you might find some further info about debugging this > on dnsmasq-discuss: > > http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss > > --Michael > and there you go helping even more. Thanks!!!! -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From drew.einhorn at gmail.com Fri Aug 10 15:38:44 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 09:38:44 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BB41AC.1050009@redhat.com> Message-ID: It's definitely a dnsmasq problem. If I go into /var/lib/cobbler/settings and change from dnsmasq back to isc, then It's pxe booting!!! And I move on to the next problem. After the VM boots, it complains about an Error Downloading Kickstart file. The url it gives me to tweak is the correct url, checking with a browser I see the kickstart is where it's supposed to be. Ctl-Alt-F3 to look at the log, there's the problem: 10:18:44 INFO : getting kickstart file 10:18:44 ERROR : no network device in choose network device 10:18:44:ERROR : no network drivers for doing kickstart 10:18:44:ERROR : unable to bring up network Not sure what to do next. On 8/9/07, drew einhorn wrote: > > > > On 8/9/07, Michael DeHaan wrote: > > > > Individual networking issues are beyond my ability to debug over > > > email, and probably the scope of this list. > > > > > > You might want to see if you get the same problem with ISC and whether > > > you get DHCP offers for your VMware > > > guests when using your existing DHCP config. > > > It was working before I switched to dnsmaq from whatever the default was, > I assume it was ISC, but I'll have to go look. > > > If any of those remove the problem, you have a dnsmasq config problem > > > -- otherwise, it might be something > > > else. I would think VMware would PXE just like anything else, but I > > > have no idea what all the pieces > > > to the puzzle are. > > > > > > Sorry I can't be of any more help. > > > > > > > You've already been a great help, > > I forgot to add, you might find some further info about debugging this > > on dnsmasq-discuss: > > > > http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss > > > > --Michael > > > > > and there you go helping even more. > > Thanks!!!! > > -- > Drew Einhorn -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From wright at imageworks.com Fri Aug 10 16:57:10 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 10 Aug 2007 09:57:10 -0700 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46B7340A.30901@redhat.com><46B754FE.40801@redhat.com><46BA0CCE.7050406@redhat.com><46BB4170.60909@redhat.com> <46BB41AC.1050009@redhat.com> Message-ID: <46BC98E6.30700@imageworks.com> drew einhorn wrote: > > > After the VM boots, it complains about an Error Downloading Kickstart > file. > The url it gives me to tweak is the correct url, checking with a > browser I > see the kickstart is where it's supposed to be. > > Ctl-Alt-F3 to look at the log, there's the problem: > > 10:18:44 INFO : getting kickstart file > 10:18:44 ERROR : no network device in choose network device > 10:18:44:ERROR : no network drivers for doing kickstart > 10:18:44:ERROR : unable to bring up network > the driver error sounds like an issue with the vmlinuz/initrd.img you are using. do these kernels find the divers for your NIC's when booting w/o cobbler (like booting from CD and loading your kickstart from a USB keychain for example)? if your kickstart is setup to do a network install (NFS for example) you'll know very quickly if your kernel + initrd has the correct drivers ;) -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From drew.einhorn at gmail.com Fri Aug 10 17:15:36 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 11:15:36 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BC98E6.30700@imageworks.com> References: <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BB41AC.1050009@redhat.com> <46BC98E6.30700@imageworks.com> Message-ID: On 8/10/07, Peter Wright wrote: > > drew einhorn wrote: > > > > > > After the VM boots, it complains about an Error Downloading Kickstart > > file. > > The url it gives me to tweak is the correct url, checking with a > > browser I > > see the kickstart is where it's supposed to be. > > > > Ctl-Alt-F3 to look at the log, there's the problem: > > > > 10:18:44 INFO : getting kickstart file > > 10:18:44 ERROR : no network device in choose network device > > 10:18:44:ERROR : no network drivers for doing kickstart > > 10:18:44:ERROR : unable to bring up network > > > > the driver error sounds like an issue with the vmlinuz/initrd.img you > are using. do these kernels find the divers for your NIC's when booting > w/o cobbler (like booting from CD and loading your kickstart from a USB > keychain for example)? if your kickstart is setup to do a network > install (NFS for example) you'll know very quickly if your kernel + > initrd has the correct drivers ;) Bet that's it. I have not idea where my vmlinuz/initrd.img came from. I'm not sure where my network driver came from, where it is, or exactly how it got there. It's probably a VMware proprietary driver, VMware says you have to get it from a "VMware Tools" package, but I thing it was just there when I installed to OS, but I might have done a VMware Tools install that I've forgotten about. Anyway it's probably not in the default vmlinuz/initrd.img. Gotta go figure out what I need to read to learn how to build and install a vmlinuz/initrd.img with the necessary driver. -pete > > > -- > Peter Wright > Systems Administrator > Sony Pictures Imageworks > wright at imageworks.com > www.imageworks.com > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From drew.einhorn at gmail.com Fri Aug 10 17:30:12 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 11:30:12 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BB4170.60909@redhat.com> References: <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> Message-ID: Think there's at least 2 parts to my dnsmasq problem. Got past the firstone. It was a permissions problem. Had to do a chmod o+rx /var/lib/cobbler which was owned by root, but dnsmasq was running as nobody and could not read /var/lib/cobbler/cobbler_hosts Don't know if the permission on /var/lib/cobbler was a default from the cobbler installation, or something I mucked up. Haven't gotten started on the second on. Peter Wright suggested that it looks like my vmlinuz/initrd.img don't support my virtual nic. He's probably right, I've got to figure out how to build the right vmlinz/initrd.img before I try again and we see if there's a third problem lurking behind these two. On 8/9/07, Michael DeHaan wrote: > > drew einhorn wrote: > > > > > > On 8/8/07, *Michael DeHaan* > > wrote: > > > > It's hard to tell for sure from what you've posted (lots of > networking > > things could be afoot), but what does your > > /etc/cobbler/dnsmasq.template > > look like? How about the generated /etc/dnsmasq.conf? Those are > a > > few things to start with. Since you didn't mention editing > > /etc/cobbler/dnsmasq.conf it might not be correct for your > > environment. > > > > There are some setup instructions including information at logging > > here: > > http://www.thekelleys.org.uk/dnsmasq/docs/setup.html -- keep in mind > > that cobbler will generate > > it's own dnsmasq.conf based on the template, and the same for > > /etc/ethers and /var/lib/cobbler/cobbler_hosts -- but that should > give > > you some idea of how it all works. > > > > > > > > --Michael > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > > > > I've tried a few variations on /etc/cobbler/dnsmasq.template > > Here's my latest: > > > > $cat /etc/cobbler/dnsmasq.template > > # Cobbler generated configuration file for dnsmasq > > # $date > > # > > > > # resolve.conf .. ? > > #no-poll > > #enable-dbus > > read-ethers > > addn-hosts = /var/lib/cobbler/cobbler_hosts > > > > dhcp-range=10.0.0.128 , 10.0.0.254 > > > dhcp-option=3,$next_server > > dhcp-lease-max=1000 > > dhcp-authoritative > > dhcp-boot=pxelinux.0 > > dhcp-boot=net:normalarch,pxelinux.0 > > dhcp-boot=net:ia64,$elilo > > > > $insert_cobbler_system_definitions > > > > And the resulting > > > > $cat /etc/dnsmasq.conf > > > > # Cobbler generated configuration file for dnsmasq > > # Thu Aug 9 15:42:24 2007 > > > > # resolve.conf .. ? > > #no-poll > > #enable-dbus > > read-ethers > > addn-hosts = /var/lib/cobbler/cobbler_hosts > > > > dhcp-range= 10.0.0.128 , 10.0.0.254 < > http://10.0.0.254> > > dhcp-option=3,10.0.0.21 > > dhcp-lease-max=1000 > > dhcp-authoritative > > dhcp-boot=pxelinux.0 > > dhcp-boot=net:normalarch, pxelinux.0 > > dhcp-boot=net:ia64,elilo-3.6-ia64.efi > > > > and > > $ cat /etc/ethers > > 00:50:56:00:00:FF 10.0.0.66 > > > > and > > $ cat /var/lib/cobbler/cobbler_hosts > > 10.0.0.66 ctl01.sms > > > > Still pxeboot complains: No DHCP or proxyDHCP offers were received. > > > > -- > > Drew Einhorn > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > Individual networking issues are beyond my ability to debug over email, > and probably the scope of this list. > > You might want to see if you get the same problem with ISC and whether > you get DHCP offers for your VMware > guests when using your existing DHCP config. > > If any of those remove the problem, you have a dnsmasq config problem -- > otherwise, it might be something > else. I would think VMware would PXE just like anything else, but I > have no idea what all the pieces > to the puzzle are. > > Sorry I can't be of any more help. > > --Michael > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From mdehaan at redhat.com Fri Aug 10 17:59:40 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 10 Aug 2007 13:59:40 -0400 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> Message-ID: <46BCA78C.2060604@redhat.com> drew einhorn wrote: > Think there's at least 2 parts to my dnsmasq problem. > > Got past the firstone. It was a permissions problem. > Had to do a chmod o+rx /var/lib/cobbler > > which was owned by root, but dnsmasq was running as nobody Will fix the RPM, thanks. The proper fix is probably just to chmod the cobbler_hosts file +r and leave the rest as is. > > > Don't know if the permission on /var/lib/cobbler was a default > from the cobbler installation, or something I mucked up. Looks like something I mucked up :) > > Haven't gotten started on the second on. > > Peter Wright suggested that it looks like my vmlinuz/initrd.img > don't support my virtual nic. He's probably right, I've got to > figure out how to build the right vmlinz/initrd.img before I try > again and we see if there's a third problem lurking behind these > two. cobbler import will pick the kernel/initrd from the "os/images/pxe" directory for TFTP installs. If you used "cobbler import" and your distro/profile didn't contain the word "Xen" in it, that's the one you are using. The Xen pair definitely probably won't work. If those don't work, the NIC is probably the issue -- though I would kind of expect them to be simulating something relatively standard. Could be wrong :) From drew.einhorn at gmail.com Fri Aug 10 18:33:40 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 12:33:40 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BCA78C.2060604@redhat.com> References: <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BCA78C.2060604@redhat.com> Message-ID: On 8/10/07, Michael DeHaan wrote: > > drew einhorn wrote: > > Think there's at least 2 parts to my dnsmasq problem. > > > > Got past the firstone. It was a permissions problem. > > Had to do a chmod o+rx /var/lib/cobbler > > > > which was owned by root, but dnsmasq was running as nobody > > Will fix the RPM, thanks. The proper fix is probably just to chmod the > cobbler_hosts file +r and leave the rest as is. Think /var/lib/cobbler needs o+x otherwise "nobody" will not even be able to look in the directory and see the cobbler hosts much less read it. > > > > > Don't know if the permission on /var/lib/cobbler was a default > > from the cobbler installation, or something I mucked up. > > Looks like something I mucked up :) > > > > Haven't gotten started on the second on. > > > > Peter Wright suggested that it looks like my vmlinuz/initrd.img > > don't support my virtual nic. He's probably right, I've got to > > figure out how to build the right vmlinz/initrd.img before I try > > again and we see if there's a third problem lurking behind these > > two. > > cobbler import will pick the kernel/initrd from the "os/images/pxe" > directory for > TFTP installs. If you used "cobbler import" and your distro/profile > didn't contain > the word "Xen" in it, that's the one you are using. The Xen pair > definitely > probably won't work. If those don't work, the NIC is probably the > issue -- though > I would kind of expect them to be simulating something relatively > standard. Could be wrong :) I'm trying to get things running VMware, will amost certainly have to roll my own. _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From wright at imageworks.com Fri Aug 10 18:55:08 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 10 Aug 2007 11:55:08 -0700 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BCA78C.2060604@redhat.com> References: <46AE51C9.3000706@redhat.com> <46B7340A.30901@redhat.com> <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BCA78C.2060604@redhat.com> Message-ID: <46BCB48C.7010703@imageworks.com> Michael DeHaan wrote: > > > > > > Peter Wright suggested that it looks like my vmlinuz/initrd.img > > don't support my virtual nic. He's probably right, I've got to > > figure out how to build the right vmlinz/initrd.img before I try > > again and we see if there's a third problem lurking behind these > > two. > > cobbler import will pick the kernel/initrd from the "os/images/pxe" > directory for > TFTP installs. If you used "cobbler import" and your distro/profile > didn't contain > the word "Xen" in it, that's the one you are using. The Xen pair > definitely > probably won't work. If those don't work, the NIC is probably the > issue -- though > I would kind of expect them to be simulating something relatively > standard. Could be wrong :) > I memory serves me correctly VMWare emulates either an NE100 NIC or a generic AMD NIC (can't remember the divers specifically off the top of my head though). you do not need the vm-tools package though to get the NIC working though - i believe the tools package addresses some performance problems with the virtual NIC, but i do not think it modifies the kernel module used. while googling it seems they do emulate a generic AMD NIC, what's interesting is that the OpenBSD folks have developed a specific VMware driver named "vic" which "supports the vmxnet driver protocol". -p -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From drew.einhorn at gmail.com Fri Aug 10 19:27:49 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 13:27:49 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: <46BCB48C.7010703@imageworks.com> References: <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BCA78C.2060604@redhat.com> <46BCB48C.7010703@imageworks.com> Message-ID: On 8/10/07, Peter Wright wrote: > > Michael DeHaan wrote: > > > > > > > > > > Peter Wright suggested that it looks like my vmlinuz/initrd.img > > > don't support my virtual nic. He's probably right, I've got to > > > figure out how to build the right vmlinz/initrd.img before I try > > > again and we see if there's a third problem lurking behind these > > > two. > > > > cobbler import will pick the kernel/initrd from the "os/images/pxe" > > directory for > > TFTP installs. If you used "cobbler import" and your distro/profile > > didn't contain > > the word "Xen" in it, that's the one you are using. The Xen pair > > definitely > > probably won't work. If those don't work, the NIC is probably the > > issue -- though > > I would kind of expect them to be simulating something relatively > > standard. Could be wrong :) > > > > > I memory serves me correctly VMWare emulates either an NE100 NIC or a > generic AMD NIC (can't remember the divers specifically off the top of > my head though). you do not need the vm-tools package though to get the > NIC working though - i believe the tools package addresses some > performance problems with the virtual NIC, but i do not think it > modifies the kernel module used. > > while googling it seems they do emulate a generic AMD NIC, what's > interesting is that the OpenBSD folks have developed a specific VMware > driver named "vic" which "supports the vmxnet driver protocol". VMware has 3 network virtual devices/drivers the one you are thinking of they call the vlance driver which is based on the AMD PCnet 3, there are issues with this virtual nic/driver and it is recommended only if you have bigger issues with the better drivers and can't get them to work. the vmxnet virtual device/driver is a custom virtual nic, don't believe there are any physical implementations. This one is recommended for 32-bit VMs the e1000 is a virtual Intel Pro 1000 and is recommended for 64-bit VMs. Don't think any of these are included in the stock vmlinuz/initrd.img, VMware Tools includes all this and more (virtual graphics adapter, monitoring tools, etc.) But I think some distributions take some if not all of these pieces and loads them up with all the other drivers. Did a find on the wrong box (I was confused), and found a whole of versions of vmnet.o and vmmon.o for different kernels. Wonder if vmnet.o has all three drivers bundled together? So far I finding new questions faster than answers. I'll figured it out sooner or later. -p > > -- > Peter Wright > Systems Administrator > Sony Pictures Imageworks > wright at imageworks.com > www.imageworks.com > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From berrange at redhat.com Fri Aug 10 21:03:55 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 10 Aug 2007 22:03:55 +0100 Subject: [et-mgmt-tools] improved virtinst mouse support Message-ID: <20070810210355.GE25416@redhat.com> FYI, I just pushed a change to virtinst so that it always adds a USB tablet for fullyvirtualized guests. Being a tablet, it works in absolute coordinate mode so the guest pointer will track 1-for-1 with the host pointer. Windows should automatically configure itself. For Linux you can add a input device section to the Xorg config using the evdev driver to match the USB device. Other OS, well they can just ignore if it they don't support it - the PS2 mouse is still there & operational Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From drew.einhorn at gmail.com Sat Aug 11 02:42:02 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 20:42:02 -0600 Subject: [et-mgmt-tools] dnsmasq doc suggestion Message-ID: Tell folks that when they switch from isc to dnsmasq, or vice versa, they will have to kill the currentlys running dhcpd or dnsmasq daemon (or something more graceful if they want to dig around in their init.d or whatever to find it) The new one won't start till the old one is dead. Remind them that the almost certainly want to dig around in their init.d or whatever and figure out how to make sure dhcpd or dnsmasq as appropriate starts on boot. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From drew.einhorn at gmail.com Sat Aug 11 02:43:57 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Fri, 10 Aug 2007 20:43:57 -0600 Subject: [et-mgmt-tools] cobbler system add In-Reply-To: References: <46B754FE.40801@redhat.com> <46BA0CCE.7050406@redhat.com> <46BB4170.60909@redhat.com> <46BCA78C.2060604@redhat.com> <46BCB48C.7010703@imageworks.com> Message-ID: On 8/10/07, drew einhorn wrote: > > > > On 8/10/07, Peter Wright < wright at imageworks.com> wrote: > > > > Michael DeHaan wrote: > > > > > > > > > > > > > > Peter Wright suggested that it looks like my vmlinuz/initrd.img > > > > don't support my virtual nic. He's probably right, I've got to > > > > figure out how to build the right vmlinz/initrd.img before I try > > > > again and we see if there's a third problem lurking behind these > > > > two. > > > > > > cobbler import will pick the kernel/initrd from the "os/images/pxe" > > > directory for > > > TFTP installs. If you used "cobbler import" and your distro/profile > > > didn't contain > > > the word "Xen" in it, that's the one you are using. The Xen pair > > > definitely > > > probably won't work. If those don't work, the NIC is probably the > > > issue -- though > > > I would kind of expect them to be simulating something relatively > > > standard. Could be wrong :) > > > > > > > > > I memory serves me correctly VMWare emulates either an NE100 NIC or a > > generic AMD NIC (can't remember the divers specifically off the top of > > my head though). you do not need the vm-tools package though to get the > > > > NIC working though - i believe the tools package addresses some > > performance problems with the virtual NIC, but i do not think it > > modifies the kernel module used. > > > > while googling it seems they do emulate a generic AMD NIC, what's > > interesting is that the OpenBSD folks have developed a specific VMware > > driver named "vic" which "supports the vmxnet driver protocol". > > > VMware has 3 network virtual devices/drivers > > the one you are thinking of they call the vlance driver which is based on > the > AMD PCnet 3, there are issues with this virtual nic/driver and it is > recommended > only if you have bigger issues with the better drivers and can't get them > to work. > > the vmxnet virtual device/driver is a custom virtual nic, don't believe > there > are any physical implementations. This one is recommended for 32-bit VMs > > the e1000 is a virtual Intel Pro 1000 and is recommended for 64-bit VMs. > > Don't think any of these are included in the stock vmlinuz/initrd.img, > > VMware Tools includes all this and more (virtual graphics adapter, > monitoring tools, etc.) > > But I think some distributions take some if not all of these pieces and > loads them up with all the other drivers. > > Did a find on the wrong box (I was confused), and found a whole of > versions > of vmnet.o and vmmon.o for different kernels. > > Wonder if vmnet.o has all three drivers bundled together? > > So far I finding new questions faster than answers. > > I'll figured it out sooner or later. > I was wrong about one thing the low end vlance driver is included in the default vmlinuz/initrd.img And if I build a vm using the vlance virtualdevice things work!!! And I do need to build a custom 32-bit vmlinuz/initrd.img with the vmxnet driver and a 64-bit vmlinuz/initrd.img with the e1000 driver. Hmm, Maybe it's a stock e1000 driver, and maybe that one is already supported. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From fj0588di at aa.jp.fujitsu.com Mon Aug 13 09:41:33 2007 From: fj0588di at aa.jp.fujitsu.com (S.Sakamoto) Date: Mon, 13 Aug 2007 18:41:33 +0900 Subject: [et-mgmt-tools] [PATCH] check if the specified diskusingvirt-clone is already used in another guests. In-Reply-To: <200708031705.GAI26568.JGE9K960@aa.jp.fujitsu.com> References: <200707300954.AFB05251.9EKJ60G9@aa.jp.fujitsu.com> <200708031705.GAI26568.JGE9K960@aa.jp.fujitsu.com> Message-ID: <200708131841.FCD13070.JEGK6909@aa.jp.fujitsu.com> Hi I want to do a duplication check similar to virt-install in virt-clone. Please apply this patch. Thanks, Shigeki Sakamoto. > Hi > > Would you give me a comment on this patch? > If not, please apply it. > > > Thanks, > Shigeki Sakamoto. > > > Hi > > > > Current virt-clone does not confirm if the specified disk is already > > used in another guests or not. So, I create the patch to check if the > > specified disk using virt-clone is already used in another guests. > > > > > > Signed-off-by: Shigeki Sakamoto > > > > Thanks, > > Shigeki Sakamoto. > > > > ---------------------------------------------------------------------- > > diff -r 645217bda13b virtinst/CloneManager.py > > --- a/virtinst/CloneManager.py Sat Jul 21 13:03:07 2007 -0400 > > +++ b/virtinst/CloneManager.py Fri Jul 27 17:38:47 2007 +0900 > > @@ -29,6 +29,7 @@ import commands > > import commands > > import libvirt > > import Guest > > +import cli > > from virtinst import _virtinst as _ > > > > # > > @@ -106,6 +107,13 @@ class CloneDesign(object): > > def set_clone_devices(self, devices): > > if len(devices) == 0: > > raise ValueError, _("New file to use for disk image is required") > > + cdev = [] > > + cdev_size = [] > > + cdev_type = [] > > + cdev.append(devices) > > + cdev_size,\ > > + cdev_type = self._get_clone_devices_info(cdev) > > + devices = self._check_file(self._hyper_conn, devices, cdev_size) > > self._clone_devices.append(devices) > > def get_clone_devices(self): > > return self._clone_devices > > @@ -320,6 +328,45 @@ class CloneDesign(object): > > except libvirt.libvirtError, e: > > pass > > return check > > + > > + # > > + # check used file func > > + # ret : Use File Path > > + # > > + def _check_file(self, conn, disk, size): > > + retryFlg = False > > + while 1: > > + if disk == None: > > + msg = _("What would you like to use as the disk (path)?") > > + disk = cli.prompt_for_input(msg, disk) > > + > > + try: > > + d = Guest.VirtualDisk(disk, size) > > + if d.is_conflict_disk(conn) is True: > > + while 1: > > + retryFlg = False > > + warnmsg = _("Disk %s is already in use by another guest!\n") % disk > > + res = cli.prompt_for_input(warnmsg + _("Do you really want to use the disk (yes or no)? ")) > > + try: > > + if cli.yes_or_no(res) is True: > > + break > > + else: > > + retryFlg = True > > + break > > + except ValueError, e: > > + print _("ERROR: "), e > > + continue > > + if retryFlg is True: > > + disk = None > > + continue > > + except ValueError, e: > > + print _("ERROR: "), e > > + disk = None > > + continue > > + > > + break > > + > > + return disk > > > > # > > # check used mac func > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Mon Aug 13 15:26:29 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 13 Aug 2007 11:26:29 -0400 Subject: [et-mgmt-tools] Cobbler article in Red Hat Magazine Message-ID: <46C07825.7070503@redhat.com> Shouldn't be anything new for folks on this list, but I figured I would pass it along anyway :) http://www.redhatmagazine.com/2007/08/10/cobbler-how-to-set-up-a-network-boot-server-in-10-minutes/ --Michael From mdehaan at redhat.com Mon Aug 13 15:43:21 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 13 Aug 2007 11:43:21 -0400 Subject: [et-mgmt-tools] Status Message-ID: <46C07C19.8040809@redhat.com> Last week: -- lots of KVM hacking/testing with virt-factory, provisioning module changes -- distro/repo mirroring workflow overhaul -- brought Virt-Factory up to F-7 (with KVM) for default test config -- released cobbler 0.6.0 -- meeting with RHX + Luke (puppet) + skvidal, re: RHX + config management + virt -- Cobbler Red Hat magazine article was published (yipee!) This week -- meeting with jmh re: dreamworks -- meet with Don Fischer re: EC2 -- Cobbler presentation for virtual-Fudcon (IRC) -- another round of RHX discussions re: virt-factory, profiles, how we can collaborate -- Uli's parallel programming + hw talks in RDU (Thursday/Friday) -- squeeze in time to do actual work :) From mdehaan at redhat.com Mon Aug 13 15:46:18 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 13 Aug 2007 11:46:18 -0400 Subject: [et-mgmt-tools] Status In-Reply-To: <46C07C19.8040809@redhat.com> References: <46C07C19.8040809@redhat.com> Message-ID: <46C07CCA.1060407@redhat.com> Michael DeHaan wrote: ... ugh, wrong list. we have too many lists starting with the same thing. please ignore. From mdehaan at redhat.com Mon Aug 13 22:47:33 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 13 Aug 2007 18:47:33 -0400 Subject: [et-mgmt-tools] Networking modes from koan ("user" vs "bridged") Message-ID: <46C0DF85.1020002@redhat.com> Hi folks, Apparently I had koan set up to install Xen guests in "user" networking mode instead of "bridged" mode, which means the IP's the virtual guests would be not be accessible from outside the virtual host. This does not seem to be what most folks would want, so I've changed it to "bridged". I've fixed this upstream (see git). Basically if you want to change this yourself, just change "user" to "bridged" in xencreate.py (/usr/lib/python2.4/site-packages/koan/xencreate.py). Or, here's the patch to do the same: http://git.fedoraproject.org/?p=hosted/koan;a=commitdiff;h=3cf1bcd2a967fc8dfb84cff7b5e55234a96ed445 That same one-line change is not working for me with KVM though I'm going to investigate and see what it takes to make KVM work this way as well -- assuming it's doable -- which it most likely is. --Michael From berrange at redhat.com Mon Aug 13 23:17:52 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 14 Aug 2007 00:17:52 +0100 Subject: [et-mgmt-tools] Networking modes from koan ("user" vs "bridged") In-Reply-To: <46C0DF85.1020002@redhat.com> References: <46C0DF85.1020002@redhat.com> Message-ID: <20070813231752.GA13194@redhat.com> On Mon, Aug 13, 2007 at 06:47:33PM -0400, Michael DeHaan wrote: > Hi folks, > > Apparently I had koan set up to install Xen guests in "user" networking > mode instead of "bridged" mode, which means the IP's the virtual guests > would be not be accessible > from outside the virtual host. This does not seem to be what most > folks would want, so I've changed it to "bridged". > > I've fixed this upstream (see git). Basically if you want to change > this yourself, just change "user" to "bridged" in xencreate.py > (/usr/lib/python2.4/site-packages/koan/xencreate.py). > > Or, here's the patch to do the same: > > http://git.fedoraproject.org/?p=hosted/koan;a=commitdiff;h=3cf1bcd2a967fc8dfb84cff7b5e55234a96ed445 > > That same one-line change is not working for me with KVM though I'm > going to investigate and see what it takes to make KVM work this way as > well -- assuming it's doable -- > which it most likely is. It should work with KVM - you just need to have something to bridge to and you need to specify what it is. There's basically 3 networking modes - type="bridge" - bridge the virtual interface to the physical LAN. Needs to have host setup so that one or more of the physical ethernet devices are members of bridges. Xen's network script does this automatically for eth0/xenbr0. I don't recommend using that though. Instead follow example number #2 here http://watzmann.net/blog/index.php/2007/04/27/networking_with_kvm_and_libvirt When passing type='bridge' you also need to specufy what device you want to bridge to - ie the name of the bridge shown by brctl show. If following the link above, this would be 'eth0' (your physical device is peth0). - type="network" - connect the virtual interface to a virtual network. This virtual network consists of an isolated bridge device with no physical devices attached. Connectivity to the outside world is via NAT. This is xample #1 in the URL above When passing type='user' you also need to specify what network you wanto to attach to. Libvirt provides one called 'default' out of the box (this is where the virbr0 comes from). - type="user" - this is only relevant for QEMU guests - it is not supported for Xen. It is a special NAT style connetion that doesn't rely on the host networking stack at all. It is basically only intended for people who don't have root who still want to use QEMU. If it works at all with Xen that is pure luck & certainly not intended - it is a bug that libvirt does not reject this config in the Xen case The type='bridge' is the only one which really makes much sense for koan I think, perhaps type='network' in some real niche cases & certainly not type='user'. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From jim at rossberry.com Mon Aug 13 22:29:32 2007 From: jim at rossberry.com (Jim Wildman) Date: Mon, 13 Aug 2007 18:29:32 -0400 (EDT) Subject: [et-mgmt-tools] Status In-Reply-To: <46C07CCA.1060407@redhat.com> References: <46C07C19.8040809@redhat.com> <46C07CCA.1060407@redhat.com> Message-ID: On Mon, 13 Aug 2007, Michael DeHaan wrote: > Michael DeHaan wrote: > > ... > > ugh, wrong list. we have too many lists starting with the same thing. > > please ignore. But it was interesting!! ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com "Society in every state is a blessing, but Government, even in its best state, is a necessary evil; in its worst state, an intolerable one." Thomas Paine From hbrock at redhat.com Tue Aug 14 16:43:59 2007 From: hbrock at redhat.com (Hugh Brock) Date: Tue, 14 Aug 2007 12:43:59 -0400 Subject: [et-mgmt-tools] multiple connections in virt-manager first draft now on tip Message-ID: <46C1DBCF.8090100@redhat.com> Hi all. I have just committed the first cut at a UI for handling multiple connections from a single manager window in virt-manager. Opening a new connection with File->Open Connection... will now display that connection and its guests (if any) in a tree layout in the manager window. The "new" button is gone, in favor of a right-click option for creating guests for a particular connection. If you click the expand/collapse arrow to collapse a connection, virt-manager will avoid polling that connection for resource updates until you expand the arrow again (handy for avoiding excessive network traffic in the remote connection case). Things that still need to be dealt with: 1. I have temporarily disabled restoring guests from the UI. It's necessary to specify a connection when restoring a guest, and this brings up issues around how to choose that connection and the whole guest checkpointing question. 2. There should really be a little "new" or "create" button on each connection line along with the right-click menu. 3. We should store connection info in gconf rather than having them vanish on disconnection. This way when you start virt-manager all your connections would come up inactive, and actually connect only when you activate them. 4. Creating a new guest still works only on local connections. Guest creation for remote connections depends on resource discovery which we still haven't worked out yet. Please give the new UI a spin and let me know if you find anything broken. In particular, I haven't tested it with remote connections, although (aside from creating a guest) there is no reason it shouldn't work. Enjoy, --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From tomg at mcclatchyinteractive.com Tue Aug 14 19:18:43 2007 From: tomg at mcclatchyinteractive.com (Tom Georgoulias) Date: Tue, 14 Aug 2007 15:18:43 -0400 Subject: [et-mgmt-tools] changing webdir and ks config file locations in 0.6.0 In-Reply-To: <20070813231752.GA13194@redhat.com> References: <46C0DF85.1020002@redhat.com> <20070813231752.GA13194@redhat.com> Message-ID: <46C20013.6010101@mcclatchyinteractive.com> Hi, I'm new to this list and just started working to replace our current RHEL3 provisioning system with cobbler on RHEL5. So far so good. :) I need a sanity check, though: I wanted to change the default file locations for the webdir and kickstart config files ( currently /var/www/cobbler and /etc/cobbler/*) during the RPM install and found myself making so many changes to the source code files that I ended up requiring a patch for the RPM and lots of edits to cobbler.spec. Getting the webdir filepath changed required me to edit these files: setup.py config/cobbler.conf scripts/findks.cgi scripts/nopxe.cgi cobbler/settings.py I couldn't find a way to move just the kickstart files elsewhere without impacting the entire cobbler etcpath. Did I over complicate this or is this what is required at this time? Thanks, Tom -- Tom Georgoulias Sr. Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com From mdehaan at redhat.com Tue Aug 14 19:34:20 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 14 Aug 2007 15:34:20 -0400 Subject: [et-mgmt-tools] changing webdir and ks config file locations in 0.6.0 In-Reply-To: <46C20013.6010101@mcclatchyinteractive.com> References: <46C0DF85.1020002@redhat.com> <20070813231752.GA13194@redhat.com> <46C20013.6010101@mcclatchyinteractive.com> Message-ID: <46C203BC.205@redhat.com> Tom Georgoulias wrote: > Hi, > > I'm new to this list and just started working to replace our current > RHEL3 provisioning system with cobbler on RHEL5. So far so good. :) > > I need a sanity check, though: I wanted to change the default file > locations for the webdir and kickstart config files ( currently > /var/www/cobbler and /etc/cobbler/*) during the RPM install and found > myself making so many changes to the source code files that I ended up > requiring a patch for the RPM and lots of edits to cobbler.spec. > Getting the webdir filepath changed required me to edit these files: > > setup.py > config/cobbler.conf > scripts/findks.cgi > scripts/nopxe.cgi > cobbler/settings.py There are some places though (koan, templates, etc) that assume some relative http structure exists -- so I'd caution against moving the web dir. A better question would be, why do you want to move the webdir? If it is to conserve space, you can use "--available-as" on the cobbler import to specify that the URLs you import are available externally and do not need to be mirrored. Then the content to be kept on the cobbler server would be fairly minimal. It's true that the yum mirroring support doesn't really know about this yet, which brings me to option 2... Another option would be to make /var/www/cobbler (or even /var/www) a symlink to a directory on another partition/volume -- or an NFS mount point. Be careful though, NFS + symlinks + Apache + SELinux can require a bit of hoop jumping. (You can use mount with a security context instead). The templates files in /etc can probably stay there. If you want to add some more, you can path them anywhere you want and don't have to use those. Those templates are just starter ideas :) Syntax for changing what template a profile uses: cobbler profile edit --name=foo --kickstart=/opt/special/mykickstarts/x.ks # etc Hopefully that makes sense, if not, maybe you can elaborate a bit on why you wanted to move things around? --Michael From tomg at mcclatchyinteractive.com Tue Aug 14 20:24:13 2007 From: tomg at mcclatchyinteractive.com (Tom Georgoulias) Date: Tue, 14 Aug 2007 16:24:13 -0400 Subject: [et-mgmt-tools] changing webdir and ks config file locations in 0.6.0 In-Reply-To: <46C203BC.205@redhat.com> References: <46C0DF85.1020002@redhat.com> <20070813231752.GA13194@redhat.com> <46C20013.6010101@mcclatchyinteractive.com> <46C203BC.205@redhat.com> Message-ID: <46C20F6D.3030205@mcclatchyinteractive.com> Michael DeHaan wrote: > There are some places though (koan, templates, etc) that assume some > relative http structure exists -- so I'd caution against moving the web > dir. A better question would be, why do you want to move the webdir? > > If it is to conserve space, you can use "--available-as" on the cobbler > import to specify that the URLs you import are available externally and > do not need to be mirrored. > Another option would be to make /var/www/cobbler (or even /var/www) a > symlink to a directory on another partition/volume It was disk space related--I have access to much more disk space if I use a partition other than /var, so creating the /var/www/cobbler symlink is a good option for me. > Syntax for changing what template a profile uses: > > cobbler profile edit --name=foo > --kickstart=/opt/special/mykickstarts/x.ks # etc Excellent. I wanted to continue to make use of our existing kickstart config file generator script and provide cobbler with access those ks.cfg files, so that takes care of that. I guess I mistakenly thought that cobbler was going to look for them in /etc/cobbler instead of whereever I specified them. > Hopefully that makes sense, if not, maybe you can elaborate a bit on why > you wanted to move things around? It makes perfect sense. I've gone back to the standard RPM and implemented your suggestions. Tom -- Tom Georgoulias Sr. Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com From mdehaan at redhat.com Tue Aug 14 21:03:11 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 14 Aug 2007 17:03:11 -0400 Subject: [et-mgmt-tools] Using Cobbler to erase systems Message-ID: <46C2188F.5070404@redhat.com> Harry Hoffman wrote up a good Wiki article on using Cobbler with DBAN here: https://hosted.fedoraproject.org/projects/cobbler/wiki/SystemRetirement The same kind of tricks can also be played with "deploying" memtest ... though I believe you may have to feed cobbler a bogus (empty) initrd file because memtest doesn't have or use one. (You can also just insert memtest by hand into the top of the PXE menu templates). --Michael From mdehaan at redhat.com Tue Aug 14 21:37:59 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 14 Aug 2007 17:37:59 -0400 Subject: [et-mgmt-tools] Networking modes from koan ("user" vs "bridged") In-Reply-To: <20070813231752.GA13194@redhat.com> References: <46C0DF85.1020002@redhat.com> <20070813231752.GA13194@redhat.com> Message-ID: <46C220B7.8010502@redhat.com> Daniel P. Berrange wrote: > On Mon, Aug 13, 2007 at 06:47:33PM -0400, Michael DeHaan wrote: > >> Hi folks, >> >> Apparently I had koan set up to install Xen guests in "user" networking >> mode instead of "bridged" mode, which means the IP's the virtual guests >> would be not be accessible >> from outside the virtual host. This does not seem to be what most >> folks would want, so I've changed it to "bridged". >> >> I've fixed this upstream (see git). Basically if you want to change >> this yourself, just change "user" to "bridged" in xencreate.py >> (/usr/lib/python2.4/site-packages/koan/xencreate.py). >> >> Or, here's the patch to do the same: >> >> http://git.fedoraproject.org/?p=hosted/koan;a=commitdiff;h=3cf1bcd2a967fc8dfb84cff7b5e55234a96ed445 >> >> That same one-line change is not working for me with KVM though I'm >> going to investigate and see what it takes to make KVM work this way as >> well -- assuming it's doable -- >> which it most likely is. >> > > It should work with KVM - you just need to have something to bridge to > and you need to specify what it is. There's basically 3 networking modes > > - type="bridge" - bridge the virtual interface to the physical LAN. Needs > to have host setup so that one or more of the physical ethernet devices > are members of bridges. Xen's network script does this automatically for > eth0/xenbr0. I don't recommend using that though. Instead follow example > number #2 here > > http://watzmann.net/blog/index.php/2007/04/27/networking_with_kvm_and_libvirt > > > When passing type='bridge' you also need to specufy what device you want to > bridge to - ie the name of the bridge shown by brctl show. If following the > link above, this would be 'eth0' (your physical device is peth0). > > - type="network" - connect the virtual interface to a virtual network. This > virtual network consists of an isolated bridge device with no physical > devices attached. Connectivity to the outside world is via NAT. This is > xample #1 in the URL above > > When passing type='user' you also need to specify what network you wanto > to attach to. Libvirt provides one called 'default' out of the box (this > is where the virbr0 comes from). > > - type="user" - this is only relevant for QEMU guests - it is not supported > for Xen. It is a special NAT style connetion that doesn't rely on the > host networking stack at all. It is basically only intended for people > who don't have root who still want to use QEMU. If it works at all with > Xen that is pure luck & certainly not intended - it is a bug that libvirt > does not reject this config in the Xen case > > The type='bridge' is the only one which really makes much sense for koan > I think, perhaps type='network' in some real niche cases & certainly not > type='user'. > > > Regards, > Dan. > Ok, with Dan's help (thanks!), I've hammered out a solution to this... You can read about how it works on the Wiki: https://hosted.fedoraproject.org/projects/cobbler/wiki/VirtNetworkingSetupForUseWithKoan Basically this is going to be in koan 0.6.1 and if you want to try this out now, pull from the latest git source, and build a version of koan. As this is rather important/useful, I'll probably push this out in about a week or so rather than including this in a more evolved "testing" release. However, you don't have to wait -- testing is still very much appreciated :) --Michael From mazurekd at cpsc.ucalgary.ca Tue Aug 14 23:20:31 2007 From: mazurekd at cpsc.ucalgary.ca (Debbie Mazurek) Date: Tue, 14 Aug 2007 17:20:31 -0600 Subject: [et-mgmt-tools] Cobbler version Message-ID: <1187133631.18482.181.camel@caliente> Hi, I am only seeing cobbler-0.4.8-1 on the epel-testing mirrors. Should 0.6 be there, or is it only available as a src.rpm? Or, is there another yum repo out there with it? thanks, dm -- d e b b i e m a z u r e k mazurekd at cpsc.ucalgary.ca system admin . department of computer science . university of calgary From 1.cobbler at msquared.id.au Wed Aug 15 01:06:25 2007 From: 1.cobbler at msquared.id.au (Msquared) Date: Wed, 15 Aug 2007 09:06:25 +0800 Subject: [et-mgmt-tools] changing webdir and ks config file locations in 0.6.0 In-Reply-To: <46C203BC.205@redhat.com> References: <46C0DF85.1020002@redhat.com> <20070813231752.GA13194@redhat.com> <46C20013.6010101@mcclatchyinteractive.com> <46C203BC.205@redhat.com> Message-ID: <20070815010625.GU19300@sliderule.msquared.com.au> On Tue, Aug 14, 2007 at 03:34:20PM -0400, Michael DeHaan wrote: > There are some places though (koan, templates, etc) that assume some > relative http structure exists -- so I'd caution against moving the web > dir. A better question would be, why do you want to move the webdir? Different from Tom, I may want to move it to a subfolder, perhaps. If I plan to use a webserver for virtual hosting, I will create a subfolder within /var/www for each virtual host. eg: /var/www/www.mydomain.com/ /var/www/www.yourdomain.com/ /var/www/personal.mydomain.com/ I think this would be less work and cause less issues than moving the entire /var/www somewhere else, but are there still any 'gotchas' I should be aware of? Regards, Msquared... From sakaia at jp.fujitsu.com Wed Aug 15 02:42:48 2007 From: sakaia at jp.fujitsu.com (Atsushi SAKAI) Date: Wed, 15 Aug 2007 11:42:48 +0900 Subject: [et-mgmt-tools] [patch] fix typo in virtmanager Message-ID: <200708150242.l7F2gtOS022324@fjmscan501.ms.jp.fujitsu.com> Since I detect typo in libvirt and virtinstall, To make sure, I try to apply ispell to virtmanager to make sure. Signed-off-by: Atsushi SAKAI Thanks Atsushi SAKAI -------------- next part -------------- A non-text attachment was scrubbed... Name: fix_virtmgr_typo.patch Type: application/octet-stream Size: 4808 bytes Desc: not available URL: From sakaia at jp.fujitsu.com Wed Aug 15 02:42:51 2007 From: sakaia at jp.fujitsu.com (Atsushi SAKAI) Date: Wed, 15 Aug 2007 11:42:51 +0900 Subject: [et-mgmt-tools] [patch] fix typo in virtinstall Message-ID: <200708150243.l7F2h1a5019450@fjmscan502.ms.jp.fujitsu.com> Since I detect typo in libvirt, I try to apply ispell to virtinstall to make sure. Anyway Changelog seems many typos, but I am not touch it. Signed-off-by: Atsushi SAKAI Thanks Atsushi SAKAI -------------- next part -------------- A non-text attachment was scrubbed... Name: fix_virtinst_typo.patch Type: application/octet-stream Size: 9886 bytes Desc: not available URL: From berrange at redhat.com Wed Aug 15 03:46:19 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Wed, 15 Aug 2007 04:46:19 +0100 Subject: [et-mgmt-tools] multiple connections in virt-manager first draft now on tip In-Reply-To: <46C1DBCF.8090100@redhat.com> References: <46C1DBCF.8090100@redhat.com> Message-ID: <20070815034619.GA19360@redhat.com> On Tue, Aug 14, 2007 at 12:43:59PM -0400, Hugh Brock wrote: > Hi all. > > I have just committed the first cut at a UI for handling multiple > connections from a single manager window in virt-manager. Opening a new > connection with File->Open Connection... will now display that > connection and its guests (if any) in a tree layout in the manager > window. The "new" button is gone, in favor of a right-click option for > creating guests for a particular connection. If you click the > expand/collapse arrow to collapse a connection, virt-manager will avoid > polling that connection for resource updates until you expand the arrow > again (handy for avoiding excessive network traffic in the remote > connection case). > > Things that still need to be dealt with: > > 1. I have temporarily disabled restoring guests from the UI. It's > necessary to specify a connection when restoring a guest, and this > brings up issues around how to choose that connection and the whole > guest checkpointing question. Ultimately I think the save/restore stuff needs to be re-thought, perhaps with extra APIs at the libvirt layer. The current stuff is what I term 'unmanaged save/restore' - ie once you save it the image, libvirt (or the underlying tools like XenD) have no tracking of the saved VM. As of Xen 3.1, or QEMU 0.9 there is a concept of 'managed save/restore' (and even checkpoints) where the save images are permanently associated with the VM. So for an inactive VM, you can list available saved images, and pick one to restore to. This would work quite nicely in the virt-manager UI, since we'd just see the inactive VM and when starting, one could let the user start it from scratch, or pick a saved image to start from. > 2. There should really be a little "new" or "create" button on each > connection line along with the right-click menu. Definitely add it to the right-click menu. We should try having it inline too just to see whether it works as a UI concept. I'm in half a mind to say that we should kill the row of buttons along the bottom of the main window completely. Double-clicking lets you easily display the console window, and from there you can view details. It might even make sense to put the console & details into one window - just as separate tabs. > 3. We should store connection info in gconf rather than having them > vanish on disconnection. This way when you start virt-manager all your > connections would come up inactive, and actually connect only when you > activate them. Yes, that'd work nicely. Future ideas also include having the libvirt daemon advertise itself using Ahavi, so virt-manager could browse for virt capable hosts on the LAN. > 4. Creating a new guest still works only on local connections. Guest > creation for remote connections depends on resource discovery which we > still haven't worked out yet. Tested it very quickly and it seemed to work in a couple of cases - good stuff ! Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From mdehaan at redhat.com Wed Aug 15 10:59:32 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 15 Aug 2007 06:59:32 -0400 Subject: [et-mgmt-tools] changing webdir and ks config file locations in 0.6.0 In-Reply-To: <20070815010625.GU19300@sliderule.msquared.com.au> References: <46C0DF85.1020002@redhat.com> <20070813231752.GA13194@redhat.com> <46C20013.6010101@mcclatchyinteractive.com> <46C203BC.205@redhat.com> <20070815010625.GU19300@sliderule.msquared.com.au> Message-ID: <46C2DC94.1000804@redhat.com> Msquared wrote: > On Tue, Aug 14, 2007 at 03:34:20PM -0400, Michael DeHaan wrote: > > >> There are some places though (koan, templates, etc) that assume some >> relative http structure exists -- so I'd caution against moving the web >> dir. A better question would be, why do you want to move the webdir? >> > > Different from Tom, I may want to move it to a subfolder, perhaps. If I > plan to use a webserver for virtual hosting, I will create a subfolder > within /var/www for each virtual host. eg: > > /var/www/www.mydomain.com/ > /var/www/www.yourdomain.com/ > /var/www/personal.mydomain.com/ > > I think this would be less work and cause less issues than moving the > entire /var/www somewhere else, but are there still any 'gotchas' I should > be aware of? > > Regards, Msquared... > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > That shouldn't be too difficult. If you set "server" to (for example) bootserver.mydomain.com in /var/lib/cobbler/settings and then "webdir" to /var/www/bootserver.mydomain.com/cobbler things should work out fine. That being said, it's not something I test, so if you encounter any weirdness, let me know, and I'll fix it. You will definitely have to modify /etc/httpd.d/cobbler.conf to modify cobbler's Apache configuration -- though it's marked "config/noreplace" in the RPM so it shouldn't change in an upgrade once you get it the way you want. The main thing is ensuring that the URL structure cobbler expects ("http://$server/cobbler/") remains the same from http:// -- there shouldn't be any hardcodes of "/var/www/cobbler" anywhere in the Python code anymore -- and if you ever run across one, it's a bug. Since the RPM will still want to own directory entries in /var/www/cobbler make sure you copy the directory structure over to the new location. You can see that information by looking for "/var/www/cobbler" in the RPM spec file. Or, you can do what Tom did and build your own patched RPM -- in your case, I don't think you shouldn't have to do that -- but if you want to, you can. Those directories are: %dir /var/log/cobbler %dir /var/log/cobbler/kicklog %dir /var/www/cobbler/ %dir /var/www/cobbler/localmirror %dir /var/www/cobbler/kickstarts %dir /var/www/cobbler/kickstarts_sys %dir /var/www/cobbler/repo_mirror %dir /var/www/cobbler/ks_mirror %dir /var/www/cobbler/ks_mirror/config %dir /var/www/cobbler/images %dir /var/www/cobbler/distros %dir /var/www/cobbler/profiles %dir /var/www/cobbler/systems %dir /var/www/cobbler/links --Michael From mdehaan at redhat.com Wed Aug 15 11:24:51 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 15 Aug 2007 07:24:51 -0400 Subject: [et-mgmt-tools] Cobbler version In-Reply-To: <1187133631.18482.181.camel@caliente> References: <1187133631.18482.181.camel@caliente> Message-ID: <46C2E283.3060903@redhat.com> Debbie Mazurek wrote: > Hi, > > I am only seeing cobbler-0.4.8-1 on the epel-testing mirrors. > Should 0.6 be there, or is it only available as a src.rpm? Or, is there > another yum repo out there with it? > > thanks, > > dm > I asked about this last night, and EPEL hasn't done a push yet. Once that happens, 0.6.0 will be in both EL 4 and EL 5. Rebuilding the source RPM should be fine, though you'll likely have to find a few dependencies -- which you can get from EPEL. Sidenote: I believe EPEL-4 is still waiting for the inclusion of yum and yum-utils in EPEL4 (ETA: soon), so cobbler won't move into EL 4 "non-testing" until yum-utils is part of EL4. Cobbler requires yum-utils as a dependency for the repo management features. This isn't an issue for EL5. From John.Sanabria at ece.uprm.edu Wed Aug 15 13:58:38 2007 From: John.Sanabria at ece.uprm.edu (John Sanabria) Date: Wed, 15 Aug 2007 09:58:38 -0400 (AST) Subject: [et-mgmt-tools] How to create XEN virtual machine from iso file Message-ID: <47576.136.145.116.114.1187186318.squirrel@ece.uprm.edu> Hi, Here at UPRM, we are working in to build a Linux distro to create a grid computational nodes. Now, we have a preliminar distro image. I wanna know how to create a xen virtual machine using cobbler from an iso image? Thanks a lot, regards. http://ece.uprm.edu/~s047267 http://del.icio.us/josanabr http://blog-grid.blogspot.com From mdehaan at redhat.com Wed Aug 15 14:26:39 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 15 Aug 2007 10:26:39 -0400 Subject: [et-mgmt-tools] How to create XEN virtual machine from iso file In-Reply-To: <47576.136.145.116.114.1187186318.squirrel@ece.uprm.edu> References: <47576.136.145.116.114.1187186318.squirrel@ece.uprm.edu> Message-ID: <46C30D1F.7090202@redhat.com> John Sanabria wrote: > Hi, > > Here at UPRM, we are working in to build a Linux distro to create a grid > computational nodes. > > Now, we have a preliminar distro image. I wanna know how to create a xen > virtual machine using cobbler from an iso image? > > If the distro follows the regular ISO structure that Fedora, RHEL, and derivatives follow, the following will work: cobbler import --name=mydistro --path=/mnt/mounted_iso_location If not, you can use "cobbler distro add" and just copy your install tree from the DVD to a location in your http tree. The path "local_mirror" is reserved for this purpose. /var/www/cobbler/local_mirror/mydistro The distro add command is documented in the manpage. You would then update your kickstarts to reference the install tree at the new location. This could be done by changing the URL line in the kickstart or, if the template contains the "$tree" variable like the cobbler templates, you can just feed it in by changing the template parameter that describes where the install tree is... cobbler distro edit --name=mydistro-i386 --ksmeta="tree=http://server.example.org/cobbler/local_mirror/mydistro/.../tree" OR just cobbler profile edit --name=mydistro-i386 --kickstart=/opt/mystuff/newkickstart.cfg FYI -- I believe you were one of the people who were using the "enchant" function for remote installation of nodes. Since this is just a wrapper around ssh (ssh target.example.com koan --virt --profile=foo --server=bootserver.example.com) I will probably be removing this function in the next couple of releases. You may want to update your code to just invoke koan over SSH and if neccessary deploy koan with scp or to pull it in with yum or up2date. Looking over the "action_enchant.py" module should give you a good idea what is going on. > Thanks a lot, regards. > > http://ece.uprm.edu/~s047267 > http://del.icio.us/josanabr > http://blog-grid.blogspot.com > > > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From rigg0022 at umn.edu Wed Aug 15 15:56:00 2007 From: rigg0022 at umn.edu (Ben Riggs) Date: Wed, 15 Aug 2007 10:56:00 -0500 Subject: [et-mgmt-tools] [PATCH] Changed collection.find() to take a dictionary argument instead of name, allowing searches by ip_address, mac_address, etc. Message-ID: <11871933602647-git-send-email-rigg0022@umn.edu> From: root diff --git a/cobbler/action_enchant.py b/cobbler/action_enchant.py index 3bf80f8..4a54ebe 100644 --- a/cobbler/action_enchant.py +++ b/cobbler/action_enchant.py @@ -40,9 +40,9 @@ class Enchant: raise CX(_("enchant failed. no address specified")) if system is None and profile is None: raise CX(_("enchant failed. no profile specified")) - if system is not None and self.config.systems().find(system) is None: + if system is not None and self.config.systems().find(name=system) is None: raise CX(_("enchant failed. system not found")) - if profile is not None and self.config.profiles().find(profile) is None: + if profile is not None and self.config.profiles().find(name=profile) is None: raise CX(_("enchant failed. profile name not found")) diff --git a/cobbler/action_import.py b/cobbler/action_import.py index 5789ae0..c74bc81 100644 --- a/cobbler/action_import.py +++ b/cobbler/action_import.py @@ -155,7 +155,7 @@ class Importer: """ for profile in self.profiles: - distro = self.distros.find(profile.distro) + distro = self.distros.find(name=profile.distro) if distro is None or not (distro in self.distros_added): print _("- skipping distro %s since it wasn't imported this time") % profile.distro continue @@ -429,7 +429,7 @@ class Importer: pxe_arch = self.get_pxe_arch(dirname) name = self.get_proposed_name(dirname, pxe_arch) - existing_distro = self.distros.find(name) + existing_distro = self.distros.find(name=name) if existing_distro is not None: print _("- modifying existing distro: %s") % name @@ -446,7 +446,7 @@ class Importer: self.distros.add(distro) self.distros_added.append(distro) - existing_profile = self.profiles.find(name) + existing_profile = self.profiles.find(name=name) if existing_profile is None: print _("- creating new profile: %s") % name diff --git a/cobbler/action_litesync.py b/cobbler/action_litesync.py index 21e6c62..2560d8d 100644 --- a/cobbler/action_litesync.py +++ b/cobbler/action_litesync.py @@ -53,7 +53,7 @@ class BootLiteSync: def add_single_distro(self, name): # get the distro record - distro = self.distros.find(name) + distro = self.distros.find(name=name) if distro is None: raise CX(_("error in distro lookup: %s") % name) # generate YAML file in distros/$name in webdir @@ -73,7 +73,7 @@ class BootLiteSync: def add_single_profile(self, name): # get the profile object: - profile = self.profiles.find(name) + profile = self.profiles.find(name=name) if profile is None: raise CX(_("error in profile lookup")) # rebuild profile_list YAML file in webdir @@ -93,7 +93,7 @@ class BootLiteSync: def add_single_system(self, name): # get the system object: - system = self.systems.find(name) + system = self.systems.find(name=name) if system is None: raise CX(_("error in system lookup for %s") % name) # rebuild system_list file in webdir @@ -106,7 +106,7 @@ class BootLiteSync: self.sync.validate_kickstart_for_specific_system(system) def remove_single_system(self, name): - system_record = self.systems.find(name) + system_record = self.systems.find(name=name) # rebuild system_list file in webdir self.sync.write_listings() # delete system YAML file in systems/$name in webdir @@ -118,11 +118,11 @@ class BootLiteSync: # delete PXE Linux configuration file (which might be in one of two places) itanic = False - system_record = self.systems.find(name) - profile = self.profiles.find(system_record.profile) + system_record = self.systems.find(name=name) + profile = self.profiles.find(name=system_record.profile) # allow cobbler deletes to still work in the cobbler config is discombobulated if profile is not None: - distro = self.distros.find(profile.distro) + distro = self.distros.find(name=profile.distro) if distro is not None and distro in [ "ia64", "IA64"]: itanic = True if not itanic: diff --git a/cobbler/action_status.py b/cobbler/action_status.py index 158ee30..660fb42 100644 --- a/cobbler/action_status.py +++ b/cobbler/action_status.py @@ -17,6 +17,7 @@ import os import os.path import glob import time +import api as cobbler_api from rhpl.translate import _, N_, textdomain, utf8 @@ -99,6 +100,8 @@ class BootStatusReport: tracking will be incomplete. This should be noted in the docs. """ + api = cobbler_api.BootAPI() + apache_results = self.scan_apache_logfiles() syslog_results = self.scan_syslog_logfiles() ips = apache_results.keys() @@ -112,9 +115,9 @@ class BootStatusReport: last_recorded_time = 0 time_collisions = 0 - #header = ("Address", "State", "Started", "Last Request", "Seconds", "Log Entries") + #header = ("Name", "State", "Started", "Last Request", "Seconds", "Log Entries") print "%-20s | %-15s | %-25s | %-25s | %-10s | %-6s" % ( - _("Address"), + _("Name"), _("State"), _("Last Request"), _("Started"), @@ -150,20 +153,19 @@ class BootStatusReport: else: entries[logtime] = "1" - - self.generate_report(entries,ip) + name = api.systems().find(ip_address=ip).name + self.generate_report(entries,name) return True #----------------------------------------- - def generate_report(self,entries,ip): + def generate_report(self,entries,name): """ Given the information about transferred files and kickstart finish times, attempt to produce a report that most describes the state of the system. """ - # sort the access times rtimes = entries.keys() rtimes.sort() @@ -175,7 +177,7 @@ class BootStatusReport: fcount = 0 if len(rtimes) == 0: - print _("%s: ?") % ip + print _("%s: ?") % name return # for each request time the machine has made @@ -215,7 +217,7 @@ class BootStatusReport: # print the status line for this IP address print "%-20s | %-15s | %-25s | %-25s | %-10s | %-6s" % ( - ip, + name, install_state, display_start, display_last, diff --git a/cobbler/action_sync.py b/cobbler/action_sync.py index ae65ccc..1e12aa8 100644 --- a/cobbler/action_sync.py +++ b/cobbler/action_sync.py @@ -367,7 +367,7 @@ class BootSync: buf = "" repos = profile.repos for r in repos: - repo = self.repos.find(r) + repo = self.repos.find(name=r) if repo is None: continue http_url = "http://%s/cblr/repo_mirror/%s" % (self.settings.server, repo.name) @@ -389,7 +389,7 @@ class BootSync: repos = profile.repos buf = "" for r in repos: - repo = self.repos.find(r) + repo = self.repos.find(name=r) if repo is None: continue repo.local_filename = repo.local_filename.replace(".repo","") diff --git a/cobbler/cobbler.py b/cobbler/cobbler.py index 13daabd..0f34eee 100755 --- a/cobbler/cobbler.py +++ b/cobbler/cobbler.py @@ -200,7 +200,7 @@ class BootCLI: def __list_names2(self, collection, args): for p in args: - obj = collection.find(p) + obj = collection.find(name=p) if obj is not None: print obj.printable() return True @@ -269,7 +269,7 @@ class BootCLI: control_fn(args,obj) def __generic_edit(self,args,collection_fn,control_fn,exc_msg): - obj = collection_fn().find(self.find_arg(args,"--name")) + obj = collection_fn().find(name=self.find_arg(args,"--name")) name2 = self.find_arg(args,"--newname") if name2 is not None: raise CX("objects cannot be renamed with the edit command, use 'rename'") @@ -278,7 +278,7 @@ class BootCLI: control_fn(args,obj) def __generic_copy(self,args,collection_fn,control_fn,exc_msg): - obj = collection_fn().find(self.find_arg(args,"--name")) + obj = collection_fn().find(name=self.find_arg(args,"--name")) obj2 = self.find_arg(args,"--newname") if obj is None: raise CX(exc_msg) diff --git a/cobbler/cobblerd.py b/cobbler/cobblerd.py index 54294cb..d9a5ced 100644 --- a/cobbler/cobblerd.py +++ b/cobbler/cobblerd.py @@ -71,10 +71,11 @@ def do_syslog(bootapi, settings, port, logger): while 1: data, addr = s.recvfrom(buf) (ip, port) = addr - if not data: + name = bootapi.systems().find(ip_address = ip) + if not data and name: break else: - logfile = open("/var/log/cobbler/syslog/%s" % ip, "a+") + logfile = open("/var/log/cobbler/syslog/%s" % name, "a+") t = time.localtime() # write numeric time seconds = str(time.mktime(t)) @@ -118,7 +119,7 @@ class CobblerXMLRPCInterface: # feature disabled! return False systems = self.api.systems() - obj = systems.find(name) + obj = systems.find(name=name) if obj == None: # system not found! return False @@ -148,7 +149,7 @@ class CobblerXMLRPCInterface: def __get_specific(self,collection,name): self.api.clear() self.api.deserialize() - item = collection.find(name) + item = collection.find(name=name) if item is None: return self.fix_none({}) return self.fix_none(item.to_datastruct()) @@ -169,7 +170,7 @@ class CobblerXMLRPCInterface: def get_distro_for_koan(self,name): self.api.clear() self.api.deserialize() - obj = cobbler_api.distros().find(name) + obj = cobbler_api.distros().find(name=name) if obj is not None: return self.fix_none(utils.blender(True, obj)) return self.fix_none({}) @@ -177,7 +178,7 @@ class CobblerXMLRPCInterface: def get_profile_for_koan(self,name): self.api.clear() self.api.deserialize() - obj = self.api.profiles().find(name) + obj = self.api.profiles().find(name=name) if obj is not None: return self.fix_none(utils.blender(True, obj)) return self.fix_none({}) @@ -185,7 +186,7 @@ class CobblerXMLRPCInterface: def get_system_for_koan(self,name): self.api.clear() self.api.deserialize() - obj = self.api.systems().find(name) + obj = self.api.systems().find(name=name) if obj is not None: return self.fix_none(utils.blender(True, obj)) return self.fix_none({}) @@ -193,7 +194,7 @@ class CobblerXMLRPCInterface: def get_repo_for_koan(self,name): self.api.clear() self.api.deserialize() - obj = self.api.repos().find(name) + obj = self.api.repos().find(name=name) if obj is not None: return self.fix_none(utils.blender(True, obj)) return self.fix_none({}) diff --git a/cobbler/collection.py b/cobbler/collection.py index 0c743f4..b2e3628 100644 --- a/cobbler/collection.py +++ b/cobbler/collection.py @@ -30,8 +30,8 @@ class Collection(serializable.Serializable): def __init__(self,config): """ - Constructor. - """ + Constructor. + """ self.config = config self.clear() @@ -54,17 +54,17 @@ class Collection(serializable.Serializable): """ self.listing = {} - def find(self,name): + def find(self, **kargs): """ - Return anything named 'name' in the collection, else return None if - no objects can be found. + Return first object in the collection that maches all item='value' + pairs passed, else return None if no objects can be found. """ - n1 = name.lower() - listing = self.listing - if listing.has_key(n1): - return self.listing[n1] - else: - return None + kargs = dict((k.lower(), v.lower()) for k, v in kargs.items()) + for entry in self.to_datastruct(): + match = dict((k.lower(), v.lower()) for k, v in entry.items() if kargs.has_key(k)) + if match == kargs: + return self.listing[entry['name'].lower()] + return None def to_datastruct(self): """ diff --git a/cobbler/collection_distros.py b/cobbler/collection_distros.py index 70029eb..9ab9e3c 100644 --- a/cobbler/collection_distros.py +++ b/cobbler/collection_distros.py @@ -51,7 +51,7 @@ class Distros(collection.Collection): for v in self.config.profiles(): if v.distro == name: raise CX(_("removal would orphan profile: %s") % v.name) - if self.find(name): + if self.find(name=name): if with_delete: self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/distro/pre/*") lite_sync = action_litesync.BootLiteSync(self.config) diff --git a/cobbler/collection_profiles.py b/cobbler/collection_profiles.py index ac1798e..7331b08 100644 --- a/cobbler/collection_profiles.py +++ b/cobbler/collection_profiles.py @@ -49,7 +49,7 @@ class Profiles(collection.Collection): for v in self.config.systems(): if v.profile == name: raise CX(_("removal would orphan system: %s") % v.name) - if self.find(name): + if self.find(name=name): if with_delete: self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/profile/pre/*") lite_sync = action_litesync.BootLiteSync(self.config) diff --git a/cobbler/collection_repos.py b/cobbler/collection_repos.py index 8ffa1bf..ed62d88 100644 --- a/cobbler/collection_repos.py +++ b/cobbler/collection_repos.py @@ -53,7 +53,7 @@ class Repos(collection.Collection): # NOTE: with_delete isn't currently meaningful for repos # but is left in for consistancy in the API. Unused. name = name.lower() - if self.find(name): + if self.find(name=name): if with_delete: self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/repo/pre/*") self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/repo/post/*") diff --git a/cobbler/collection_systems.py b/cobbler/collection_systems.py index 5ffcea5..58f9e13 100644 --- a/cobbler/collection_systems.py +++ b/cobbler/collection_systems.py @@ -49,7 +49,7 @@ class Systems(collection.Collection): Remove element named 'name' from the collection """ name = name.lower() - if self.find(name): + if self.find(name=name): if with_delete: self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/system/pre/*") lite_sync = action_litesync.BootLiteSync(self.config) diff --git a/cobbler/item_distro.py b/cobbler/item_distro.py index 4bbf32f..4a97fcf 100644 --- a/cobbler/item_distro.py +++ b/cobbler/item_distro.py @@ -55,7 +55,7 @@ class Distro(item.Item): if self.parent is None or self.parent == '': return None else: - return self.config.distros().find(self.parent) + return self.config.distros().find(name=self.parent) def from_datastruct(self,seed_data): """ diff --git a/cobbler/item_profile.py b/cobbler/item_profile.py index c7c9814..4062826 100644 --- a/cobbler/item_profile.py +++ b/cobbler/item_profile.py @@ -91,7 +91,7 @@ class Profile(item.Item): # check must be done in two places as set_parent could be called before/after # set_name... raise CX(_("self parentage is weird")) - found = self.config.profiles().find(parent_name) + found = self.config.profiles().find(name=parent_name) if found is None: raise CX(_("profile %s not found, inheritance not possible") % parent_name) self.parent = parent_name @@ -102,7 +102,7 @@ class Profile(item.Item): Sets the distro. This must be the name of an existing Distro object in the Distros collection. """ - d = self.config.distros().find(distro_name) + d = self.config.distros().find(name=distro_name) if d is not None: self.distro = distro_name self.depth = d.depth +1 # reset depth if previously a subprofile and now top-level @@ -125,7 +125,7 @@ class Profile(item.Item): except: pass for r in repolist: - if not self.config.repos().find(r): + if not self.config.repos().find(name=r): ok = False break if ok: @@ -203,9 +203,9 @@ class Profile(item.Item): Return object next highest up the tree. """ if self.parent is None or self.parent == '': - result = self.config.distros().find(self.distro) + result = self.config.distros().find(name=self.distro) else: - result = self.config.profiles().find(self.parent) + result = self.config.profiles().find(name=self.parent) return result def is_valid(self): diff --git a/cobbler/item_system.py b/cobbler/item_system.py index cb7d02f..3575d79 100644 --- a/cobbler/item_system.py +++ b/cobbler/item_system.py @@ -93,9 +93,9 @@ class System(item.Item): Return object next highest up the tree. """ if self.parent is None or self.parent == '': - return self.config.profiles().find(self.profile) + return self.config.profiles().find(name=self.profile) else: - return self.config.systems().find(self.parent) + return self.config.systems().find(name=self.parent) def set_name(self,name): """ @@ -180,10 +180,10 @@ class System(item.Item): def set_profile(self,profile_name): """ - Set the system to use a certain named profile. The profile - must have already been loaded into the Profiles collection. - """ - p = self.config.profiles().find(profile_name) + Set the system to use a certain named profile. The profile + must have already been loaded into the Profiles collection. + """ + p = self.config.profiles().find(name=profile_name) if p is not None: self.profile = profile_name self.depth = p.depth + 1 # subprofiles have varying depths. @@ -229,8 +229,8 @@ class System(item.Item): def is_valid(self): """ - A system is valid when it contains a valid name and a profile. - """ + A system is valid when it contains a valid name and a profile. + """ # NOTE: this validation code does not support inheritable distros at this time. # this is by design as inheritable systems don't make sense. if self.name is None: diff --git a/tests/tests.py b/tests/tests.py index 2f082e7..97b685c 100644 --- a/tests/tests.py +++ b/tests/tests.py @@ -84,20 +84,20 @@ class BootTest(unittest.TestCase): self.assertTrue(distro.set_kernel(self.fk_kernel)) self.assertTrue(distro.set_initrd(self.fk_initrd)) self.assertTrue(self.api.distros().add(distro)) - self.assertTrue(self.api.distros().find("testdistro0")) + self.assertTrue(self.api.distros().find(name="testdistro0")) profile = self.api.new_profile() self.assertTrue(profile.set_name("testprofile0")) self.assertTrue(profile.set_distro("testdistro0")) self.assertTrue(profile.set_kickstart(FAKE_KICKSTART)) self.assertTrue(self.api.profiles().add(profile)) - self.assertTrue(self.api.profiles().find("testprofile0")) + self.assertTrue(self.api.profiles().find(name="testprofile0")) system = self.api.new_system() self.assertTrue(system.set_name(self.hostname)) self.assertTrue(system.set_profile("testprofile0")) self.assertTrue(self.api.systems().add(system)) - self.assertTrue(self.api.systems().find(self.hostname)) + self.assertTrue(self.api.systems().find(name=self.hostname)) class Utilities(BootTest): @@ -149,7 +149,7 @@ class Additions(BootTest): self.failUnlessRaises(CobblerException,distro.set_kernel,"filedoesntexist") self.assertTrue(distro.set_initrd(self.fk_initrd)) self.failUnlessRaises(CobblerException, self.api.distros().add, distro) - self.assertFalse(self.api.distros().find("testdistro2")) + self.assertFalse(self.api.distros().find(name="testdistro2")) def test_invalid_distro_non_referenced_initrd(self): distro = self.api.new_distro() @@ -157,7 +157,7 @@ class Additions(BootTest): self.assertTrue(distro.set_kernel(self.fk_kernel)) self.failUnlessRaises(CobblerException, distro.set_initrd, "filedoesntexist") self.failUnlessRaises(CobblerException, self.api.distros().add, distro) - self.assertFalse(self.api.distros().find("testdistro3")) + self.assertFalse(self.api.distros().find(name="testdistro3")) def test_invalid_profile_non_referenced_distro(self): profile = self.api.new_profile() @@ -165,7 +165,7 @@ class Additions(BootTest): self.failUnlessRaises(CobblerException, profile.set_distro, "distrodoesntexist") self.assertTrue(profile.set_kickstart(FAKE_KICKSTART)) self.failUnlessRaises(CobblerException, self.api.profiles().add, profile) - self.assertFalse(self.api.profiles().find("testprofile2")) + self.assertFalse(self.api.profiles().find(name="testprofile2")) def test_invalid_profile_kickstart_not_url(self): profile = self.api.new_profile() @@ -174,7 +174,7 @@ class Additions(BootTest): self.failUnlessRaises(CobblerException, profile.set_kickstart, "kickstartdoesntexist") # since kickstarts are optional, you can still add it self.assertTrue(self.api.profiles().add(profile)) - self.assertTrue(self.api.profiles().find("testprofile12")) + self.assertTrue(self.api.profiles().find(name="testprofile12")) # now verify the other kickstart forms would still work self.assertTrue(profile.set_kickstart("http://bar")) self.assertTrue(profile.set_kickstart("ftp://bar")) @@ -199,7 +199,7 @@ class Additions(BootTest): self.assertTrue(system.set_name(name)) self.assertTrue(system.set_profile("testprofile0")) self.assertTrue(self.api.systems().add(system)) - self.assertTrue(self.api.systems().find(name)) + self.assertTrue(self.api.systems().find(name=name)) def test_system_name_is_an_IP(self): system = self.api.new_system() @@ -207,7 +207,7 @@ class Additions(BootTest): self.assertTrue(system.set_name(name)) self.assertTrue(system.set_profile("testprofile0")) self.assertTrue(self.api.systems().add(system)) - self.assertTrue(self.api.systems().find(name)) + self.assertTrue(self.api.systems().find(name=name)) def test_invalid_system_non_referenced_profile(self): system = self.api.new_system() @@ -241,9 +241,9 @@ class Deletions(BootTest): self.api.serialize() self.assertTrue(self.api.profiles().remove("testprofile0")) self.assertTrue(self.api.distros().remove("testdistro0")) - self.assertFalse(self.api.systems().find(self.hostname)) - self.assertFalse(self.api.profiles().find("testprofile0")) - self.assertFalse(self.api.distros().find("testdistro0")) + self.assertFalse(self.api.systems().find(name=self.hostname)) + self.assertFalse(self.api.profiles().find(name="testprofile0")) + self.assertFalse(self.api.distros().find(name="testdistro0")) class TestCheck(BootTest): -- 1.5.2.2 From mazurekd at cpsc.ucalgary.ca Wed Aug 15 18:41:05 2007 From: mazurekd at cpsc.ucalgary.ca (Debbie Mazurek) Date: Wed, 15 Aug 2007 12:41:05 -0600 Subject: [et-mgmt-tools] Cobbler version In-Reply-To: <46C2E283.3060903@redhat.com> References: <1187133631.18482.181.camel@caliente> <46C2E283.3060903@redhat.com> Message-ID: <1187203265.18482.203.camel@caliente> > I asked about this last night, and EPEL hasn't done a push yet. Once > that happens, 0.6.0 will be in > both EL 4 and EL 5. > > Rebuilding the source RPM should be fine, though you'll likely have to > find a few dependencies -- which you can > get from EPEL. > > Sidenote: I believe EPEL-4 is still waiting for the inclusion of yum > and yum-utils in EPEL4 (ETA: soon), so cobbler won't > move into EL 4 "non-testing" until yum-utils is part of EL4. Cobbler > requires yum-utils as a dependency > for the repo management features. This isn't an issue for EL5. > > OK, thanks for the info. This is for EL4. I'm having a bit of trouble sorting out the deps for cheetah, but that for sure is not a cobbler problem. Also, is there a minimum or specific version of yum and yum-utils that cobbler-0.6 needs? I didn't see anything in cobbler's spec file. -- d e b b i e m a z u r e k mazurekd at cpsc.ucalgary.ca system admin . department of computer science . university of calgary From hyclak at math.ohiou.edu Wed Aug 15 20:16:07 2007 From: hyclak at math.ohiou.edu (Matt Hyclak) Date: Wed, 15 Aug 2007 16:16:07 -0400 Subject: [et-mgmt-tools] Cobbler version In-Reply-To: <1187203265.18482.203.camel@caliente> References: <1187133631.18482.181.camel@caliente> <46C2E283.3060903@redhat.com> <1187203265.18482.203.camel@caliente> Message-ID: <20070815201607.GP5268@math.ohiou.edu> On Wed, Aug 15, 2007 at 12:41:05PM -0600, Debbie Mazurek enlightened us: > > I asked about this last night, and EPEL hasn't done a push yet. Once > > that happens, 0.6.0 will be in > > both EL 4 and EL 5. > > > > Rebuilding the source RPM should be fine, though you'll likely have to > > find a few dependencies -- which you can > > get from EPEL. > > > > Sidenote: I believe EPEL-4 is still waiting for the inclusion of yum > > and yum-utils in EPEL4 (ETA: soon), so cobbler won't > > move into EL 4 "non-testing" until yum-utils is part of EL4. Cobbler > > requires yum-utils as a dependency > > for the repo management features. This isn't an issue for EL5. > > > > > > OK, thanks for the info. This is for EL4. > I'm having a bit of trouble sorting out the deps for cheetah, > but that for sure is not a cobbler problem. > http://www.math.ohiou.edu/pub/casit/et-tools/ I've pretty much got everything for EL4 in there, and what I have is in the process of being moved into the CentOS distribution system. We're still working out some details there of how that will happen, but it will happen :-) > Also, is there a minimum or specific version of yum and yum-utils that > cobbler-0.6 needs? I didn't see anything in cobbler's spec file. I use the latest from CentOS and it works just fine. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263 From rigg0022 at umn.edu Wed Aug 15 20:17:37 2007 From: rigg0022 at umn.edu (Ben Riggs) Date: Wed, 15 Aug 2007 15:17:37 -0500 Subject: [et-mgmt-tools] [PATCH] Fixed naming error for log files. Message-ID: <11872090571865-git-send-email-rigg0022@umn.edu> Signed-off-by: Ben Riggs --- cobbler/cobblerd.py | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/cobbler/cobblerd.py b/cobbler/cobblerd.py index d9a5ced..d2ef336 100644 --- a/cobbler/cobblerd.py +++ b/cobbler/cobblerd.py @@ -71,11 +71,11 @@ def do_syslog(bootapi, settings, port, logger): while 1: data, addr = s.recvfrom(buf) (ip, port) = addr - name = bootapi.systems().find(ip_address = ip) - if not data and name: + system = bootapi.systems().find(ip_address = ip) + if not data and system: break else: - logfile = open("/var/log/cobbler/syslog/%s" % name, "a+") + logfile = open("/var/log/cobbler/syslog/%s" % system.name, "a+") t = time.localtime() # write numeric time seconds = str(time.mktime(t)) -- 1.5.2.2 From mdehaan at redhat.com Wed Aug 15 20:46:16 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 15 Aug 2007 16:46:16 -0400 Subject: [et-mgmt-tools] Cobbler version In-Reply-To: <1187203265.18482.203.camel@caliente> References: <1187133631.18482.181.camel@caliente> <46C2E283.3060903@redhat.com> <1187203265.18482.203.camel@caliente> Message-ID: <46C36618.3080509@redhat.com> Debbie Mazurek wrote: >> I asked about this last night, and EPEL hasn't done a push yet. Once >> that happens, 0.6.0 will be in >> both EL 4 and EL 5. >> >> Rebuilding the source RPM should be fine, though you'll likely have to >> find a few dependencies -- which you can >> get from EPEL. >> >> Sidenote: I believe EPEL-4 is still waiting for the inclusion of yum >> and yum-utils in EPEL4 (ETA: soon), so cobbler won't >> move into EL 4 "non-testing" until yum-utils is part of EL4. Cobbler >> requires yum-utils as a dependency >> for the repo management features. This isn't an issue for EL5. >> >> >> > > OK, thanks for the info. This is for EL4. > I'm having a bit of trouble sorting out the deps for cheetah, > but that for sure is not a cobbler problem. > python-cheetah should also be in EPEL. So if you add the yum repository and install that from yum, life should be in good. If not, I'd be interested in what's not good -- that would be a bug in Cheetah for EPEL. What Matt said also works. > Also, is there a minimum or specific version of yum and yum-utils that > cobbler-0.6 needs? I didn't see anything in cobbler's spec file. > > > Not really. For RHEL the versions out of Centos are good -- and we'll have those in EPEL soon enough. From mdehaan at redhat.com Wed Aug 15 20:51:42 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 15 Aug 2007 16:51:42 -0400 Subject: [et-mgmt-tools] [PATCH] Changed collection.find() to take a dictionary argument instead of name, allowing searches by ip_address, mac_address, etc. In-Reply-To: <11871933602647-git-send-email-rigg0022@umn.edu> References: <11871933602647-git-send-email-rigg0022@umn.edu> Message-ID: <46C3675E.5020205@redhat.com> Ben Riggs wrote: > From: root > > diff --git a/cobbler/action_enchant.py b/cobbler/action_enchant.py > index 3bf80f8..4a54ebe 100644 > --- a/cobbler/action_enchant.py > +++ b/cobbler/action_enchant.py > @@ -40,9 +40,9 @@ class Enchant: > raise CX(_("enchant failed. no address specified")) > if system is None and profile is None: > raise CX(_("enchant failed. no profile specified")) > - if system is not None and self.config.systems().find(system) is None: > + if system is not None and self.config.systems().find(name=system) is None: > raise CX(_("enchant failed. system not found")) > - if profile is not None and self.config.profiles().find(profile) is None: > + if profile is not None and self.config.profiles().find(name=profile) is None: > raise CX(_("enchant failed. profile name not found")) > > > diff --git a/cobbler/action_import.py b/cobbler/action_import.py > index 5789ae0..c74bc81 100644 > --- a/cobbler/action_import.py > +++ b/cobbler/action_import.py > @@ -155,7 +155,7 @@ class Importer: > """ > > for profile in self.profiles: > - distro = self.distros.find(profile.distro) > + distro = self.distros.find(name=profile.distro) > if distro is None or not (distro in self.distros_added): > print _("- skipping distro %s since it wasn't imported this time") % profile.distro > continue > @@ -429,7 +429,7 @@ class Importer: > pxe_arch = self.get_pxe_arch(dirname) > name = self.get_proposed_name(dirname, pxe_arch) > > - existing_distro = self.distros.find(name) > + existing_distro = self.distros.find(name=name) > > if existing_distro is not None: > print _("- modifying existing distro: %s") % name > @@ -446,7 +446,7 @@ class Importer: > self.distros.add(distro) > self.distros_added.append(distro) > > - existing_profile = self.profiles.find(name) > + existing_profile = self.profiles.find(name=name) > > if existing_profile is None: > print _("- creating new profile: %s") % name > diff --git a/cobbler/action_litesync.py b/cobbler/action_litesync.py > index 21e6c62..2560d8d 100644 > --- a/cobbler/action_litesync.py > +++ b/cobbler/action_litesync.py > @@ -53,7 +53,7 @@ class BootLiteSync: > > def add_single_distro(self, name): > # get the distro record > - distro = self.distros.find(name) > + distro = self.distros.find(name=name) > if distro is None: > raise CX(_("error in distro lookup: %s") % name) > # generate YAML file in distros/$name in webdir > @@ -73,7 +73,7 @@ class BootLiteSync: > > def add_single_profile(self, name): > # get the profile object: > - profile = self.profiles.find(name) > + profile = self.profiles.find(name=name) > if profile is None: > raise CX(_("error in profile lookup")) > # rebuild profile_list YAML file in webdir > @@ -93,7 +93,7 @@ class BootLiteSync: > > def add_single_system(self, name): > # get the system object: > - system = self.systems.find(name) > + system = self.systems.find(name=name) > if system is None: > raise CX(_("error in system lookup for %s") % name) > # rebuild system_list file in webdir > @@ -106,7 +106,7 @@ class BootLiteSync: > self.sync.validate_kickstart_for_specific_system(system) > > def remove_single_system(self, name): > - system_record = self.systems.find(name) > + system_record = self.systems.find(name=name) > # rebuild system_list file in webdir > self.sync.write_listings() > # delete system YAML file in systems/$name in webdir > @@ -118,11 +118,11 @@ class BootLiteSync: > > # delete PXE Linux configuration file (which might be in one of two places) > itanic = False > - system_record = self.systems.find(name) > - profile = self.profiles.find(system_record.profile) > + system_record = self.systems.find(name=name) > + profile = self.profiles.find(name=system_record.profile) > # allow cobbler deletes to still work in the cobbler config is discombobulated > if profile is not None: > - distro = self.distros.find(profile.distro) > + distro = self.distros.find(name=profile.distro) > if distro is not None and distro in [ "ia64", "IA64"]: > itanic = True > if not itanic: > diff --git a/cobbler/action_status.py b/cobbler/action_status.py > index 158ee30..660fb42 100644 > --- a/cobbler/action_status.py > +++ b/cobbler/action_status.py > @@ -17,6 +17,7 @@ import os > import os.path > import glob > import time > +import api as cobbler_api > > from rhpl.translate import _, N_, textdomain, utf8 > > @@ -99,6 +100,8 @@ class BootStatusReport: > tracking will be incomplete. This should be noted in the docs. > """ > > + api = cobbler_api.BootAPI() > + > apache_results = self.scan_apache_logfiles() > syslog_results = self.scan_syslog_logfiles() > ips = apache_results.keys() > @@ -112,9 +115,9 @@ class BootStatusReport: > last_recorded_time = 0 > time_collisions = 0 > > - #header = ("Address", "State", "Started", "Last Request", "Seconds", "Log Entries") > + #header = ("Name", "State", "Started", "Last Request", "Seconds", "Log Entries") > print "%-20s | %-15s | %-25s | %-25s | %-10s | %-6s" % ( > - _("Address"), > + _("Name"), > _("State"), > _("Last Request"), > _("Started"), > @@ -150,20 +153,19 @@ class BootStatusReport: > else: > entries[logtime] = "1" > > - > - self.generate_report(entries,ip) > + name = api.systems().find(ip_address=ip).name > + self.generate_report(entries,name) > > > return True > > #----------------------------------------- > > - def generate_report(self,entries,ip): > + def generate_report(self,entries,name): > """ > Given the information about transferred files and kickstart finish times, attempt > to produce a report that most describes the state of the system. > """ > - > # sort the access times > rtimes = entries.keys() > rtimes.sort() > @@ -175,7 +177,7 @@ class BootStatusReport: > fcount = 0 > > if len(rtimes) == 0: > - print _("%s: ?") % ip > + print _("%s: ?") % name > return > > # for each request time the machine has made > @@ -215,7 +217,7 @@ class BootStatusReport: > > # print the status line for this IP address > print "%-20s | %-15s | %-25s | %-25s | %-10s | %-6s" % ( > - ip, > + name, > install_state, > display_start, > display_last, > diff --git a/cobbler/action_sync.py b/cobbler/action_sync.py > index ae65ccc..1e12aa8 100644 > --- a/cobbler/action_sync.py > +++ b/cobbler/action_sync.py > @@ -367,7 +367,7 @@ class BootSync: > buf = "" > repos = profile.repos > for r in repos: > - repo = self.repos.find(r) > + repo = self.repos.find(name=r) > if repo is None: > continue > http_url = "http://%s/cblr/repo_mirror/%s" % (self.settings.server, repo.name) > @@ -389,7 +389,7 @@ class BootSync: > repos = profile.repos > buf = "" > for r in repos: > - repo = self.repos.find(r) > + repo = self.repos.find(name=r) > if repo is None: > continue > repo.local_filename = repo.local_filename.replace(".repo","") > diff --git a/cobbler/cobbler.py b/cobbler/cobbler.py > index 13daabd..0f34eee 100755 > --- a/cobbler/cobbler.py > +++ b/cobbler/cobbler.py > @@ -200,7 +200,7 @@ class BootCLI: > > def __list_names2(self, collection, args): > for p in args: > - obj = collection.find(p) > + obj = collection.find(name=p) > if obj is not None: > print obj.printable() > return True > @@ -269,7 +269,7 @@ class BootCLI: > control_fn(args,obj) > > def __generic_edit(self,args,collection_fn,control_fn,exc_msg): > - obj = collection_fn().find(self.find_arg(args,"--name")) > + obj = collection_fn().find(name=self.find_arg(args,"--name")) > name2 = self.find_arg(args,"--newname") > if name2 is not None: > raise CX("objects cannot be renamed with the edit command, use 'rename'") > @@ -278,7 +278,7 @@ class BootCLI: > control_fn(args,obj) > > def __generic_copy(self,args,collection_fn,control_fn,exc_msg): > - obj = collection_fn().find(self.find_arg(args,"--name")) > + obj = collection_fn().find(name=self.find_arg(args,"--name")) > obj2 = self.find_arg(args,"--newname") > if obj is None: > raise CX(exc_msg) > diff --git a/cobbler/cobblerd.py b/cobbler/cobblerd.py > index 54294cb..d9a5ced 100644 > --- a/cobbler/cobblerd.py > +++ b/cobbler/cobblerd.py > @@ -71,10 +71,11 @@ def do_syslog(bootapi, settings, port, logger): > while 1: > data, addr = s.recvfrom(buf) > (ip, port) = addr > - if not data: > + name = bootapi.systems().find(ip_address = ip) > + if not data and name: > break > else: > - logfile = open("/var/log/cobbler/syslog/%s" % ip, "a+") > + logfile = open("/var/log/cobbler/syslog/%s" % name, "a+") > t = time.localtime() > # write numeric time > seconds = str(time.mktime(t)) > @@ -118,7 +119,7 @@ class CobblerXMLRPCInterface: > # feature disabled! > return False > systems = self.api.systems() > - obj = systems.find(name) > + obj = systems.find(name=name) > if obj == None: > # system not found! > return False > @@ -148,7 +149,7 @@ class CobblerXMLRPCInterface: > def __get_specific(self,collection,name): > self.api.clear() > self.api.deserialize() > - item = collection.find(name) > + item = collection.find(name=name) > if item is None: > return self.fix_none({}) > return self.fix_none(item.to_datastruct()) > @@ -169,7 +170,7 @@ class CobblerXMLRPCInterface: > def get_distro_for_koan(self,name): > self.api.clear() > self.api.deserialize() > - obj = cobbler_api.distros().find(name) > + obj = cobbler_api.distros().find(name=name) > if obj is not None: > return self.fix_none(utils.blender(True, obj)) > return self.fix_none({}) > @@ -177,7 +178,7 @@ class CobblerXMLRPCInterface: > def get_profile_for_koan(self,name): > self.api.clear() > self.api.deserialize() > - obj = self.api.profiles().find(name) > + obj = self.api.profiles().find(name=name) > if obj is not None: > return self.fix_none(utils.blender(True, obj)) > return self.fix_none({}) > @@ -185,7 +186,7 @@ class CobblerXMLRPCInterface: > def get_system_for_koan(self,name): > self.api.clear() > self.api.deserialize() > - obj = self.api.systems().find(name) > + obj = self.api.systems().find(name=name) > if obj is not None: > return self.fix_none(utils.blender(True, obj)) > return self.fix_none({}) > @@ -193,7 +194,7 @@ class CobblerXMLRPCInterface: > def get_repo_for_koan(self,name): > self.api.clear() > self.api.deserialize() > - obj = self.api.repos().find(name) > + obj = self.api.repos().find(name=name) > if obj is not None: > return self.fix_none(utils.blender(True, obj)) > return self.fix_none({}) > diff --git a/cobbler/collection.py b/cobbler/collection.py > index 0c743f4..b2e3628 100644 > --- a/cobbler/collection.py > +++ b/cobbler/collection.py > @@ -30,8 +30,8 @@ class Collection(serializable.Serializable): > > def __init__(self,config): > """ > - Constructor. > - """ > + Constructor. > + """ > self.config = config > self.clear() > > @@ -54,17 +54,17 @@ class Collection(serializable.Serializable): > """ > self.listing = {} > > - def find(self,name): > + def find(self, **kargs): > """ > - Return anything named 'name' in the collection, else return None if > - no objects can be found. > + Return first object in the collection that maches all item='value' > + pairs passed, else return None if no objects can be found. > """ > - n1 = name.lower() > - listing = self.listing > - if listing.has_key(n1): > - return self.listing[n1] > - else: > - return None > + kargs = dict((k.lower(), v.lower()) for k, v in kargs.items()) > + for entry in self.to_datastruct(): > + match = dict((k.lower(), v.lower()) for k, v in entry.items() if kargs.has_key(k)) > + if match == kargs: > + return self.listing[entry['name'].lower()] > + return None > > def to_datastruct(self): > """ > diff --git a/cobbler/collection_distros.py b/cobbler/collection_distros.py > index 70029eb..9ab9e3c 100644 > --- a/cobbler/collection_distros.py > +++ b/cobbler/collection_distros.py > @@ -51,7 +51,7 @@ class Distros(collection.Collection): > for v in self.config.profiles(): > if v.distro == name: > raise CX(_("removal would orphan profile: %s") % v.name) > - if self.find(name): > + if self.find(name=name): > if with_delete: > self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/distro/pre/*") > lite_sync = action_litesync.BootLiteSync(self.config) > diff --git a/cobbler/collection_profiles.py b/cobbler/collection_profiles.py > index ac1798e..7331b08 100644 > --- a/cobbler/collection_profiles.py > +++ b/cobbler/collection_profiles.py > @@ -49,7 +49,7 @@ class Profiles(collection.Collection): > for v in self.config.systems(): > if v.profile == name: > raise CX(_("removal would orphan system: %s") % v.name) > - if self.find(name): > + if self.find(name=name): > if with_delete: > self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/profile/pre/*") > lite_sync = action_litesync.BootLiteSync(self.config) > diff --git a/cobbler/collection_repos.py b/cobbler/collection_repos.py > index 8ffa1bf..ed62d88 100644 > --- a/cobbler/collection_repos.py > +++ b/cobbler/collection_repos.py > @@ -53,7 +53,7 @@ class Repos(collection.Collection): > # NOTE: with_delete isn't currently meaningful for repos > # but is left in for consistancy in the API. Unused. > name = name.lower() > - if self.find(name): > + if self.find(name=name): > if with_delete: > self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/repo/pre/*") > self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/repo/post/*") > diff --git a/cobbler/collection_systems.py b/cobbler/collection_systems.py > index 5ffcea5..58f9e13 100644 > --- a/cobbler/collection_systems.py > +++ b/cobbler/collection_systems.py > @@ -49,7 +49,7 @@ class Systems(collection.Collection): > Remove element named 'name' from the collection > """ > name = name.lower() > - if self.find(name): > + if self.find(name=name): > if with_delete: > self._run_triggers(self.listing[name], "/var/lib/cobbler/triggers/delete/system/pre/*") > lite_sync = action_litesync.BootLiteSync(self.config) > diff --git a/cobbler/item_distro.py b/cobbler/item_distro.py > index 4bbf32f..4a97fcf 100644 > --- a/cobbler/item_distro.py > +++ b/cobbler/item_distro.py > @@ -55,7 +55,7 @@ class Distro(item.Item): > if self.parent is None or self.parent == '': > return None > else: > - return self.config.distros().find(self.parent) > + return self.config.distros().find(name=self.parent) > > def from_datastruct(self,seed_data): > """ > diff --git a/cobbler/item_profile.py b/cobbler/item_profile.py > index c7c9814..4062826 100644 > --- a/cobbler/item_profile.py > +++ b/cobbler/item_profile.py > @@ -91,7 +91,7 @@ class Profile(item.Item): > # check must be done in two places as set_parent could be called before/after > # set_name... > raise CX(_("self parentage is weird")) > - found = self.config.profiles().find(parent_name) > + found = self.config.profiles().find(name=parent_name) > if found is None: > raise CX(_("profile %s not found, inheritance not possible") % parent_name) > self.parent = parent_name > @@ -102,7 +102,7 @@ class Profile(item.Item): > Sets the distro. This must be the name of an existing > Distro object in the Distros collection. > """ > - d = self.config.distros().find(distro_name) > + d = self.config.distros().find(name=distro_name) > if d is not None: > self.distro = distro_name > self.depth = d.depth +1 # reset depth if previously a subprofile and now top-level > @@ -125,7 +125,7 @@ class Profile(item.Item): > except: > pass > for r in repolist: > - if not self.config.repos().find(r): > + if not self.config.repos().find(name=r): > ok = False > break > if ok: > @@ -203,9 +203,9 @@ class Profile(item.Item): > Return object next highest up the tree. > """ > if self.parent is None or self.parent == '': > - result = self.config.distros().find(self.distro) > + result = self.config.distros().find(name=self.distro) > else: > - result = self.config.profiles().find(self.parent) > + result = self.config.profiles().find(name=self.parent) > return result > > def is_valid(self): > diff --git a/cobbler/item_system.py b/cobbler/item_system.py > index cb7d02f..3575d79 100644 > --- a/cobbler/item_system.py > +++ b/cobbler/item_system.py > @@ -93,9 +93,9 @@ class System(item.Item): > Return object next highest up the tree. > """ > if self.parent is None or self.parent == '': > - return self.config.profiles().find(self.profile) > + return self.config.profiles().find(name=self.profile) > else: > - return self.config.systems().find(self.parent) > + return self.config.systems().find(name=self.parent) > > def set_name(self,name): > """ > @@ -180,10 +180,10 @@ class System(item.Item): > > def set_profile(self,profile_name): > """ > - Set the system to use a certain named profile. The profile > - must have already been loaded into the Profiles collection. > - """ > - p = self.config.profiles().find(profile_name) > + Set the system to use a certain named profile. The profile > + must have already been loaded into the Profiles collection. > + """ > + p = self.config.profiles().find(name=profile_name) > if p is not None: > self.profile = profile_name > self.depth = p.depth + 1 # subprofiles have varying depths. > @@ -229,8 +229,8 @@ class System(item.Item): > > def is_valid(self): > """ > - A system is valid when it contains a valid name and a profile. > - """ > + A system is valid when it contains a valid name and a profile. > + """ > # NOTE: this validation code does not support inheritable distros at this time. > # this is by design as inheritable systems don't make sense. > if self.name is None: > diff --git a/tests/tests.py b/tests/tests.py > index 2f082e7..97b685c 100644 > --- a/tests/tests.py > +++ b/tests/tests.py > @@ -84,20 +84,20 @@ class BootTest(unittest.TestCase): > self.assertTrue(distro.set_kernel(self.fk_kernel)) > self.assertTrue(distro.set_initrd(self.fk_initrd)) > self.assertTrue(self.api.distros().add(distro)) > - self.assertTrue(self.api.distros().find("testdistro0")) > + self.assertTrue(self.api.distros().find(name="testdistro0")) > > profile = self.api.new_profile() > self.assertTrue(profile.set_name("testprofile0")) > self.assertTrue(profile.set_distro("testdistro0")) > self.assertTrue(profile.set_kickstart(FAKE_KICKSTART)) > self.assertTrue(self.api.profiles().add(profile)) > - self.assertTrue(self.api.profiles().find("testprofile0")) > + self.assertTrue(self.api.profiles().find(name="testprofile0")) > > system = self.api.new_system() > self.assertTrue(system.set_name(self.hostname)) > self.assertTrue(system.set_profile("testprofile0")) > self.assertTrue(self.api.systems().add(system)) > - self.assertTrue(self.api.systems().find(self.hostname)) > + self.assertTrue(self.api.systems().find(name=self.hostname)) > > class Utilities(BootTest): > > @@ -149,7 +149,7 @@ class Additions(BootTest): > self.failUnlessRaises(CobblerException,distro.set_kernel,"filedoesntexist") > self.assertTrue(distro.set_initrd(self.fk_initrd)) > self.failUnlessRaises(CobblerException, self.api.distros().add, distro) > - self.assertFalse(self.api.distros().find("testdistro2")) > + self.assertFalse(self.api.distros().find(name="testdistro2")) > > def test_invalid_distro_non_referenced_initrd(self): > distro = self.api.new_distro() > @@ -157,7 +157,7 @@ class Additions(BootTest): > self.assertTrue(distro.set_kernel(self.fk_kernel)) > self.failUnlessRaises(CobblerException, distro.set_initrd, "filedoesntexist") > self.failUnlessRaises(CobblerException, self.api.distros().add, distro) > - self.assertFalse(self.api.distros().find("testdistro3")) > + self.assertFalse(self.api.distros().find(name="testdistro3")) > > def test_invalid_profile_non_referenced_distro(self): > profile = self.api.new_profile() > @@ -165,7 +165,7 @@ class Additions(BootTest): > self.failUnlessRaises(CobblerException, profile.set_distro, "distrodoesntexist") > self.assertTrue(profile.set_kickstart(FAKE_KICKSTART)) > self.failUnlessRaises(CobblerException, self.api.profiles().add, profile) > - self.assertFalse(self.api.profiles().find("testprofile2")) > + self.assertFalse(self.api.profiles().find(name="testprofile2")) > > def test_invalid_profile_kickstart_not_url(self): > profile = self.api.new_profile() > @@ -174,7 +174,7 @@ class Additions(BootTest): > self.failUnlessRaises(CobblerException, profile.set_kickstart, "kickstartdoesntexist") > # since kickstarts are optional, you can still add it > self.assertTrue(self.api.profiles().add(profile)) > - self.assertTrue(self.api.profiles().find("testprofile12")) > + self.assertTrue(self.api.profiles().find(name="testprofile12")) > # now verify the other kickstart forms would still work > self.assertTrue(profile.set_kickstart("http://bar")) > self.assertTrue(profile.set_kickstart("ftp://bar")) > @@ -199,7 +199,7 @@ class Additions(BootTest): > self.assertTrue(system.set_name(name)) > self.assertTrue(system.set_profile("testprofile0")) > self.assertTrue(self.api.systems().add(system)) > - self.assertTrue(self.api.systems().find(name)) > + self.assertTrue(self.api.systems().find(name=name)) > > def test_system_name_is_an_IP(self): > system = self.api.new_system() > @@ -207,7 +207,7 @@ class Additions(BootTest): > self.assertTrue(system.set_name(name)) > self.assertTrue(system.set_profile("testprofile0")) > self.assertTrue(self.api.systems().add(system)) > - self.assertTrue(self.api.systems().find(name)) > + self.assertTrue(self.api.systems().find(name=name)) > > def test_invalid_system_non_referenced_profile(self): > system = self.api.new_system() > @@ -241,9 +241,9 @@ class Deletions(BootTest): > self.api.serialize() > self.assertTrue(self.api.profiles().remove("testprofile0")) > self.assertTrue(self.api.distros().remove("testdistro0")) > - self.assertFalse(self.api.systems().find(self.hostname)) > - self.assertFalse(self.api.profiles().find("testprofile0")) > - self.assertFalse(self.api.distros().find("testdistro0")) > + self.assertFalse(self.api.systems().find(name=self.hostname)) > + self.assertFalse(self.api.profiles().find(name="testprofile0")) > + self.assertFalse(self.api.distros().find(name="testdistro0")) > > class TestCheck(BootTest): > > Thanks root at xt23.math.umn.edu :) As I've talked with Ben on IRC, this is really good, and I've applied a minor variant of this. We keep API compatibility so collection.find("foo") works like collection.find(name="foo"), added exception raising when searching on an argument that cobbler doesn't understand, and added the ability to return a list in the API (collection.find(distro="xyz",return_list=True) -> []). I will add information on this to the API page on the Wiki. In all fairness, my implementation of find is /much/ less elegant than the version posted above :) The change end users will see out of this so far is that "cobbler status" will show the system name instead of the IP if it can figure it out -- which is great. Nothing else is effected. --Michael From mdehaan at redhat.com Wed Aug 15 20:52:01 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 15 Aug 2007 16:52:01 -0400 Subject: [et-mgmt-tools] [PATCH] Fixed naming error for log files. In-Reply-To: <11872090571865-git-send-email-rigg0022@umn.edu> References: <11872090571865-git-send-email-rigg0022@umn.edu> Message-ID: <46C36771.8080000@redhat.com> Ben Riggs wrote: > Signed-off-by: Ben Riggs > --- > cobbler/cobblerd.py | 6 +++--- > 1 files changed, 3 insertions(+), 3 deletions(-) > > diff --git a/cobbler/cobblerd.py b/cobbler/cobblerd.py > index d9a5ced..d2ef336 100644 > --- a/cobbler/cobblerd.py > +++ b/cobbler/cobblerd.py > @@ -71,11 +71,11 @@ def do_syslog(bootapi, settings, port, logger): > while 1: > data, addr = s.recvfrom(buf) > (ip, port) = addr > - name = bootapi.systems().find(ip_address = ip) > - if not data and name: > + system = bootapi.systems().find(ip_address = ip) > + if not data and system: > break > else: > - logfile = open("/var/log/cobbler/syslog/%s" % name, "a+") > + logfile = open("/var/log/cobbler/syslog/%s" % system.name, "a+") > t = time.localtime() > # write numeric time > seconds = str(time.mktime(t)) > Applied, thanks! From rigg0022 at umn.edu Wed Aug 15 21:02:54 2007 From: rigg0022 at umn.edu (Riggs, Ben) Date: Wed, 15 Aug 2007 16:02:54 -0500 Subject: [et-mgmt-tools] [PATCH] Changed collection.find() to take a dictionary argument instead of name, allowing searches by ip_address, mac_address, etc. In-Reply-To: <46C3675E.5020205@redhat.com> References: <11871933602647-git-send-email-rigg0022@umn.edu> <46C3675E.5020205@redhat.com> Message-ID: <46C369FE.6080502@umn.edu> Michael DeHaan wrote: > > As I've talked with Ben on IRC, this is really good, and I've applied a > minor variant of this. Minor? ;) > The change end users will see out of this so far is that "cobbler > status" will show the system name instead of the IP if it can figure it > out -- which is great. > Nothing else is effected. The logs files in /var/log/cobbler/syslog/ will also be listed by system name instead of ip as well. Ben From berrange at redhat.com Wed Aug 15 21:05:33 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Wed, 15 Aug 2007 22:05:33 +0100 Subject: [et-mgmt-tools] [patch] fix typo in virtinstall In-Reply-To: <200708150243.l7F2h1a5019450@fjmscan502.ms.jp.fujitsu.com> References: <200708150243.l7F2h1a5019450@fjmscan502.ms.jp.fujitsu.com> Message-ID: <20070815210533.GQ22579@redhat.com> On Wed, Aug 15, 2007 at 11:42:51AM +0900, Atsushi SAKAI wrote: > Since I detect typo in libvirt, > I try to apply ispell to virtinstall to make sure. > > Anyway Changelog seems many typos, but I am not touch it. > > Signed-off-by: Atsushi SAKAI Applied this - thanks for the patch. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Wed Aug 15 21:05:58 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Wed, 15 Aug 2007 22:05:58 +0100 Subject: [et-mgmt-tools] [patch] fix typo in virtmanager In-Reply-To: <200708150242.l7F2gtOS022324@fjmscan501.ms.jp.fujitsu.com> References: <200708150242.l7F2gtOS022324@fjmscan501.ms.jp.fujitsu.com> Message-ID: <20070815210558.GR22579@redhat.com> On Wed, Aug 15, 2007 at 11:42:48AM +0900, Atsushi SAKAI wrote: > Since I detect typo in libvirt and virtinstall, > To make sure, I try to apply ispell to virtmanager to make sure. > > Signed-off-by: Atsushi SAKAI Applied this - thanks for the patch Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From mazurekd at cpsc.ucalgary.ca Wed Aug 15 22:11:53 2007 From: mazurekd at cpsc.ucalgary.ca (Debbie Mazurek) Date: Wed, 15 Aug 2007 16:11:53 -0600 Subject: [et-mgmt-tools] Cobbler version In-Reply-To: <46C36618.3080509@redhat.com> References: <1187133631.18482.181.camel@caliente> <46C2E283.3060903@redhat.com> <1187203265.18482.203.camel@caliente> <46C36618.3080509@redhat.com> Message-ID: <1187215913.18482.208.camel@caliente> On Wed, 2007-08-15 at 16:46 -0400, Michael DeHaan wrote: > python-cheetah should also be in EPEL. So if you add the yum > repository and install that from yum, life should be in good. > If not, I'd be interested in what's not good -- that would be a bug in > Cheetah for EPEL. > > What Matt said also works. > > > Also, is there a minimum or specific version of yum and yum-utils that > > cobbler-0.6 needs? I didn't see anything in cobbler's spec file. > Not really. For RHEL the versions out of Centos are good -- and we'll > have those in EPEL soon enough. > cobbler-0.6 is in epel-testing now; I got it all sorted out. Thanks for the help, sorry for the noise. Deb -- d e b b i e m a z u r e k mazurekd at cpsc.ucalgary.ca system admin . department of computer science . university of calgary From karl at klxsystems.net Wed Aug 15 23:35:13 2007 From: karl at klxsystems.net (Karl R. Balsmeier) Date: Wed, 15 Aug 2007 16:35:13 -0700 Subject: [et-mgmt-tools] cobbler check fails: ImportError: No module named cobbler.cobbler Message-ID: <46C38DB1.1010002@klxsystems.net> Hi, Any advice on fixng this cobbler error or if there's a log I can check that will reveal the issue? I googled for it but no dice. [root at localhost ~]# cobbler check Traceback (most recent call last): File "/usr/bin/cobbler", line 18, in ? import cobbler.cobbler as app ImportError: No module named cobbler.cobbler This happens to be the second server I have tried this on, with same results, on completely different hardware, different networks, so we know this isn't a driver or network issue. Anyhow, any advice is much appreciated, i'd love to roll out a more robust provisioning system, looks like am super close once past this error. -karl From berrange at redhat.com Thu Aug 16 00:21:39 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 16 Aug 2007 01:21:39 +0100 Subject: [et-mgmt-tools] [FWD: [gtk-vnc-devel] ANNOUNCE: Release 0.1.0 of GTK-VNC] Message-ID: <20070816002139.GE20369@redhat.com> FYI, in the next couple of days I will be pushing the changes to virt-manager to use the GTK-VNC widget for the graphical console. The GTK-VNC package is currently undergoing review for inclusion into Fedora so binary packages should be available in the near future. Dan. ----- Forwarded message from "Daniel P. Berrange" ----- > Date: Thu, 16 Aug 2007 00:48:50 +0100 > From: "Daniel P. Berrange" > To: gtk-vnc-devel > Subject: [gtk-vnc-devel] ANNOUNCE: Release 0.1.0 of GTK-VNC > > The GTK-VNC development team is pleased to announce the first official > release of GTK-VNC, version 0.1.0. > > It can be downloaded from sourceforge at: > > http://downloads.sourceforge.net/gtk-vnc/gtk-vnc-0.1.0.tar.gz > > GTK-VNC is a VNC viewer widget for GTK. It is built using coroutines > allowing it to be completely asynchronous while remaining single threaded. > It supports RFB protocols 3.3 -> 3.8, and the VeNCrypt authentication > extension providing SSL/TLS encryption with x509 certificate authentication. > The core library is written in C and a binding for Python using PyGTK is > available. The networking layer supports connections over both IPv4 and > IPv6. Example code illustrates how to build a vncviewer replacement using > GTK-VNC in either C or Python. The code is licensed under the LGPL v2.1 > > Further information is available from the project website at > > http://gtk-vnc.sourceforge.net/ > > Regards, > Dan, on behalf of the GTK-VNC development team > -- > |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| > |=- Perl modules: http://search.cpan.org/~danberr/ -=| > |=- Projects: http://freshmeat.net/~danielpb/ -=| > |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > Gtk-vnc-devel mailing list > Gtk-vnc-devel at lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/gtk-vnc-devel > ----- End forwarded message ----- -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Thu Aug 16 01:41:28 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 16 Aug 2007 02:41:28 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1 Message-ID: <20070816014128.GA20064@redhat.com> Following on from the recent release of GTK-VNC, I'm pleased to be able to announce the first official release of "Virtual Machine Viewer" (virt-viewer) version 0.0.1. http://virt-manager.org/download/sources/virt-viewer/virt-viewer-0.0.1.tar.gz virt-viewer provides a lightweight, minimal client for connecting to the graphical display associated with a virtuaized guest OS. It currently supports connecting with the VNC protocol, and uses the GTK-VNC widget to provide the display. It uses libvirt to looking up VNC server configuration of the specified guest OS. It is intended as a replacement for any use of 'vncviewer'. Thus instead of having to first run 'virsh vncdisplay' to find the display info, and then feed it to vncviewer, you can simply run virt-viewer. Since it uses GTK-VNC, it also supports IPv6 and SSL/TLS encryption with x509 certificate auth via the VeNCrypt extension. To connect to the guest called "demo" running under Xen virt-viewer demo To connect to the guest with ID 7 running under QEMU virt-viewer --connect qemu:///system 7 To wait for the Xen guest with UUID 66ab33c0-6919-a3f7-e659-16c82d248521 to startup and then connect to its console virt-viewer --wait 66ab33c0-6919-a3f7-e659-16c82d248521 In the near future it will also be able to lookup details of remote hosts: For example using SSL/TLS: virt-viewer --connect xen:///example.org/ demo Or tunnelled over SSH: virt-viewer --connect xen+ssh:///example.org/ demo The virt-viewer app is provided under the GPL license. It depends on libvirt, GTK2, GTK-VNC, and libxml2 Virtual Machine Viewer is part of the 'Virt Manager' family of applications whose website is: http://virt-manager.org/ To contact the developers use the project mailing list: http://virt-manager.org/mailinglist.html Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From Roger.Goodman at js.pentagon.mil Thu Aug 16 15:51:07 2007 From: Roger.Goodman at js.pentagon.mil (Goodman, Roger L CTR JCS J8/SAMD) Date: Thu, 16 Aug 2007 11:51:07 -0400 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No module named cobbler.cobbler Message-ID: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> I'm new here, and I just tried setting up Cobbler and Koan yesterday, and I am getting the same error, but the missing file is Cheetah.Template. The traceback looks like this: File "/usr/bin/cobbler", line 18, in ? import cobbler.cobbler as app File "usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 18, in ? import api File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 17, in ? import config File "/usr/lib/python2.4/site-packages/cobbler/config.py", line 23, in ? import collection_distros as distros File "/usr/lib/python2.4/site-packages/cobbler/collection_distros.py", line 17, in ? import collection File "/usr/lib/python2.4/site-packages/cobbler/collection.py", line 22, in ? import action_litesync File "/usr/lib/python2.4/site-packages/cobbler/action_litesync.py", line 28, in ? import action_sync File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line 34, in ? from Cheetah.Template import Template ImportError: No module named Cheetah.Template I have checked the entire filesystem, and there is no file named Cheetah.Template. So, did something not get installed, or did I do something in the wrong order, or am I missing something else? Any help appreciated. Thanks, Roger [et-mgmt-tools] cobbler check fails: ImportError: No module named cobbler.cobbler * From: "Karl R. Balsmeier" * To: et-mgmt-tools redhat com * Subject: [et-mgmt-tools] cobbler check fails: ImportError: No module named cobbler.cobbler * Date: Wed, 15 Aug 2007 16:35:13 -0700 Hi, Any advice on fixng this cobbler error or if there's a log I can check that will reveal the issue? I googled for it but no dice. [root localhost ~]# cobbler check Traceback (most recent call last): File "/usr/bin/cobbler", line 18, in ? import cobbler.cobbler as app ImportError: No module named cobbler.cobbler This happens to be the second server I have tried this on, with same results, on completely different hardware, different networks, so we know this isn't a driver or network issue. Anyhow, any advice is much appreciated, i'd love to roll out a more robust provisioning system, looks like am super close once past this error. -karl From wright at imageworks.com Thu Aug 16 15:55:50 2007 From: wright at imageworks.com (Peter Wright) Date: Thu, 16 Aug 2007 08:55:50 -0700 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No modulenamed cobbler.cobbler In-Reply-To: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> References: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> Message-ID: <46C47386.7070800@imageworks.com> Goodman, Roger L CTR JCS J8/SAMD wrote: > > I'm new here, and I just tried setting up Cobbler and Koan yesterday, > and I am getting the same error, but the missing file is > Cheetah.Template. The traceback looks like this: > > File "/usr/bin/cobbler", line 18, in ? > import cobbler.cobbler as app > File "usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 18, in ? > import api > File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 17, in ? > import config > File "/usr/lib/python2.4/site-packages/cobbler/config.py", line 23, in ? > import collection_distros as distros > File "/usr/lib/python2.4/site-packages/cobbler/collection_distros.py", > line 17, in ? > import collection > File "/usr/lib/python2.4/site-packages/cobbler/collection.py", line > 22, in ? > import action_litesync > File "/usr/lib/python2.4/site-packages/cobbler/action_litesync.py", > line 28, in ? > import action_sync > File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line > 34, in ? > from Cheetah.Template import Template > ImportError: No module named Cheetah.Template > > I have checked the entire filesystem, and there is no file > named Cheetah.Template. So, did something not get installed, or did I > do something in the wrong order, or am I missing something else? Any > help appreciated. > Hi Roger, python-cheetah is a dependency of cobbler. please make sure you have this RPM installed (if you are running a RHEL5 system): python-cheetah-2.0-0.6.rc8.el5 This package is available via the EPEL yum repo (as is cobbler/koan itself): http://fedoraproject.org/wiki/EPEL -pete > Thanks, Roger > > [et-mgmt-tools] cobbler check fails: ImportError: No module named > cobbler.cobbler > > * From: "Karl R. Balsmeier" > * To: et-mgmt-tools redhat com > * Subject: [et-mgmt-tools] cobbler check fails: ImportError: No > module named cobbler.cobbler > * Date: Wed, 15 Aug 2007 16:35:13 -0700 > > Hi, > > Any advice on fixng this cobbler error or if there's a log I can check > that will reveal the issue? I googled for it but no dice. > > [root localhost ~]# cobbler check > Traceback (most recent call last): > File "/usr/bin/cobbler", line 18, in ? > import cobbler.cobbler as app > ImportError: No module named cobbler.cobbler > > This happens to be the second server I have tried this on, with same > results, on completely different hardware, different networks, so we > know this isn't a driver or network issue. > > Anyhow, any advice is much appreciated, i'd love to roll out a more > robust provisioning system, looks like am super close once past this > error. > > -karl > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From rigg0022 at umn.edu Thu Aug 16 16:29:58 2007 From: rigg0022 at umn.edu (Riggs, Ben) Date: Thu, 16 Aug 2007 11:29:58 -0500 Subject: [et-mgmt-tools] cobbler check fails: ImportError: No module named cobbler.cobbler In-Reply-To: <46C38DB1.1010002@klxsystems.net> References: <46C38DB1.1010002@klxsystems.net> Message-ID: <46C47B86.5060607@umn.edu> Karl, The error looks like the cobbler directory isn't showing up in your python path. The default install is in /usr/lib/python#.#/site-packages/cobbler/files.py, which is consistent with the default python path. Specifically, the import call is looking for cobbler/cobbler.py Ben Karl R. Balsmeier wrote: > Hi, > > Any advice on fixng this cobbler error or if there's a log I can check > that will reveal the issue? I googled for it but no dice. > > [root at localhost ~]# cobbler check > Traceback (most recent call last): > File "/usr/bin/cobbler", line 18, in ? > import cobbler.cobbler as app > ImportError: No module named cobbler.cobbler > > This happens to be the second server I have tried this on, with same > results, on completely different hardware, different networks, so we > know this isn't a driver or network issue. > > Anyhow, any advice is much appreciated, i'd love to roll out a more > robust provisioning system, looks like am super close once past this error. > > -karl > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From mdehaan at redhat.com Thu Aug 16 20:40:30 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 16 Aug 2007 16:40:30 -0400 Subject: [et-mgmt-tools] cobbler check fails: ImportError: No module named cobbler.cobbler In-Reply-To: <46C47B86.5060607@umn.edu> References: <46C38DB1.1010002@klxsystems.net> <46C47B86.5060607@umn.edu> Message-ID: <46C4B63E.60605@redhat.com> Riggs, Ben wrote: > Karl, > > The error looks like the cobbler directory isn't showing up in your > python path. The default install is in > /usr/lib/python#.#/site-packages/cobbler/files.py, which is consistent > with the default python path. Specifically, the import call is looking > for cobbler/cobbler.py > > Ben It sounds like Karl didn't install from the RPM. *DO* install from the RPM. I can't/won't support it otherwise, there are just too many things that the RPM does that are important. From mdehaan at redhat.com Thu Aug 16 22:00:31 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 16 Aug 2007 18:00:31 -0400 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No module named cobbler.cobbler In-Reply-To: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> References: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> Message-ID: <46C4C8FF.4080205@redhat.com> Goodman, Roger L CTR JCS J8/SAMD wrote: > I'm new here, and I just tried setting up Cobbler and Koan yesterday, and I am getting the same error, but the missing file is Cheetah.Template. The traceback looks like this: > > File "/usr/bin/cobbler", line 18, in ? > import cobbler.cobbler as app > File "usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 18, in ? > import api > File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 17, in ? > import config > File "/usr/lib/python2.4/site-packages/cobbler/config.py", line 23, in ? > import collection_distros as distros > File "/usr/lib/python2.4/site-packages/cobbler/collection_distros.py", line 17, in ? > import collection > File "/usr/lib/python2.4/site-packages/cobbler/collection.py", line 22, in ? > import action_litesync > File "/usr/lib/python2.4/site-packages/cobbler/action_litesync.py", line 28, in ? > import action_sync > File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line 34, in ? > from Cheetah.Template import Template > ImportError: No module named Cheetah.Template I'm guessing this is also a case of using RPM with "--nodeps" or "--force" -- which is a bad thing to do. If I'm wrong, I'd be interested from what repository this came from or what your install process was. You possibly force installed the python-cheetah from another distro rather than one that was built for your distro and version of Python? --Michael From berrange at redhat.com Thu Aug 16 22:44:20 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 16 Aug 2007 23:44:20 +0100 Subject: [et-mgmt-tools] [PATCH] check if the specified disk using virt-clone is already used in another guests. In-Reply-To: <200707300954.AFB05251.9EKJ60G9@aa.jp.fujitsu.com> References: <200707300954.AFB05251.9EKJ60G9@aa.jp.fujitsu.com> Message-ID: <20070816224420.GB21661@redhat.com> On Mon, Jul 30, 2007 at 09:54:01AM +0900, S.Sakamoto wrote: > Hi > > Current virt-clone does not confirm if the specified disk is already > used in another guests or not. So, I create the patch to check if the > specified disk using virt-clone is already used in another guests. Sorry for the delay in this - it is applied now. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From martin.minka at gmail.com Fri Aug 17 08:16:44 2007 From: martin.minka at gmail.com (Martin Minka) Date: Fri, 17 Aug 2007 10:16:44 +0200 Subject: [et-mgmt-tools] TypeError: sort() takes no keyword arguments Message-ID: <46C5596C.3@gmail.com> hello, I installed Cobbler on CentOS 4.5, but "cobbler check" (but also "cobbler") returns error: "TypeError: sort() takes no keyword arguments". # cobbler Traceback (most recent call last): File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 719, in main BootCLI(sys.argv).run() File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 42, in __init__ self.api = api.BootAPI() File "/usr/lib/python2.3/site-packages/cobbler/api.py", line 40, in __init__ self.deserialize() File "/usr/lib/python2.3/site-packages/cobbler/api.py", line 170, in deserialize return self._config.deserialize() File "/usr/lib/python2.3/site-packages/cobbler/config.py", line 169, in deserialize if not serializer.deserialize(x,topological=True): File "/usr/lib/python2.3/site-packages/cobbler/serializer.py", line 79, in deserialize datastruct.sort(cmp=__depth_cmp) TypeError: sort() takes no keyword arguments zsh: exit 1 cobbler Please help me to fix this problem. Best regards, Martin Minka From mdehaan at redhat.com Fri Aug 17 12:57:49 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 17 Aug 2007 08:57:49 -0400 Subject: [et-mgmt-tools] TypeError: sort() takes no keyword arguments In-Reply-To: <46C5596C.3@gmail.com> References: <46C5596C.3@gmail.com> Message-ID: <46C59B4D.9050807@redhat.com> Martin Minka wrote: > hello, > I installed Cobbler on CentOS 4.5, but "cobbler check" (but also > "cobbler") returns error: "TypeError: sort() takes no keyword arguments". > > # cobbler > Traceback (most recent call last): > File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 719, > in main > BootCLI(sys.argv).run() > File "/usr/lib/python2.3/site-packages/cobbler/cobbler.py", line 42, > in __init__ > self.api = api.BootAPI() > File "/usr/lib/python2.3/site-packages/cobbler/api.py", line 40, in > __init__ > self.deserialize() > File "/usr/lib/python2.3/site-packages/cobbler/api.py", line 170, in > deserialize > return self._config.deserialize() > File "/usr/lib/python2.3/site-packages/cobbler/config.py", line 169, > in deserialize > if not serializer.deserialize(x,topological=True): > File "/usr/lib/python2.3/site-packages/cobbler/serializer.py", line > 79, in deserialize > datastruct.sort(cmp=__depth_cmp) > TypeError: sort() takes no keyword arguments > zsh: exit 1 cobbler > > Please help me to fix this problem. > > Best regards, > Martin Minka > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools If I recall correctly, this appeared in one of the 0.5.x testing builds (possibly 0.5.1). 0.6.0 is out now, please upgrade to that and this will go away. You can add EPEL as a repository if you like and just use yum to update cobbler. http://download.fedora.redhat.com/pub/epel/testing/4/i386/ From kraxel at redhat.com Fri Aug 17 14:10:52 2007 From: kraxel at redhat.com (Gerd Hoffmann) Date: Fri, 17 Aug 2007 16:10:52 +0200 Subject: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1 In-Reply-To: <20070816014128.GA20064@redhat.com> References: <20070816014128.GA20064@redhat.com> Message-ID: <46C5AC6C.30206@redhat.com> Hi, > To connect to the guest with ID 7 running under QEMU > > virt-viewer --connect qemu:///system 7 Can we *please* think about a standardized environment variable to set the default for --connect and which all tools are looking at? virsh has VIRSH_DEFAULT_CONNECT_URI for that. I'd like to have *one* variable used by *all* libvirt-based tools, virt-viewer, virt-manager, virsh, ... We could simply reuse the virsh variable, or we could pick some other name without virsh in there, preferable also shorter and thus easier to remember. VIRT_URI maybe? cheers, Gerd From berrange at redhat.com Fri Aug 17 14:13:13 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 17 Aug 2007 15:13:13 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1 In-Reply-To: <46C5AC6C.30206@redhat.com> References: <20070816014128.GA20064@redhat.com> <46C5AC6C.30206@redhat.com> Message-ID: <20070817141313.GA3792@redhat.com> On Fri, Aug 17, 2007 at 04:10:52PM +0200, Gerd Hoffmann wrote: > Hi, > > > To connect to the guest with ID 7 running under QEMU > > > > virt-viewer --connect qemu:///system 7 > > Can we *please* think about a standardized environment variable to set > the default for --connect and which all tools are looking at? Yes, this would be useful. > virsh has VIRSH_DEFAULT_CONNECT_URI for that. > > I'd like to have *one* variable used by *all* libvirt-based tools, > virt-viewer, virt-manager, virsh, ... We could simply reuse the virsh > variable, or we could pick some other name without virsh in there, > preferable also shorter and thus easier to remember. VIRT_URI maybe? We should probably apply the change to libvirt, so that if 'NULL' is used for the connect URI for virConnectOpen, and VIRT_CONNECT_URI is set, then NULL gets substituted with the value of the env variable. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From Roger.Goodman at js.pentagon.mil Fri Aug 17 15:00:55 2007 From: Roger.Goodman at js.pentagon.mil (Goodman, Roger L CTR JCS J8/SAMD) Date: Fri, 17 Aug 2007 11:00:55 -0400 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No modulenamed cobbler.cobbler In-Reply-To: <46C4C8FF.4080205@redhat.com> References: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> <46C4C8FF.4080205@redhat.com> Message-ID: <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> Michael, Regarding your questions, I downloaded a file called cobbler-0.6.0.tar.gz from the website http://cobbler.et.redhat.com/ . I extracted it, ran some kind of a build command, to end up with cobbler-0.6.0.linux-x86_64.tar.gz, which I untarred from the root directory. At no point did I see a requirement/prerequisite of Cheetah.Template (not on the website, not in the .tar.gz file, not during the build phase). When I ran the "cobbler check" command, and I got the failure, is when I discovered that something was missing. I have since then loaded Cheetah templated from an RPM file on the RHEL 5 disks, and cobbler has fewer complaints now. Still having problems getting a tftpd to stay running. The tftp-server RPM is installed, but when I start tftp, it immediately exits. When I first looked, I didn't see an RPM for the x86_64 architecture, which is why I went with the source file. So, I did not do a --force RPM install of cobbler, as I didn't have an RPM file, and the Cheetah.Template file came from the RHEL5 disks (disk 4, as I recall). I did not have to do a --force on it, either, but I did install it _after_ cobbler had failed. My suggestions: 1. On the webpage, mention any/all prerequisites, and maybe even how to test if they are installed. 2. On the download page, link directly to the file wanted (x86, x86_64, etc.) directly, rather than dropping people into the Fedora EPEL 5 (x86_64) first page. I did not catch that I could jump to a letter at the top of the page, and thought that the page showing was all that was available. I did look in the links on the left, but didn't see it anywhere there, either. Thanks, Roger Goodman -----Original Message----- From: et-mgmt-tools-bounces at redhat.com [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan Sent: Thursday, August 16, 2007 6:01 PM To: Fedora/Linux Management Tools Subject: Re: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No modulenamed cobbler.cobbler Goodman, Roger L CTR JCS J8/SAMD wrote: > I'm new here, and I just tried setting up Cobbler and Koan yesterday, and I am getting the same error, but the missing file is Cheetah.Template. The traceback looks like this: > > File "/usr/bin/cobbler", line 18, in ? > import cobbler.cobbler as app > File "usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 18, in ? > import api > File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 17, in ? > import config > File "/usr/lib/python2.4/site-packages/cobbler/config.py", line 23, in ? > import collection_distros as distros > File "/usr/lib/python2.4/site-packages/cobbler/collection_distros.py", line 17, in ? > import collection > File "/usr/lib/python2.4/site-packages/cobbler/collection.py", line 22, in ? > import action_litesync > File "/usr/lib/python2.4/site-packages/cobbler/action_litesync.py", line 28, in ? > import action_sync > File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line 34, in ? > from Cheetah.Template import Template > ImportError: No module named Cheetah.Template I'm guessing this is also a case of using RPM with "--nodeps" or "--force" -- which is a bad thing to do. If I'm wrong, I'd be interested from what repository this came from or what your install process was. You possibly force installed the python-cheetah from another distro rather than one that was built for your distro and version of Python? --Michael _______________________________________________ et-mgmt-tools mailing list et-mgmt-tools at redhat.com https://www.redhat.com/mailman/listinfo/et-mgmt-tools From mdehaan at redhat.com Fri Aug 17 15:14:53 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 17 Aug 2007 11:14:53 -0400 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No modulenamed cobbler.cobbler In-Reply-To: <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> References: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> <46C4C8FF.4080205@redhat.com> <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> Message-ID: <46C5BB6D.9080807@redhat.com> Goodman, Roger L CTR JCS J8/SAMD wrote: > Michael, > Regarding your questions, I downloaded a file called cobbler-0.6.0.tar.gz from the website http://cobbler.et.redhat.com/ . I extracted it, ran some kind of a build command, to end up with cobbler-0.6.0.linux-x86_64.tar.gz, which I untarred from the root directory. At no point did I see a requirement/prerequisite of Cheetah.Template (not on the website, not in the .tar.gz file, not during the build phase). When I ran the "cobbler check" command, and I got the failure, is when I discovered that something was missing. > RHEL and Fedora express package dependencies in RPMs, so if you don't build from the RPMs, you will get burned. If you look at http://cobbler.et.redhat.com/download/ there are source RPMs there. You can make an installable RPM from a src RPM using rpmbuild. Or you could just add EPEL 5 as a repository and all of this is done for you with your package install tool. If you build from Cobbler's source, you type "make" and make outputs an rpm. --Michael From mdehaan at redhat.com Fri Aug 17 15:16:08 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 17 Aug 2007 11:16:08 -0400 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: No modulenamed cobbler.cobbler In-Reply-To: <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> References: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local> <46C4C8FF.4080205@redhat.com> <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> Message-ID: <46C5BBB8.7000100@redhat.com> > I have since then loaded Cheetah templated from an RPM file on the RHEL 5 disks, and cobbler has fewer complaints now. Still having problems getting a tftpd to stay running. The tftp-server RPM is installed, but when I start tftp, it immediately exits. When I first looked, I didn't see an RPM for the x86_64 architecture, which is why I went with the source file. tftp is launched by xinetd -- it is only started when a connection is needed, as is true with everything under /etc/xinetd.d/ From veillard at redhat.com Fri Aug 17 15:40:35 2007 From: veillard at redhat.com (Daniel Veillard) Date: Fri, 17 Aug 2007 11:40:35 -0400 Subject: [Libvir] Re: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1 In-Reply-To: <20070817141313.GA3792@redhat.com> References: <20070816014128.GA20064@redhat.com> <46C5AC6C.30206@redhat.com> <20070817141313.GA3792@redhat.com> Message-ID: <20070817154035.GE24223@redhat.com> On Fri, Aug 17, 2007 at 03:13:13PM +0100, Daniel P. Berrange wrote: > On Fri, Aug 17, 2007 at 04:10:52PM +0200, Gerd Hoffmann wrote: > > Hi, > > > > > To connect to the guest with ID 7 running under QEMU > > > > > > virt-viewer --connect qemu:///system 7 > > > > Can we *please* think about a standardized environment variable to set > > the default for --connect and which all tools are looking at? > > Yes, this would be useful. > > > virsh has VIRSH_DEFAULT_CONNECT_URI for that. > > > > I'd like to have *one* variable used by *all* libvirt-based tools, > > virt-viewer, virt-manager, virsh, ... We could simply reuse the virsh > > variable, or we could pick some other name without virsh in there, > > preferable also shorter and thus easier to remember. VIRT_URI maybe? > > We should probably apply the change to libvirt, so that if 'NULL' is used > for the connect URI for virConnectOpen, and VIRT_CONNECT_URI is set, then > NULL gets substituted with the value of the env variable. Yup that sounds reasonnable, we should update the documentation too in that case, but sounds fine to me, Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard at redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/ From kraxel at redhat.com Fri Aug 17 16:22:59 2007 From: kraxel at redhat.com (Gerd Hoffmann) Date: Fri, 17 Aug 2007 18:22:59 +0200 Subject: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1 In-Reply-To: <20070817141313.GA3792@redhat.com> References: <20070816014128.GA20064@redhat.com> <46C5AC6C.30206@redhat.com> <20070817141313.GA3792@redhat.com> Message-ID: <46C5CB63.8070507@redhat.com> Hi, > We should probably apply the change to libvirt, so that if 'NULL' is used > for the connect URI for virConnectOpen, and VIRT_CONNECT_URI is set, then > NULL gets substituted with the value of the env variable. Sounds good. thanks, Gerd From wright at imageworks.com Fri Aug 17 16:57:17 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 17 Aug 2007 09:57:17 -0700 Subject: [et-mgmt-tools] Re: Cobbler Check fails: ImportError: Nomodulenamed cobbler.cobbler In-Reply-To: <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> References: <8CE494D0333AF84585D2020474DEEB11BB5DA8@PJSINUEXCH03.jsinu.jsu.local><46C4C8FF.4080205@redhat.com> <8CE494D0333AF84585D2020474DEEB11BB5DAD@PJSINUEXCH03.jsinu.jsu.local> Message-ID: <46C5D36D.6050300@imageworks.com> Goodman, Roger L CTR JCS J8/SAMD wrote: > > Michael, > Regarding your questions, I downloaded a file called > cobbler-0.6.0.tar.gz from the website http://cobbler.et.redhat.com/ . > I extracted it, ran some kind of a build command, to end up with > cobbler-0.6.0.linux-x86_64.tar.gz, which I untarred from the root > directory. At no point did I see a requirement/prerequisite of > Cheetah.Template (not on the website, not in the .tar.gz file, not > during the build phase). When I ran the "cobbler check" command, and > I got the failure, is when I discovered that something was missing. > I have since then loaded Cheetah templated from an RPM file on > the RHEL 5 disks, and cobbler has fewer complaints now. Still having > problems getting a tftpd to stay running. The tftp-server RPM is > installed, but when I start tftp, it immediately exits. When I first > looked, I didn't see an RPM for the x86_64 architecture, which is why > I went with the source file. > Cobbler and Koan can safely be built as "noarch" rpms, as they are written in python and I do not think they require any platform specific libs. So, you should be able to install the same RPM on i386 as well as x86_64 (at least I am ;). In fact, I think if run "make" in the cobbler and koan source directories "noarch" rpm's get built by default and end up in the "rpm-build" directory. > So, I did not do a --force RPM install of cobbler, as I didn't > have an RPM file, and the Cheetah.Template file came from the RHEL5 > disks (disk 4, as I recall). I did not have to do a --force on it, > either, but I did install it _after_ cobbler had failed. > My suggestions: > 1. On the webpage, mention any/all prerequisites, and maybe even how > to test if they are installed. > Yea, I agree it could be a little more explicit. I got things up and running by using the HOWTO build a Cobbler Server from the wiki: https://hosted.fedoraproject.org/projects/cobbler/wiki/HowToCobblerServer If I had more time I'd submit a patch for the documentation section on the cobbler website.... HTH, -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From mdehaan at redhat.com Fri Aug 17 17:53:58 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 17 Aug 2007 13:53:58 -0400 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi Message-ID: <46C5E0B6.1030206@redhat.com> I've just implemented some changes in cobbler (cobblerd) and koan that allow for automatic remote service discovery using Avahi. This allows for building of tools that rely on koan to be written in a way that doesn't need to be aware of where the cobbler server is, which is particularly useful in multi-site operations. The ultimate usage of this is the koan live CD or USB disk. Now, with auto-discovery enabled, we can build a truly universal installer image. One image can install RHEL-4, F-7, whichever -- regardless of your location. The live installer CD will boot in any (supported) bare metal machine, check the machine's MAC address, find the cobbler server automatically, and install whatever profile is mapped to the MAC address inside of cobbler. None of this requires a PXE setup, either -- so it's easy to set up "bonus" install servers even if your organization already has a PXE infrastructure. To use this, install cobbler and koan from source, and also install the avahi-tools package on both ends. That's all you need to do. This will be included in the 0.6.1 release, when that happens. How does this work? Cobblerd will detect the presence of avahi-tools when it starts up, and will publish the service. With koan, we specify --server=DISCOVER to tell koan it needs to probe for a cobbler server. For instance, the following command works regardless of where you are: koan --server=DISCOVER --list-profiles --Michael From mdehaan at redhat.com Fri Aug 17 18:03:37 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 17 Aug 2007 14:03:37 -0400 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <46C5E0B6.1030206@redhat.com> References: <46C5E0B6.1030206@redhat.com> Message-ID: <46C5E2F9.70002@redhat.com> Michael DeHaan wrote: > I've just implemented some changes in cobbler (cobblerd) and koan that > allow for automatic remote service discovery using Avahi. > This allows for building of tools that rely on koan to be written in a > way that doesn't need to be aware of where the cobbler server is, > which is particularly useful in multi-site operations. > > The ultimate usage of this is the koan live CD or USB disk. Now, > with auto-discovery enabled, we can build a truly universal installer > image. One image can install RHEL-4, F-7, whichever -- regardless of > your location. The live installer CD will boot in any (supported) bare > metal machine, check the machine's MAC address, find the cobbler > server automatically, and install whatever profile is mapped to the MAC > address inside of cobbler. None of this requires a PXE setup, > either -- so it's easy to set up "bonus" install servers even if your > organization > already has a PXE infrastructure. > > To use this, install cobbler and koan from source, and also install > the avahi-tools package on both ends. That's all you need to do. > This will be included in the 0.6.1 release, when that happens. > > How does this work? > > Cobblerd will detect the presence of avahi-tools when it starts up, > and will publish the service. > > With koan, we specify --server=DISCOVER to tell koan it needs to > probe for a cobbler server. > > For instance, the following command works regardless of where you are: > koan --server=DISCOVER --list-profiles > > --Michael > > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools As someone on IRC (#cobbler on freenode.net) asked, avahi-tools is currently available in Fedora and RHEL5. It may be installable on RHEL4. Either way, the tools will continue to work without it... I detect the presense/absense of the capability at run time ... they just won't be able to autodiscover/publish if that ability is not there. From mdehaan at redhat.com Fri Aug 17 18:17:46 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 17 Aug 2007 14:17:46 -0400 Subject: [et-mgmt-tools] A note about EPEL Message-ID: <46C5E64A.20405@redhat.com> EPEL is a bit slow on doing pushes lately. To ensure you get the latest bits, add EPEL /testing/ as a repository, not the main EPEL repos. http://download.fedora.redhat.com/pub/epel/testing/4/i386/ For instance, koan in EPEL is 0.4.x, not 0.6. I'll update the web page to reflect this. --Michael From berrange at redhat.com Fri Aug 17 21:45:44 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 17 Aug 2007 22:45:44 +0100 Subject: [et-mgmt-tools] [FWD: [gtk-vnc-devel] ANNOUNCE: Release 0.1.0 of GTK-VNC] In-Reply-To: <20070816002139.GE20369@redhat.com> References: <20070816002139.GE20369@redhat.com> Message-ID: <20070817214544.GF707@redhat.com> On Thu, Aug 16, 2007 at 01:21:39AM +0100, Daniel P. Berrange wrote: > FYI, in the next couple of days I will be pushing the changes to virt-manager > to use the GTK-VNC widget for the graphical console. The GTK-VNC package is > currently undergoing review for inclusion into Fedora so binary packages > should be available in the near future. The 'gtk-vnc' package is now in Fedora rawhide, and shortly Fedora 7. The latest virt-manager code in the HG repository uses GTK-VNC for its graphical display. As a reminder for anyone not using Fedora, the GTK-VNC code is available from the project site http://gtk-vnc.sourceforge.net/ and should be portable to other POSIX OS Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From rmk.list at gmail.com Sun Aug 19 20:18:17 2007 From: rmk.list at gmail.com (Radhesh Kamath) Date: Sun, 19 Aug 2007 13:18:17 -0700 Subject: [et-mgmt-tools] Need help on changing DomU kernel Message-ID: <11a2bd2d0708191318n6c2cc657me28794ecb7382b2@mail.gmail.com> Hi, I needed help with changing the kernel on a DomU using virt-manager. I had an initial install of a Fedora core 7 domU. Now I have added some patches of my own to the standard domU kernel and I want to use it as the the new DomU. How can I do it using virt-manager? Thanks, Radhesh -------------- next part -------------- An HTML attachment was scrubbed... URL: From rjones at redhat.com Mon Aug 20 10:54:50 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Mon, 20 Aug 2007 11:54:50 +0100 Subject: [et-mgmt-tools] Re: [Libvir] Set memory (decrease) from Virt Manager In-Reply-To: <46C77B46.6040401@Sun.COM> References: <46C77B46.6040401@Sun.COM> Message-ID: <46C972FA.4050807@redhat.com> [copied to et-mgmt-tools - please followup there] Bruce Evans wrote: > Hi, > > I am trying to lower the amount of memory for a Domain > using the Virt manager. > > I am denied from doing this if I want to decrease the memory > below 50MB, even though I can do it > from virsh. There appears to be a low end limit of > 50MB in the Virt Manager. > > If the current memory allocation is greater than 50MB, I can > decrease the amount from Virt Manager, but can't go under 50MB. > > But if I use virsh, I can decrease the memory to, say, 13000KB > and it works. The Virt Manger will show this new amount but will > only enable the 'increase' icon in the 'Change Allocation option box. > The 'decrease' icon is disabled. > > Is this a bug in Virt Manager? If not, why can't the memory > be decreased below 50MB? Yes, it seems to be the case that the adjustment in virt-manager is limited to min of 50MB (and to a max of 32000MB which is probably also a bug): In src/vmm-details.glade: 50 50 32000 5 10 10 50 50 32000 5 10 10 Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From rjones at redhat.com Mon Aug 20 11:53:49 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Mon, 20 Aug 2007 12:53:49 +0100 Subject: [Libvir] Re: [et-mgmt-tools] ANNOUNCE: virt-viewer release 0.0.1 In-Reply-To: <20070817141313.GA3792@redhat.com> References: <20070816014128.GA20064@redhat.com> <46C5AC6C.30206@redhat.com> <20070817141313.GA3792@redhat.com> Message-ID: <46C980CD.5040806@redhat.com> Daniel P. Berrange wrote: > On Fri, Aug 17, 2007 at 04:10:52PM +0200, Gerd Hoffmann wrote: >> Hi, >> >>> To connect to the guest with ID 7 running under QEMU >>> >>> virt-viewer --connect qemu:///system 7 >> Can we *please* think about a standardized environment variable to set >> the default for --connect and which all tools are looking at? > > Yes, this would be useful. > >> virsh has VIRSH_DEFAULT_CONNECT_URI for that. >> >> I'd like to have *one* variable used by *all* libvirt-based tools, >> virt-viewer, virt-manager, virsh, ... We could simply reuse the virsh >> variable, or we could pick some other name without virsh in there, >> preferable also shorter and thus easier to remember. VIRT_URI maybe? > > We should probably apply the change to libvirt, so that if 'NULL' is used > for the connect URI for virConnectOpen, and VIRT_CONNECT_URI is set, then > NULL gets substituted with the value of the env variable. Very good idea. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From wright at imageworks.com Mon Aug 20 16:54:24 2007 From: wright at imageworks.com (Peter Wright) Date: Mon, 20 Aug 2007 09:54:24 -0700 Subject: [et-mgmt-tools] Need help on changing DomU kernel In-Reply-To: <11a2bd2d0708191318n6c2cc657me28794ecb7382b2@mail.gmail.com> References: <11a2bd2d0708191318n6c2cc657me28794ecb7382b2@mail.gmail.com> Message-ID: <46C9C740.8060307@imageworks.com> Radhesh Kamath wrote: > Hi, > > I needed help with changing the kernel on a DomU using virt-manager. > > I had an initial install of a Fedora core 7 domU. Now I have added > some patches of my own to the standard domU kernel and I want to use > it as the the new DomU. How can I do it using virt-manager? > Have you already compiled your locally patched kernel and tested it? If you have done that bit, which is the hard part in my opinion, then you can create a new distro using this kernel and initrd pretty easily in cobbler. check out the man page for cobbler (man cobbler) and read the section on creating a new distro. if you are having specific problems let us know! -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From mdehaan at redhat.com Mon Aug 20 17:01:08 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 20 Aug 2007 13:01:08 -0400 Subject: [et-mgmt-tools] Need help on changing DomU kernel In-Reply-To: <46C9C740.8060307@imageworks.com> References: <11a2bd2d0708191318n6c2cc657me28794ecb7382b2@mail.gmail.com> <46C9C740.8060307@imageworks.com> Message-ID: <46C9C8D4.402@redhat.com> Peter Wright wrote: > Radhesh Kamath wrote: >> Hi, >> >> I needed help with changing the kernel on a DomU using virt-manager. >> >> I had an initial install of a Fedora core 7 domU. Now I have added >> some patches of my own to the standard domU kernel and I want to use >> it as the the new DomU. How can I do it using virt-manager? >> > Have you already compiled your locally patched kernel and tested it? > If you have done that bit, which is the hard part in my opinion, then > you can create a new distro using this kernel and initrd pretty easily > in cobbler. check out the man page for cobbler (man cobbler) and read > the section on creating a new distro. if you are having specific > problems let us know! > > -pete > I think he's asking about how to change the kernel of an existing virtual machine without reinstalling it. After installation is complete (i.e. on subsequent boots), Xen gets the kernel/initrd information from the bootloader, so all you should really need to do is upgrade the kernel from inside the virtual machine as if it were any other kernel -- and restart the machine. The next time it comes up, Xen should find the new kernel. --Michael From dlutter at redhat.com Mon Aug 20 19:09:24 2007 From: dlutter at redhat.com (David Lutterkort) Date: Mon, 20 Aug 2007 12:09:24 -0700 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <46C5E0B6.1030206@redhat.com> References: <46C5E0B6.1030206@redhat.com> Message-ID: <1187636964.6594.10.camel@localhost.localdomain> On Fri, 2007-08-17 at 13:53 -0400, Michael DeHaan wrote: > How does this work? > > Cobblerd will detect the presence of avahi-tools when it starts up, > and will publish the service. > > With koan, we specify --server=DISCOVER to tell koan it needs to > probe for a cobbler server. > > For instance, the following command works regardless of where you are: > > koan --server=DISCOVER --list-profiles Cool .. how about the following behavior: when koan is started without --server, it first tries to look up a host named 'cobbler' and uses that as the cobblerd; if no such host exists, try to discover using avahi; if that fails tell the user they are dumb and need to give na explicit server with --server ? David From wright at imageworks.com Mon Aug 20 19:12:39 2007 From: wright at imageworks.com (Peter Wright) Date: Mon, 20 Aug 2007 12:12:39 -0700 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <1187636964.6594.10.camel@localhost.localdomain> References: <46C5E0B6.1030206@redhat.com> <1187636964.6594.10.camel@localhost.localdomain> Message-ID: <46C9E7A7.7030308@imageworks.com> David Lutterkort wrote: > > On Fri, 2007-08-17 at 13:53 -0400, Michael DeHaan wrote: > > How does this work? > > > > Cobblerd will detect the presence of avahi-tools when it starts up, > > and will publish the service. > > > > With koan, we specify --server=DISCOVER to tell koan it needs to > > probe for a cobbler server. > > > > For instance, the following command works regardless of where you are: > > > > koan --server=DISCOVER --list-profiles > > Cool .. how about the following behavior: when koan is started without > --server, it first tries to look up a host named 'cobbler' and uses that > as the cobblerd; if no such host exists, try to discover using avahi; if > that fails tell the user they are dumb and need to give na explicit > server with --server ? > please do not put any hard coded hostnames in the koan code base. two immediate problems crop to mind: 1) a host named cobbler already exists and does not run cobblerd 2) the system becomes coupled to a working DNS system -p -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From mdehaan at redhat.com Mon Aug 20 19:37:29 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 20 Aug 2007 15:37:29 -0400 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <46C9E7A7.7030308@imageworks.com> References: <46C5E0B6.1030206@redhat.com> <1187636964.6594.10.camel@localhost.localdomain> <46C9E7A7.7030308@imageworks.com> Message-ID: <46C9ED79.7090709@redhat.com> Peter Wright wrote: > David Lutterkort wrote: >> >> On Fri, 2007-08-17 at 13:53 -0400, Michael DeHaan wrote: >> > How does this work? >> > >> > Cobblerd will detect the presence of avahi-tools when it starts >> up, >> > and will publish the service. >> > >> > With koan, we specify --server=DISCOVER to tell koan it needs to >> > probe for a cobbler server. >> > >> > For instance, the following command works regardless of where you are: >> > > koan --server=DISCOVER --list-profiles >> >> Cool .. how about the following behavior: when koan is started without >> --server, it first tries to look up a host named 'cobbler' and uses that >> as the cobblerd; if no such host exists, try to discover using avahi; if >> that fails tell the user they are dumb and need to give na explicit >> server with --server ? >> > please do not put any hard coded hostnames in the koan code base. two > immediate problems crop to mind: > > 1) a host named cobbler already exists and does not run cobblerd > 2) the system becomes coupled to a working DNS system > > > -p > > > Peter, I'm not so sure I agree a few defaults is a bad thing, since it's pretty easy to test for cobblerd's presence (xmlrpc responding to certain method on certain port) and we wouldn't have to engage this check unless --server was omitted. David's suggestion actually works pretty well to help out those running EL4 boot servers (where there is no Avahi) -- which is probably the most likely platform for a boot server at this point. Rest assured I have ruled out "nmap" as a discovery method :) And obviously if we specify --server (as was mandatory before) we can skip all of those checks (so you don't need DNS in that case). --Michael From dlutter at redhat.com Mon Aug 20 19:38:45 2007 From: dlutter at redhat.com (David Lutterkort) Date: Mon, 20 Aug 2007 12:38:45 -0700 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <46C9E7A7.7030308@imageworks.com> References: <46C5E0B6.1030206@redhat.com> <1187636964.6594.10.camel@localhost.localdomain> <46C9E7A7.7030308@imageworks.com> Message-ID: <1187638725.6594.24.camel@localhost.localdomain> On Mon, 2007-08-20 at 12:12 -0700, Peter Wright wrote: > please do not put any hard coded hostnames in the koan code base. It's generally a good practice to have CNAME's for all your services, so that clients don't need to change if you move a service to another host. > two immediate problems crop to mind: > > 1) a host named cobbler already exists and does not run cobblerd > 2) the system becomes coupled to a working DNS system True, though in both cases, people can still use '--server' to avoid either of these. In other words: for people that continue to use koan the way it is today, nothing changes. For those who want to rely on their infrastructure to provide some sort of name lookup, omitting '--server' from the command line will jsut do that. David From wright at imageworks.com Mon Aug 20 21:28:14 2007 From: wright at imageworks.com (Peter Wright) Date: Mon, 20 Aug 2007 14:28:14 -0700 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <46C9ED79.7090709@redhat.com> References: <46C5E0B6.1030206@redhat.com> <1187636964.6594.10.camel@localhost.localdomain><46C9E7A7.7030308@imageworks.com> <46C9ED79.7090709@redhat.com> Message-ID: <46CA076E.8070300@imageworks.com> Michael DeHaan wrote: > > Peter Wright wrote: > > David Lutterkort wrote: > >> > >> On Fri, 2007-08-17 at 13:53 -0400, Michael DeHaan wrote: > >> > How does this work? > >> > > >> > Cobblerd will detect the presence of avahi-tools when it starts > >> up, > >> > and will publish the service. > >> > > >> > With koan, we specify --server=DISCOVER to tell koan it needs to > >> > probe for a cobbler server. > >> > > >> > For instance, the following command works regardless of where you > are: > >> > > koan --server=DISCOVER --list-profiles > >> > >> Cool .. how about the following behavior: when koan is started without > >> --server, it first tries to look up a host named 'cobbler' and uses > that > >> as the cobblerd; if no such host exists, try to discover using > avahi; if > >> that fails tell the user they are dumb and need to give na explicit > >> server with --server ? > >> > > please do not put any hard coded hostnames in the koan code base. two > > immediate problems crop to mind: > > > > 1) a host named cobbler already exists and does not run cobblerd > > 2) the system becomes coupled to a working DNS system > > > > > > -p > > > > > > > > > Peter, > > I'm not so sure I agree a few defaults is a bad thing, since it's pretty > easy to test for cobblerd's presence (xmlrpc responding to certain > method on certain port) and we wouldn't have to engage this check unless > --server was omitted. David's suggestion actually works pretty well to > help out those running EL4 boot servers (where there is no Avahi) -- > which is probably the most likely platform for a boot server at this > point. Rest assured I have ruled out "nmap" as a discovery method :) > > And obviously if we specify --server (as was mandatory before) we can > skip all of those checks (so you don't need DNS in that case). > > --Michael > > OK, I guess I understand the reasoning here - although I get a little worried when hostnames are hard coded into applications like this, I rarely like outside forces determining naming schemes :) I'd feel more comfortable if koan looked for a COBBLER_SERVER environment variable (or something similar) to get the name of the master cobbler node on ones network, or subnet. Maybe this variable could be taken as an override of the default "cobbler" hostname. Assuming that koan is going to be called via shell scripts or some other programmatic method in most environments this would also help admins who may need to rename machines or test code against development/staging cobbler servers. maybe i'm off base in this assumption though. I've got my anal retentive admin hat on today so take this with a grain of salt ;) -p -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From mdehaan at redhat.com Tue Aug 21 16:59:48 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 21 Aug 2007 12:59:48 -0400 Subject: [et-mgmt-tools] Service discovery for cobbler and koan with Avahi In-Reply-To: <46CA076E.8070300@imageworks.com> References: <46C5E0B6.1030206@redhat.com> <1187636964.6594.10.camel@localhost.localdomain><46C9E7A7.7030308@imageworks.com> <46C9ED79.7090709@redhat.com> <46CA076E.8070300@imageworks.com> Message-ID: <46CB1A04.1090004@redhat.com> Peter Wright wrote: > Michael DeHaan wrote: >> > > > I'd feel more comfortable if koan looked for a COBBLER_SERVER > environment variable (or something similar) to get the name of the > master cobbler node on ones network, or subnet. Maybe this variable > could be taken as an override of the default "cobbler" hostname. That also can be added and used before auto discovery if it exists. Good idea. I also like auto-discovery though, as there wouldn't be anything automatic about COBBLER_SERVER (unless, possibly, it was set to "DISCOVER"). From bkearney at redhat.com Tue Aug 21 18:50:43 2007 From: bkearney at redhat.com (Bryan Kearney) Date: Tue, 21 Aug 2007 14:50:43 -0400 Subject: [et-mgmt-tools] import existing vms into virt-manager Message-ID: <46CB3403.9080800@redhat.com> I have been trying to get xen and kvm up and running, and have eventually done so by setting up a bridge and defining the vm through command line. I launch them now with the following commands: qemu-kvm -hda /dev/mapper/VolGroup00-virtFactory -boot c -net nic,macaddr=52:64:60:12:34:78,vlan0 -net tap,vlan0 -m 256 I would like to start using the virt-* tools directly. Is there a way for me to import this information into virt manager? Both the wizard and virt-install assume I am starting with the bare install. -- bk From berrange at redhat.com Tue Aug 21 19:01:06 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 21 Aug 2007 20:01:06 +0100 Subject: [et-mgmt-tools] import existing vms into virt-manager In-Reply-To: <46CB3403.9080800@redhat.com> References: <46CB3403.9080800@redhat.com> Message-ID: <20070821190106.GB27671@redhat.com> On Tue, Aug 21, 2007 at 02:50:43PM -0400, Bryan Kearney wrote: > I have been trying to get xen and kvm up and running, and have > eventually done so by setting up a bridge and defining the vm through > command line. I launch them now with the following commands: > > qemu-kvm > -hda /dev/mapper/VolGroup00-virtFactory > -boot c > -net nic,macaddr=52:64:60:12:34:78,vlan0 > -net tap,vlan0 > -m 256 > > I would like to start using the virt-* tools directly. Is there a way > for me to import this information into virt manager? Both the wizard and > virt-install assume I am starting with the bare install. At this time there's no official easy way to do it. If you have pre-installed disk images either, manually write XML config to go wit hit and use 'virsh define foo.xml'. Or run 'virt-install' pointing it to your existing disk, and let it boot up into the installer. Then just use 'virsh destroy' to shutdown the installer. The installer won't have done any damage to your disks at that point, but virt-install will have written out a permanent config file which yuo can now use with 'virsh start'. I plan to add a '--import' flag to virt-install to make this a little more pleasant where it'd just check your disk image for a valid boot sector and if so write out the config & skip the booting of installer altogether. Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From bkearney at redhat.com Tue Aug 21 19:01:00 2007 From: bkearney at redhat.com (Bryan Kearney) Date: Tue, 21 Aug 2007 15:01:00 -0400 Subject: [et-mgmt-tools] import existing vms into virt-manager In-Reply-To: <20070821190106.GB27671@redhat.com> References: <46CB3403.9080800@redhat.com> <20070821190106.GB27671@redhat.com> Message-ID: <46CB366C.4090103@redhat.com> thanks. I did not see in the docs.. is there an xsd for the xml file? -- bk Daniel P. Berrange wrote: > On Tue, Aug 21, 2007 at 02:50:43PM -0400, Bryan Kearney wrote: > >> I have been trying to get xen and kvm up and running, and have >> eventually done so by setting up a bridge and defining the vm through >> command line. I launch them now with the following commands: >> >> qemu-kvm >> -hda /dev/mapper/VolGroup00-virtFactory >> -boot c >> -net nic,macaddr=52:64:60:12:34:78,vlan0 >> -net tap,vlan0 >> -m 256 >> >> I would like to start using the virt-* tools directly. Is there a way >> for me to import this information into virt manager? Both the wizard and >> virt-install assume I am starting with the bare install. >> > > At this time there's no official easy way to do it. > > If you have pre-installed disk images either, manually write XML config > to go wit hit and use 'virsh define foo.xml'. Or run 'virt-install' pointing > it to your existing disk, and let it boot up into the installer. Then just > use 'virsh destroy' to shutdown the installer. The installer won't have > done any damage to your disks at that point, but virt-install will have > written out a permanent config file which yuo can now use with 'virsh start'. > > I plan to add a '--import' flag to virt-install to make this a little more > pleasant where it'd just check your disk image for a valid boot sector and > if so write out the config & skip the booting of installer altogether. > > Regards, > Dan. > From berrange at redhat.com Tue Aug 21 19:05:31 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 21 Aug 2007 20:05:31 +0100 Subject: [et-mgmt-tools] import existing vms into virt-manager In-Reply-To: <46CB366C.4090103@redhat.com> References: <46CB3403.9080800@redhat.com> <20070821190106.GB27671@redhat.com> <46CB366C.4090103@redhat.com> Message-ID: <20070821190531.GC27671@redhat.com> On Tue, Aug 21, 2007 at 03:01:00PM -0400, Bryan Kearney wrote: > thanks. I did not see in the docs.. is there an xsd for the xml file? There's a reference page http://libvirt.org/format.html And a RNG grammar http://libvirt.org/libvirt.rng Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Tue Aug 21 21:20:13 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 21 Aug 2007 22:20:13 +0100 Subject: [et-mgmt-tools] FYI: virt-install console changes Message-ID: <20070821212013.GF27671@redhat.com> FYI, I have just pushed a change to virt-install which switches it from using 'vncviewer' and 'xm console' over to use virt-viewer and the generic 'virsh console'. This allows it to work with KVM, and provides a VNC viewer compatible with the extensions being added to QEMU's VNC server for authentication. Assuming you've got virt-viewer installed, you shouldn't notice any major differences in way things work - just a little prettier VNC viewer. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Wed Aug 22 01:36:05 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Wed, 22 Aug 2007 02:36:05 +0100 Subject: [et-mgmt-tools] import existing vms into virt-manager In-Reply-To: <46CB8F1D.6090103@redhat.com> References: <46CB3403.9080800@redhat.com> <20070821190106.GB27671@redhat.com> <46CB366C.4090103@redhat.com> <20070821190531.GC27671@redhat.com> <46CB8F1D.6090103@redhat.com> Message-ID: <20070822013605.GA16203@redhat.com> On Tue, Aug 21, 2007 at 09:19:25PM -0400, Bryan Kearney wrote: > Thanks.. that worked like a champ. One last question. After I loaded a > couple of files, I noticed that interface commands like this: > > > > > > > with different mac addresses still result in collisions on the fd=10 > > /usr/bin/qemu-kvm -M pc -m 250 -smp 1 -monitor pty -no-acpi -boot c -hda > /dev/mapper/VolGroup00-f7kde -net nic,macaddr=52:54:60:12:34:78,vlan=0 > -net tap,fd=10,script=,vlan=0 -usb -vnc 127.0.0.1:-5899 > > which means I can not launch more then one at a time. Is there a way to > control this? This isn't a libvirt problem, this is a bug in KVM/QEMU's emulatiuon of the rtl8139 NIC card. If you have more than one it mis-handles the MMIO reads so they all end up being directed at the same card. There's a BZ about it somewhere and we'll fix it in a future update. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From bkearney at redhat.com Wed Aug 22 17:23:59 2007 From: bkearney at redhat.com (Bryan Kearney) Date: Wed, 22 Aug 2007 13:23:59 -0400 Subject: [et-mgmt-tools] koan installing to a kvm Message-ID: <46CC712F.8050709@redhat.com> I have gotten koan to install to a kvm.. way cool. in the lib virt config, it leaves around the following: /var/tmp/virtinst-vmlinuz.Iidlr1 /var/tmp/virtinst-initrd.img.qE46sn ks=http://10.11.239.64/cblr/ki......... which means I can not start the application without export and importing the config. Is there a way to edit this directly from koan? -- bk From mdehaan at redhat.com Wed Aug 22 17:40:23 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 22 Aug 2007 13:40:23 -0400 Subject: [et-mgmt-tools] koan installing to a kvm In-Reply-To: <46CC712F.8050709@redhat.com> References: <46CC712F.8050709@redhat.com> Message-ID: <46CC7507.8030206@redhat.com> Bryan Kearney wrote: > I have gotten koan to install to a kvm.. way cool. in the lib virt > config, it leaves around the following: > > > /var/tmp/virtinst-vmlinuz.Iidlr1 > /var/tmp/virtinst-initrd.img.qE46sn > ks=http://10.11.239.64/cblr/ki......... > > > which means I can not start the application without export and > importing the config. Is there a way to edit this directly from koan? > > -- bk > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools This is fixed by a patch to virtinst that I emailed to the list a while back, and will be included in a future version of virtinst. Anyone using KVM from a tree install-tree source (as opposed to an ISO that virt-manager requests) should apply the patch: https://www.redhat.com/archives/et-mgmt-tools/2007-August/msg00045.html From satyaakam at gmail.com Thu Aug 23 14:37:22 2007 From: satyaakam at gmail.com (satyaakam goswami) Date: Thu, 23 Aug 2007 20:07:22 +0530 Subject: [et-mgmt-tools] unable to generate configure virt-manger--devel Message-ID: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> trying to run the following command to generare configure script i get the following error $ autoconf configure.ac configure.ac:2: error: possibly undefined macro: AM_INIT_AUTOMAKE If this token and others are legitimate, please use m4_pattern_allow. See the Autoconf documentation. configure.ac:6: error: possibly undefined macro: AM_GNU_GETTEXT_VERSION configure.ac:7: error: possibly undefined macro: AM_GNU_GETTEXT configure.ac:10: error: possibly undefined macro: AC_PROG_LIBTOOL configure.ac:12: error: possibly undefined macro: AC_PYTHON_DEVEL Satya From thestrider at gmail.com Thu Aug 23 14:42:25 2007 From: thestrider at gmail.com (Adam Rosenwald) Date: Thu, 23 Aug 2007 10:42:25 -0400 Subject: [et-mgmt-tools] unable to generate configure virt-manger--devel In-Reply-To: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> References: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> Message-ID: <46CD9CD1.3020309@gmail.com> Try upgrading your version of automake. satyaakam goswami wrote: > trying to run the following command to generare configure script i > get the following error > > $ autoconf configure.ac > > > configure.ac:2: error: possibly undefined macro: AM_INIT_AUTOMAKE > If this token and others are legitimate, please use m4_pattern_allow. > See the Autoconf documentation. > configure.ac:6: error: possibly undefined macro: AM_GNU_GETTEXT_VERSION > configure.ac:7: error: possibly undefined macro: AM_GNU_GETTEXT > configure.ac:10: error: possibly undefined macro: AC_PROG_LIBTOOL > configure.ac:12: error: possibly undefined macro: AC_PYTHON_DEVEL > > > > Satya > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > From apevec at redhat.com Thu Aug 23 14:54:23 2007 From: apevec at redhat.com (Alan Pevec) Date: Thu, 23 Aug 2007 16:54:23 +0200 Subject: [et-mgmt-tools] unable to generate configure virt-manger--devel In-Reply-To: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> References: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> Message-ID: <46CD9F9F.1020009@redhat.com> satyaakam goswami wrote: > trying to run the following command to generare configure script i > get the following error > $ autoconf configure.ac > configure.ac:2: error: possibly undefined macro: AM_INIT_AUTOMAKE run aclocal to generate aclocal.m4 From satyaakam at gmail.com Thu Aug 23 14:58:45 2007 From: satyaakam at gmail.com (satyaakam goswami) Date: Thu, 23 Aug 2007 20:28:45 +0530 Subject: [et-mgmt-tools] unable to generate configure virt-manger--devel In-Reply-To: <46CD9CD1.3020309@gmail.com> References: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> <46CD9CD1.3020309@gmail.com> Message-ID: <6491e1350708230758r74039b71re33d4b9a0f6ecb2d@mail.gmail.com> On 8/23/07, Adam Rosenwald wrote: > Try upgrading your version of automake. > $automake --version automake (GNU automake) 1.9.6 From satyaakam at gmail.com Thu Aug 23 15:00:58 2007 From: satyaakam at gmail.com (satyaakam goswami) Date: Thu, 23 Aug 2007 20:30:58 +0530 Subject: [et-mgmt-tools] unable to generate configure virt-manger--devel In-Reply-To: <46CD9F9F.1020009@redhat.com> References: <6491e1350708230737r7c3c301fh83424b5ebd1280f5@mail.gmail.com> <46CD9F9F.1020009@redhat.com> Message-ID: <6491e1350708230800he2337aan94f429b4b6d0fd09@mail.gmail.com> On 8/23/07, Alan Pevec wrote: > satyaakam goswami wrote: > > trying to run the following command to generare configure script i > > get the following error > > $ autoconf configure.ac > > configure.ac:2: error: possibly undefined macro: AM_INIT_AUTOMAKE > > run aclocal to generate aclocal.m4 tried that still same error From hbrock at redhat.com Thu Aug 23 21:50:05 2007 From: hbrock at redhat.com (Hugh Brock) Date: Thu, 23 Aug 2007 17:50:05 -0400 Subject: [et-mgmt-tools] more gui improvements Message-ID: <46CE010D.4080908@redhat.com> Good evening everyone. I have just pushed a change to virt-manager that adds "action" icons to the list of vms and connections in the manager window. At the moment there are icons for "new" and "delete" for each connection, and a "delete" icon for each inactive VM. Be careful where you click; I still need to add an "are you sure" dialog for the delete icon! I'm not entirely sold on the position of the "action" column the icons are in, but overall I think the change is an improvement. Let me know what you think. Take care, --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From bkearney at redhat.com Fri Aug 24 01:06:37 2007 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 23 Aug 2007 21:06:37 -0400 Subject: [et-mgmt-tools] Re: Written down In-Reply-To: <46CDD92D.8070402@redhat.com> References: <46CDCA24.1010403@redhat.com> <46CDD92D.8070402@redhat.com> Message-ID: <46CE2F1D.4030107@redhat.com> I was assuming first that we do the automated testing of the image... does it boot? But, we should dig into that.. are there more things we should check. For example... network access, security access (e.g. telnet is off), and prehaps some others. I think the engineering and security guys could provide some good insight. We need to dive into this more, record the ideas and start to plan out what we will implement and in what order. -- bk Joey Boggs wrote: > Bryan, > > You mean from an already installed point of view for testing or from > bare metal/virt-factory all the way up to installed and tested? > > > Bryan Kearney wrote: >> Here is the info from the board: >> >> http://rhxdev.usersys.redhat.com/projects/rhx/wiki/QAProcess >> >> and in a preso. Lets discuss via the wiki. First of all, I would be >> curious if I recorded ir correctly and if the stage names are >> something we can live with. As for next steps, this is what I remember: >> >> - David would own Building/Deploying/Delivery and investigate what >> tools exist, and what this would look like. >> - Darryl would talk with the symphony dude about implementing >> running, and doing manual provisioning >> >> Joey.. can you take up Provisioning? Specifically.. since you have >> been working on the images.. can you reach out to the ET guys >> (Michael, Adrian, etc) to see what exists to automate hte process and >> verify that the images start up / shut down? >> >> -- bk > > From satyaakam at gmail.com Fri Aug 24 07:44:48 2007 From: satyaakam at gmail.com (satyaakam goswami) Date: Fri, 24 Aug 2007 13:14:48 +0530 Subject: [et-mgmt-tools] error when i run autobuild.sh in virt-manager-devel Message-ID: <6491e1350708240044w262b873am1d2b4987cf01447b@mail.gmail.com> gcc -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE=\"virt-manager\" -DVERSION=\"0.4.0\" -DENABLE_NLS=1 -DHAVE_GETTEXT=1 -DHAVE_DCGETTEXT=1 -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=1 -I. -I../../../src/graphWidgets -fPIC -Wall -DUSE_CAIRO=1 -I/usr/include/pygtk-2.0 -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/gtk-2.0 -I/usr/lib64/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -I/usr/include/python2.4 -g -O2 -MT sparkline_la-pysparklinemodule.lo -MD -MP -MF .deps/sparkline_la-pysparklinemodule.Tpo -c ../../../src/graphWidgets/pysparklinemodule.c -fPIC -DPIC -o .libs/sparkline_la-pysparklinemodule.o pysparklinemodule.c:52: error: redefinition of 'PyTypeObject PyGtkCellRendererSparkline_Type' pysparklinemodule.c:23: error: 'PyTypeObject PyGtkCellRendererSparkline_Type' previously declared here pysparklinemodule.c:121: error: redefinition of 'PyTypeObject PyGtkSparkline_Type' pysparklinemodule.c:24: error: 'PyTypeObject PyGtkSparkline_Type' previously declared here make[2]: *** [sparkline_la-pysparklinemodule.lo] Error 1 make[2]: Leaving directory `/root/virt-manager--devel/build/src/graphWidgets' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/root/virt-manager--devel/build/src' make: *** [all-recursive] Error 1 i am trying to build virt-manager on RHEL 5.0 From tomg at mcclatchyinteractive.com Fri Aug 24 14:37:51 2007 From: tomg at mcclatchyinteractive.com (Tom Georgoulias) Date: Fri, 24 Aug 2007 10:37:51 -0400 Subject: [et-mgmt-tools] koan not grabbing updated ks.cfg from cobbler server after a botched kickstart? Message-ID: <46CEED3F.8080401@mcclatchyinteractive.com> Started experimenting with koan today and got myself into a bit of circle. Used koan with --replace-self and everything went smoothly until anaconda choked on a syntax error in my ks.cfg. I corrected the error in the ks.cfg on the cobbler server, used "cobbler sync" to get the fixed ks.cfg into the profile, then rebooted the client to reattempt teh install. Unfortunately for me, koan didn't attempt to get the corrected ks.cfg from the cobbler server, it used the one already on the client. I tried to boot up the server I was trying to reinstall and manually edit /var/spool/koan/ks.cfg, but that didn't work either. Where does koan keep the ks.cfg? Thanks, Tom -- Tom Georgoulias Sr. Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com From mdehaan at redhat.com Fri Aug 24 14:44:54 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 24 Aug 2007 10:44:54 -0400 Subject: [et-mgmt-tools] koan not grabbing updated ks.cfg from cobbler server after a botched kickstart? In-Reply-To: <46CEED3F.8080401@mcclatchyinteractive.com> References: <46CEED3F.8080401@mcclatchyinteractive.com> Message-ID: <46CEEEE6.5040307@redhat.com> Tom Georgoulias wrote: > Started experimenting with koan today and got myself into a bit of > circle. > > Used koan with --replace-self and everything went smoothly until > anaconda choked on a syntax error in my ks.cfg. I corrected the error > in the ks.cfg on the cobbler server, used "cobbler sync" to get the > fixed ks.cfg into the profile, then rebooted the client to reattempt > teh install. > > Unfortunately for me, koan didn't attempt to get the corrected ks.cfg > from the cobbler server, it used the one already on the client. I > tried to boot up the server I was trying to reinstall and manually > edit /var/spool/koan/ks.cfg, but that didn't work either. Where does > koan keep the ks.cfg? > > Thanks, > Tom Koan embeds the ks.cfg in the image, in order to work around some DHCP timeouts that are possible in quite a few network configurations. So all you needed to do was to run "--replace-self" again to add the new kickstart entry. You may also find the "cobbler validateks" command useful, which checks all of the kickstart files you have for possible errors, using the "validateks" tool. This can help find some errors (but not all) prior to install time. --Michael From bkearney at redhat.com Fri Aug 24 14:44:45 2007 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 24 Aug 2007 10:44:45 -0400 Subject: [et-mgmt-tools] Re: Written down In-Reply-To: <46CE2F1D.4030107@redhat.com> References: <46CDCA24.1010403@redhat.com> <46CDD92D.8070402@redhat.com> <46CE2F1D.4030107@redhat.com> Message-ID: <46CEEEDD.1090605@redhat.com> Sorry... did not mean to post to et-mgmt-tools at redhat.com. Sorry about that. -- bk Bryan Kearney wrote: > I was assuming first that we do the automated..... From tomg at mcclatchyinteractive.com Fri Aug 24 17:38:45 2007 From: tomg at mcclatchyinteractive.com (Tom Georgoulias) Date: Fri, 24 Aug 2007 13:38:45 -0400 Subject: [et-mgmt-tools] koan not grabbing updated ks.cfg from cobbler server after a botched kickstart? In-Reply-To: <46CEEEE6.5040307@redhat.com> References: <46CEED3F.8080401@mcclatchyinteractive.com> <46CEEEE6.5040307@redhat.com> Message-ID: <46CF17A5.7040206@mcclatchyinteractive.com> Michael DeHaan wrote: > Koan embeds the ks.cfg in the image, in order to work around some DHCP > timeouts that are possible in quite a few network configurations. > > So all you needed to do was to run "--replace-self" again to add the new > kickstart entry. That's what I had to do in the end. I thought I was doing the lazy repair by rerunning it, but sounds like that is the correct approach. At least I learned something while poking around. :) Thanks for your help. Tom -- Tom Georgoulias Sr. Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com From jason at 3dogs.us Sat Aug 25 00:52:26 2007 From: jason at 3dogs.us (Jason Hartley) Date: Fri, 24 Aug 2007 20:52:26 -0400 (EDT) Subject: [et-mgmt-tools] Xen System builds using Koan Message-ID: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> First off, I really like the design of Cobbler and Koan. I especially like the kickstart templates and the flexibility that the templates give you. The only problem is that I am having trouble with the issue below and would appreciate any kind of help. I have a Centos 5 system that I have setup as a Xen host. I have setup the EPEL repository and used it to install Cobbler version 0.4.8-1,along with koan 0.4.0-1. I have successfully used the tutorial at http://cobbler.et.redhat.com/cobbler-import.php with a few modifications to build a Xen guest using the default profile for Centos 5 that is created when you import from the Centos 5 DVD. When I attempt to build from a system definition that I have built from the default profile I get the below error. I have started to dig through the code, but would like some direction in debugging the problem and maybe a push in the right direction on what might be wrong. [root at hydra ~]# koan --server=hydra --system=sirrush --virt - fetching configuration for system: (sirrush) Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/koan/app.py", line 468, in get_profile_xmlrpc return self.xmlrpc_server.get_profile_for_koan(profile_name) File "/usr/lib/python2.4/xmlrpclib.py", line 1096, in __call__ return self.__send(self.__name, args) File "/usr/lib/python2.4/xmlrpclib.py", line 1383, in __request verbose=self.__verbose File "/usr/lib/python2.4/xmlrpclib.py", line 1147, in request return self._parse_response(h.getfile(), sock) File "/usr/lib/python2.4/xmlrpclib.py", line 1286, in _parse_response return u.close() File "/usr/lib/python2.4/xmlrpclib.py", line 744, in close raise Fault(**self._stack[0]) Fault: File "/usr/bin/koan", line 38, in ? sys.exit(app.main() or 0) File "/usr/lib/python2.4/site-packages/koan/app.py", line 108, in main k.run() File "/usr/lib/python2.4/site-packages/koan/app.py", line 175, in run self.do_virt() File "/usr/lib/python2.4/site-packages/koan/app.py", line 306, in do_virt return self.do_net_install("/var/lib/xen",after_download) File "/usr/lib/python2.4/site-packages/koan/app.py", line 266, in do_net_install profile_data = self.get_system_xmlrpc(self.system) File "/usr/lib/python2.4/site-packages/koan/app.py", line 538, in get_system_xmlrpc profile_data = self.get_profile_xmlrpc(self.safe_load(system_data,'profile')) File "/usr/lib/python2.4/site-packages/koan/app.py", line 471, in get_profile_xmlrpc self.connect_fail() File "/usr/lib/python2.4/site-packages/koan/app.py", line 454, in connect_fail print self.trace_me() File "/usr/lib/python2.4/site-packages/koan/app.py", line 449, in trace_me x = traceback.extract_stack() Could not communicate with hydra:25151 [root at hydra ~]# Thanks, Jason From wright at imageworks.com Sat Aug 25 00:44:38 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 24 Aug 2007 17:44:38 -0700 Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> Message-ID: <46CF7B76.2070606@imageworks.com> Jason Hartley wrote: > > First off, I really like the design of Cobbler and Koan. I especially > like the kickstart templates and the flexibility that the templates give > you. The only problem is that I am having trouble with the issue below > and would appreciate any kind of help. > > I have a Centos 5 system that I have setup as a Xen host. I have setup > the EPEL repository and used it to install Cobbler version 0.4.8-1,along > with koan 0.4.0-1. I have successfully used the tutorial at > http://cobbler.et.redhat.com/cobbler-import.php with a few modifications > to build a Xen guest using the default profile for Centos 5 that is > created when you import from the Centos 5 DVD. When I attempt to build > from a system definition that I have built from the default profile I get > the below error. I have started to dig through the code, but would like > some direction in debugging the problem and maybe a push in the right > direction on what might be wrong. > dumb question first - to get it out of the way - have you started cobblerd? I seem to remember that step missing from the documentation, and I've been too lazy to report that... -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From jason at 3dogs.us Sat Aug 25 01:13:54 2007 From: jason at 3dogs.us (Jason Hartley) Date: Fri, 24 Aug 2007 21:13:54 -0400 (EDT) Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <46CF7B76.2070606@imageworks.com> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <46CF7B76.2070606@imageworks.com> Message-ID: <54044.192.168.0.10.1188004434.squirrel@www.3dogs.us> On Fri, August 24, 2007 8:44 pm, Peter Wright wrote: > Jason Hartley wrote: >> >> First off, I really like the design of Cobbler and Koan. I especially >> like the kickstart templates and the flexibility that the templates give >> you. The only problem is that I am having trouble with the issue below >> and would appreciate any kind of help. >> >> I have a Centos 5 system that I have setup as a Xen host. I have setup >> the EPEL repository and used it to install Cobbler version 0.4.8-1,along >> with koan 0.4.0-1. I have successfully used the tutorial at >> http://cobbler.et.redhat.com/cobbler-import.php with a few modifications >> to build a Xen guest using the default profile for Centos 5 that is >> created when you import from the Centos 5 DVD. When I attempt to build >> from a system definition that I have built from the default profile I >> get >> the below error. I have started to dig through the code, but would like >> some direction in debugging the problem and maybe a push in the right >> direction on what might be wrong. >> > > dumb question first - to get it out of the way - have you started > cobblerd? I seem to remember that step missing from the documentation, > and I've been too lazy to report that... > > -pete > Hey Pete, Yes, I have started cobblerd. I can telnet to port 25151 and connect. Here is the output from netstat -anp: [root at hydra ~]# netstat -anp|grep 25151 tcp 0 0 0.0.0.0:25151 0.0.0.0:* LISTEN 18323/python [root at hydra ~]# ps -ef|grep 18323 root 18323 18322 0 Aug23 ? 00:00:00 /usr/bin/python /usr/bin/cobblerd root 21343 18330 0 21:00 pts/4 00:00:00 grep 18323 [root at hydra ~]# Thanks, Jason Regards, Jason From jason at 3dogs.us Sat Aug 25 01:14:45 2007 From: jason at 3dogs.us (Jason Hartley) Date: Fri, 24 Aug 2007 21:14:45 -0400 (EDT) Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <46CF7B76.2070606@imageworks.com> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <46CF7B76.2070606@imageworks.com> Message-ID: <54046.192.168.0.10.1188004485.squirrel@www.3dogs.us> On Fri, August 24, 2007 8:44 pm, Peter Wright wrote: > Jason Hartley wrote: >> >> First off, I really like the design of Cobbler and Koan. I especially >> like the kickstart templates and the flexibility that the templates give >> you. The only problem is that I am having trouble with the issue below >> and would appreciate any kind of help. >> >> I have a Centos 5 system that I have setup as a Xen host. I have setup >> the EPEL repository and used it to install Cobbler version 0.4.8-1,along >> with koan 0.4.0-1. I have successfully used the tutorial at >> http://cobbler.et.redhat.com/cobbler-import.php with a few modifications >> to build a Xen guest using the default profile for Centos 5 that is >> created when you import from the Centos 5 DVD. When I attempt to build >> from a system definition that I have built from the default profile I >> get >> the below error. I have started to dig through the code, but would like >> some direction in debugging the problem and maybe a push in the right >> direction on what might be wrong. >> > > dumb question first - to get it out of the way - have you started > cobblerd? I seem to remember that step missing from the documentation, > and I've been too lazy to report that... > > -pete > > Hey Pete, Yes, I have started cobblerd. I can telnet to port 25151 and connect. Here is the output from netstat -anp: [root at hydra ~]# netstat -anp|grep 25151 tcp 0 0 0.0.0.0:25151 0.0.0.0:* LISTEN 18323/python [root at hydra ~]# ps -ef|grep 18323 root 18323 18322 0 Aug23 ? 00:00:00 /usr/bin/python /usr/bin/cobblerd root 21343 18330 0 21:00 pts/4 00:00:00 grep 18323 [root at hydra ~]# Thanks, Jason From drew.einhorn at gmail.com Sat Aug 25 16:05:26 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Sat, 25 Aug 2007 10:05:26 -0600 Subject: [et-mgmt-tools] cobbler sync crashes Message-ID: I've been on vacation and sick and don't remember what the last thing I did when I was last working on this project, so I may have forgotten something important. Anyway things were working with a version of cobbler I built from the git repository. (I was an early adopter of the SNIPPET feature). I installed 0.6.0 and now cobbler sync dies with: Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 743, in main BootCLI(sys.argv).run() File "/usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 118, in run self.relay_args(self.args[1:], self.commands['toplevel']) File "/usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 553, in relay_args commands[args[0]](args[1:]) File "/usr/lib/python2.4/site-packages/cobbler/cobbler.py", line 575, in sync self.api.sync() File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 155, in sync sync = action_sync.BootSync(self._config) File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line 55, in __init__ self.load_snippet_cache() File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line 466, in load_snippet_cache snip_file = open(snip) IOError: [Errno 21] Is a directory Hope I figure out how to get past this one before I get an answer on Monday. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From smooge at gmail.com Sun Aug 26 00:05:37 2007 From: smooge at gmail.com (Stephen John Smoogen) Date: Sat, 25 Aug 2007 18:05:37 -0600 Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> Message-ID: <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> On 8/24/07, Jason Hartley wrote: > First off, I really like the design of Cobbler and Koan. I especially > like the kickstart templates and the flexibility that the templates give > you. The only problem is that I am having trouble with the issue below > and would appreciate any kind of help. > > I have a Centos 5 system that I have setup as a Xen host. I have setup > the EPEL repository and used it to install Cobbler version 0.4.8-1,along > with koan 0.4.0-1. I have successfully used the tutorial at > http://cobbler.et.redhat.com/cobbler-import.php with a few modifications > to build a Xen guest using the default profile for Centos 5 that is > created when you import from the Centos 5 DVD. When I attempt to build > from a system definition that I have built from the default profile I get > the below error. I have started to dig through the code, but would like > some direction in debugging the problem and maybe a push in the right > direction on what might be wrong. > I would first get the cobbler && koan from EPEL testing. They are updated and fix a lot of bugs from the EPEL-stable version. -- Stephen J Smoogen. -- CSIRT/Linux System Administrator How far that little candle throws his beams! So shines a good deed in a naughty world. = Shakespeare. "The Merchant of Venice" From sakaia at jp.fujitsu.com Mon Aug 27 05:23:32 2007 From: sakaia at jp.fujitsu.com (Atsushi SAKAI) Date: Mon, 27 Aug 2007 14:23:32 +0900 Subject: [et-mgmt-tools] [patch] fix virt-image manual typo Message-ID: <200708270523.l7R5NYek014471@fjmscan502.ms.jp.fujitsu.com> fix virt-image manual typo virt-image-xml.pod | 14 +++++++------- virt-image.1 | 6 +++--- virt-image.5 | 14 +++++++------- virt-image.pod | 6 +++--- 4 files changed, 20 insertions(+), 20 deletions(-) It seems the typo is same as that of virt-install. Thanks Atsushi SAKAI -------------- next part -------------- A non-text attachment was scrubbed... Name: fix_virt_image_man_typo.patch Type: application/octet-stream Size: 8631 bytes Desc: not available URL: From mdehaan at redhat.com Mon Aug 27 14:29:06 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 27 Aug 2007 10:29:06 -0400 Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> Message-ID: <46D2DFB2.5000501@redhat.com> Stephen John Smoogen wrote: > > I would first get the cobbler && koan from EPEL testing. They are > updated and fix a lot of bugs from the EPEL-stable version. > > > > > Exactly what I was going to say :) 0.6.0 is out now, and I'll be releasing 0.6.1 (a few more refinements) later this week. 0.4.x is pretty old. --Michael From mdehaan at redhat.com Mon Aug 27 14:33:51 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 27 Aug 2007 10:33:51 -0400 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: References: Message-ID: <46D2E0CF.8080105@redhat.com> drew einhorn wrote: > > I installed 0.6.0 and now cobbler sync dies with: > > File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line > 466, in load_snippet_cache > snip_file = open(snip) > IOError: [Errno 21] Is a directory > At first guess, it looks like you have a subdirectory in your /var/lib/cobbler/snippets directory where I wasn't expecting any. The short term solution is to not put subdirectories there (cobbler won't read them). I'll fix 0.6.1 to ignore potential subdirectories. --Michael From hyclak at math.ohiou.edu Mon Aug 27 14:51:47 2007 From: hyclak at math.ohiou.edu (Matt Hyclak) Date: Mon, 27 Aug 2007 10:51:47 -0400 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: <46D2E0CF.8080105@redhat.com> References: <46D2E0CF.8080105@redhat.com> Message-ID: <20070827145147.GI16892@math.ohiou.edu> On Mon, Aug 27, 2007 at 10:33:51AM -0400, Michael DeHaan enlightened us: > drew einhorn wrote: > > > >I installed 0.6.0 and now cobbler sync dies with: > > > > File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line > >466, in load_snippet_cache > > snip_file = open(snip) > >IOError: [Errno 21] Is a directory > > > > At first guess, it looks like you have a subdirectory in your > /var/lib/cobbler/snippets directory where I wasn't expecting any. > > The short term solution is to not put subdirectories there (cobbler > won't read them). I'll fix 0.6.1 to ignore potential subdirectories. > Might be worth emulating some other things that have directories that are read, but only read files with a particular ending - e.g. partition_select.snip and ignore all other files/directories. That way it's easy to disable a particular snippet by renaming it to something else. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263 From mdehaan at redhat.com Mon Aug 27 15:10:35 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 27 Aug 2007 11:10:35 -0400 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: <20070827145147.GI16892@math.ohiou.edu> References: <46D2E0CF.8080105@redhat.com> <20070827145147.GI16892@math.ohiou.edu> Message-ID: <46D2E96B.5030806@redhat.com> Matt Hyclak wrote: > On Mon, Aug 27, 2007 at 10:33:51AM -0400, Michael DeHaan enlightened us: > >> drew einhorn wrote: >> >>> I installed 0.6.0 and now cobbler sync dies with: >>> >>> File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line >>> 466, in load_snippet_cache >>> snip_file = open(snip) >>> IOError: [Errno 21] Is a directory >>> >>> >> At first guess, it looks like you have a subdirectory in your >> /var/lib/cobbler/snippets directory where I wasn't expecting any. >> >> The short term solution is to not put subdirectories there (cobbler >> won't read them). I'll fix 0.6.1 to ignore potential subdirectories. >> >> > > Might be worth emulating some other things that have directories that are > read, but only read files with a particular ending - e.g. > partition_select.snip and ignore all other files/directories. That way it's > easy to disable a particular snippet by renaming it to something else. > > Matt > > If a kickstart references "SNIPPET::bar" all you need to do currently to disable it is to rename "bar" to something not named "bar". Searching subdirectories (and possibly even namespacing them) is reasonable though. --Michael From lfarkas at bppiac.hu Mon Aug 27 15:21:55 2007 From: lfarkas at bppiac.hu (Farkas Levente) Date: Mon, 27 Aug 2007 17:21:55 +0200 Subject: [et-mgmt-tools] bridged network in kvm with virt-manager Message-ID: <46D2EC13.8020005@bppiac.hu> hi, is it possible to setup bridged network for kvm guest with virt-manager as described in: http://libvirt.org/format.html#KVM1? or should i 've to tune config files manualy? thanks in advance. -- Levente "Si vis pacem para bellum!" From hyclak at math.ohiou.edu Mon Aug 27 15:29:15 2007 From: hyclak at math.ohiou.edu (Matt Hyclak) Date: Mon, 27 Aug 2007 11:29:15 -0400 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: <46D2E96B.5030806@redhat.com> References: <46D2E0CF.8080105@redhat.com> <20070827145147.GI16892@math.ohiou.edu> <46D2E96B.5030806@redhat.com> Message-ID: <20070827152915.GC622@math.ohiou.edu> On Mon, Aug 27, 2007 at 11:10:35AM -0400, Michael DeHaan enlightened us: > Matt Hyclak wrote: > >On Mon, Aug 27, 2007 at 10:33:51AM -0400, Michael DeHaan enlightened us: > > > >>drew einhorn wrote: > >> > >>>I installed 0.6.0 and now cobbler sync dies with: > >>> > >>> File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line > >>>466, in load_snippet_cache > >>> snip_file = open(snip) > >>>IOError: [Errno 21] Is a directory > >>> > >>> > >>At first guess, it looks like you have a subdirectory in your > >>/var/lib/cobbler/snippets directory where I wasn't expecting any. > >> > >>The short term solution is to not put subdirectories there (cobbler > >>won't read them). I'll fix 0.6.1 to ignore potential subdirectories. > >> > >> > > > >Might be worth emulating some other things that have directories that are > >read, but only read files with a particular ending - e.g. > >partition_select.snip and ignore all other files/directories. That way it's > >easy to disable a particular snippet by renaming it to something else. > > > >Matt > > > > > If a kickstart references "SNIPPET::bar" all you need to do currently to > disable it is to rename > "bar" to something not named "bar". > True. Just hope you don't have another snippet named "not bar". :-) > Searching subdirectories (and possibly even namespacing them) is > reasonable though. > Oooh, getting advanced on us. That might be interesting... Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263 From drew.einhorn at gmail.com Mon Aug 27 15:33:50 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Mon, 27 Aug 2007 09:33:50 -0600 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: <20070827152915.GC622@math.ohiou.edu> References: <46D2E0CF.8080105@redhat.com> <20070827145147.GI16892@math.ohiou.edu> <46D2E96B.5030806@redhat.com> <20070827152915.GC622@math.ohiou.edu> Message-ID: Actually I was going in an entirely different direction with the subdirectory. It's an RCS directory. I've been using for years and never switched to cvs, or svn. It would be good if version control subdirectories were quietly ignored. On 8/27/07, Matt Hyclak wrote: > > On Mon, Aug 27, 2007 at 11:10:35AM -0400, Michael DeHaan enlightened us: > > Matt Hyclak wrote: > > >On Mon, Aug 27, 2007 at 10:33:51AM -0400, Michael DeHaan enlightened > us: > > > > > >>drew einhorn wrote: > > >> > > >>>I installed 0.6.0 and now cobbler sync dies with: > > >>> > > >>> File "/usr/lib/python2.4/site-packages/cobbler/action_sync.py", line > > >>>466, in load_snippet_cache > > >>> snip_file = open(snip) > > >>>IOError: [Errno 21] Is a directory > > >>> > > >>> > > >>At first guess, it looks like you have a subdirectory in your > > >>/var/lib/cobbler/snippets directory where I wasn't expecting any. > > >> > > >>The short term solution is to not put subdirectories there (cobbler > > >>won't read them). I'll fix 0.6.1 to ignore potential > subdirectories. > > >> > > >> > > > > > >Might be worth emulating some other things that have directories that > are > > >read, but only read files with a particular ending - e.g. > > >partition_select.snip and ignore all other files/directories. That way > it's > > >easy to disable a particular snippet by renaming it to something else. > > > > > >Matt > > > > > > > > If a kickstart references "SNIPPET::bar" all you need to do currently to > > disable it is to rename > > "bar" to something not named "bar". > > > > True. Just hope you don't have another snippet named "not bar". :-) > > > Searching subdirectories (and possibly even namespacing them) is > > reasonable though. > > > > Oooh, getting advanced on us. That might be interesting... > > Matt > > -- > Matt Hyclak > Department of Mathematics > Department of Social Work > Ohio University > (740) 593-1263 > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From hbrock at redhat.com Mon Aug 27 15:44:15 2007 From: hbrock at redhat.com (Hugh Brock) Date: Mon, 27 Aug 2007 11:44:15 -0400 Subject: [et-mgmt-tools] [patch] fix virt-image manual typo In-Reply-To: <200708270523.l7R5NYek014471@fjmscan502.ms.jp.fujitsu.com> References: <200708270523.l7R5NYek014471@fjmscan502.ms.jp.fujitsu.com> Message-ID: <46D2F14F.6020906@redhat.com> Atsushi SAKAI wrote: > fix virt-image manual typo > > virt-image-xml.pod | 14 +++++++------- > virt-image.1 | 6 +++--- > virt-image.5 | 14 +++++++------- > virt-image.pod | 6 +++--- > 4 files changed, 20 insertions(+), 20 deletions(-) > > It seems the typo is same as that of virt-install. > Thanks, I have applied this! --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From mdehaan at redhat.com Mon Aug 27 15:44:20 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 27 Aug 2007 11:44:20 -0400 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: References: <46D2E0CF.8080105@redhat.com> <20070827145147.GI16892@math.ohiou.edu> <46D2E96B.5030806@redhat.com> <20070827152915.GC622@math.ohiou.edu> Message-ID: <46D2F154.6020803@redhat.com> drew einhorn wrote: > Actually I was going in an entirely different direction with the > subdirectory. > > It's an RCS directory. I've been using for years and never switched > to cvs, > or svn. It would be good if version control subdirectories were > quietly ignored. Not a problem. From drew.einhorn at gmail.com Mon Aug 27 15:54:18 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Mon, 27 Aug 2007 09:54:18 -0600 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: <46D2F154.6020803@redhat.com> References: <46D2E0CF.8080105@redhat.com> <20070827145147.GI16892@math.ohiou.edu> <46D2E96B.5030806@redhat.com> <20070827152915.GC622@math.ohiou.edu> <46D2F154.6020803@redhat.com> Message-ID: For now I'll nuke the RCS directory. Let me know when you have a fix committed to git and and I'll build an 0.6.0+ and put the RCS directory back right away instead of waiting for 0.6.1 to be released. Thanks for all your great work on a wonderful tool. On 8/27/07, Michael DeHaan wrote: > > drew einhorn wrote: > > Actually I was going in an entirely different direction with the > > subdirectory. > > > > It's an RCS directory. I've been using for years and never switched > > to cvs, > > or svn. It would be good if version control subdirectories were > > quietly ignored. > > Not a problem. > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From mdehaan at redhat.com Mon Aug 27 15:56:52 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Mon, 27 Aug 2007 11:56:52 -0400 Subject: [et-mgmt-tools] cobbler sync crashes In-Reply-To: References: <46D2E0CF.8080105@redhat.com> <20070827145147.GI16892@math.ohiou.edu> <46D2E96B.5030806@redhat.com> <20070827152915.GC622@math.ohiou.edu> <46D2F154.6020803@redhat.com> Message-ID: <46D2F444.6070501@redhat.com> drew einhorn wrote: > For now I'll nuke the RCS directory. Let me know when you have > a fix committed to git and and I'll build an 0.6.0+ and put the RCS > directory back right away instead of waiting for 0.6.1 to be released. It's already pushed. Here's the patch if you just want to apply that: http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h=2ead5ca92d371325a67d6cef0cf38c00e88c99b1 > > Thanks for all your great work on a wonderful tool. > you're welcome. Thanks! From dlutter at redhat.com Mon Aug 27 16:22:53 2007 From: dlutter at redhat.com (David Lutterkort) Date: Mon, 27 Aug 2007 09:22:53 -0700 Subject: [et-mgmt-tools] Re: [patch] fix virt-image manual typo In-Reply-To: <200708270523.l7R5NYek014471@fjmscan502.ms.jp.fujitsu.com> References: <200708270523.l7R5NYek014471@fjmscan502.ms.jp.fujitsu.com> Message-ID: <1188231773.25884.87.camel@localhost.localdomain> On Mon, 2007-08-27 at 14:23 +0900, Atsushi SAKAI wrote: > Atsushi SAKAI Excellent - thanks for the fixes. Committed as 261:6905717ca5a3 David From dlutter at redhat.com Mon Aug 27 16:37:30 2007 From: dlutter at redhat.com (David Lutterkort) Date: Mon, 27 Aug 2007 16:37:30 +0000 Subject: [et-mgmt-tools] bridged network in kvm with virt-manager In-Reply-To: <46D2EC13.8020005@bppiac.hu> References: <46D2EC13.8020005@bppiac.hu> Message-ID: <1188232650.25884.100.camel@localhost.localdomain> On Mon, 2007-08-27 at 17:21 +0200, Farkas Levente wrote: > hi, > is it possible to setup bridged network for kvm guest with virt-manager > as described in: http://libvirt.org/format.html#KVM1? > or should i 've to tune config files manualy? > thanks in advance. Absolutely .. all you need to do is set things up so that your physical NIC is shared over a bridge. See [1] for details David [1] http://watzmann.net/blog/index.php/2007/04/27/networking_with_kvm_and_libvirt From oehmes at de.ibm.com Mon Aug 27 19:38:32 2007 From: oehmes at de.ibm.com (Sven Oehme) Date: Mon, 27 Aug 2007 21:38:32 +0200 Subject: [et-mgmt-tools] Image Corruption Possible with qemu and qemu-kvm In-Reply-To: <1188232650.25884.100.camel@localhost.localdomain> References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> Message-ID: Hi, i corrupted a couple of virtual Linux images last week as i accidentally started them two times .. i opened a bug against qemu (not sure if this was a good idea) --> https://bugzilla.redhat.com/show_bug.cgi?id=253533 are there any plans to ensure that this can't happen in future releases of virt-manager with lock files in the image directory or similar ? thanks. Sven -------------- next part -------------- An HTML attachment was scrubbed... URL: From oehmes at de.ibm.com Mon Aug 27 20:19:42 2007 From: oehmes at de.ibm.com (Sven Oehme) Date: Mon, 27 Aug 2007 22:19:42 +0200 Subject: [et-mgmt-tools] Image Corruption Possible with qemu and qemu-kvm In-Reply-To: References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> Message-ID: i thought it might be good to post this on qemu-devel as well, as i see this as a general qemu / kvm / xen /whatevercomesnext issue . are there any plans to implement a default in qemu to prevent accessing the same image multiple times ? i know that in xen they deal with this problem somehow xen specific. wouldn't it be good to prevent multiple access to the same image by default and add a switch to overwrite it ? i am sure other people had the same issue as i had (corrupting images) and would be great to prevent people in the future messing around with data loss :-) Sven Sven Oehme/Germany/IBM at IBMDE Sent by: et-mgmt-tools-bounces at redhat.com 08/27/2007 09:38 PM Please respond to Fedora/Linux Management Tools To Fedora/Linux Management Tools cc Subject [et-mgmt-tools] Image Corruption Possible with qemu and qemu-kvm Hi, i corrupted a couple of virtual Linux images last week as i accidentally started them two times .. i opened a bug against qemu (not sure if this was a good idea) --> https://bugzilla.redhat.com/show_bug.cgi?id=253533 are there any plans to ensure that this can't happen in future releases of virt-manager with lock files in the image directory or similar ? thanks. Sven _______________________________________________ et-mgmt-tools mailing list et-mgmt-tools at redhat.com https://www.redhat.com/mailman/listinfo/et-mgmt-tools -------------- next part -------------- An HTML attachment was scrubbed... URL: From berrange at redhat.com Mon Aug 27 20:27:48 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Mon, 27 Aug 2007 21:27:48 +0100 Subject: [et-mgmt-tools] Image Corruption Possible with qemu and qemu-kvm In-Reply-To: References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> Message-ID: <20070827202748.GK9043@redhat.com> On Mon, Aug 27, 2007 at 10:19:42PM +0200, Sven Oehme wrote: > i thought it might be good to post this on qemu-devel as well, as i see > this as a general qemu / kvm / xen /whatevercomesnext issue . > > are there any plans to implement a default in qemu to prevent accessing > the same image multiple times ? Each QEMU instance is unware of each other so can't directly check this. One possiblity would be for QEMU to use fcntl() to take a lock on a disk image when opening it. With raw disk images at least, it is, however, safe to let multiple QEMU instances use it at once *provided* you have a clustered filesystem installed, rather than regular FAT or ext2/3, so one wouldn't want to exclude this use case. > i know that in xen they deal with this problem somehow xen specific. > wouldn't it be good to prevent multiple access to the same image by > default and add a switch to overwrite it ? In the scenario you mention, libvirt should probably do a sanity check for this before letting you start the guest. libvirt already supports the idea of 'shared' disk images where two or more guests can be optionally configured to have write access - basically assumes the admin requesting sharing knows what they're doing. > i corrupted a couple of virtual Linux images last week as i accidentally > started them two times .. > i opened a bug against qemu (not sure if this was a good idea) --> > https://bugzilla.redhat.com/show_bug.cgi?id=253533 > > are there any plans to ensure that this can't happen in future releases of > virt-manager with lock files in the image directory or similar ? Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From veillard at redhat.com Mon Aug 27 21:26:25 2007 From: veillard at redhat.com (Daniel Veillard) Date: Mon, 27 Aug 2007 17:26:25 -0400 Subject: [Qemu-devel] Re: [et-mgmt-tools] Image Corruption Possible with qemu and qemu-kvm In-Reply-To: <1188248558.21696.1.camel@squirrel> References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> <20070827202748.GK9043@redhat.com> <1188248558.21696.1.camel@squirrel> Message-ID: <20070827212624.GK23101@redhat.com> On Mon, Aug 27, 2007 at 04:02:38PM -0500, Anthony Liguori wrote: > > On Mon, 2007-08-27 at 21:27 +0100, Daniel P. Berrange wrote: > > On Mon, Aug 27, 2007 at 10:19:42PM +0200, Sven Oehme wrote: > > > i know that in xen they deal with this problem somehow xen specific. > > > wouldn't it be good to prevent multiple access to the same image by > > > default and add a switch to overwrite it ? > > > > In the scenario you mention, libvirt should probably do a sanity check for > > this before letting you start the guest. libvirt already supports the idea > > of 'shared' disk images where two or more guests can be optionally configured > > to have write access - basically assumes the admin requesting sharing knows > > what they're doing. > > I think this is the right level myself. Advisory locks work okay but > not all filesystems support them. It's particularly nasty when you have > a clustered filesystem in the host. I think it would do more harm than > good to have a feature like that was supposed to provide a safe-guard > but then frequently didn't work. Well initially I was really of the position that checking had to be done at the hypervisor level, but as this thread pointed out we can't always do this (there is also the potential of conflict between different hypervisor though that sounds more theorical than practical). Daniel -- Red Hat Virtualization group http://redhat.com/virtualization/ Daniel Veillard | virtualization library http://libvirt.org/ veillard at redhat.com | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/ From jason at 3dogs.us Mon Aug 27 23:52:21 2007 From: jason at 3dogs.us (Jason Hartley) Date: Mon, 27 Aug 2007 19:52:21 -0400 (EDT) Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <46D2DFB2.5000501@redhat.com> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> <46D2DFB2.5000501@redhat.com> Message-ID: <49304.192.168.0.10.1188258741.squirrel@www.3dogs.us> On Mon, August 27, 2007 10:29 am, Michael DeHaan wrote: > Stephen John Smoogen wrote: >> >> I would first get the cobbler && koan from EPEL testing. They are >> updated and fix a lot of bugs from the EPEL-stable version. >> >> >> >> >> > > Exactly what I was going to say :) > > 0.6.0 is out now, and I'll be releasing 0.6.1 (a few more refinements) > later this week. > 0.4.x is pretty old. > > --Michael > Thanks, Michael. Through trail and error I discovered that Cobbler/Koan did not like me using a DNS name for the system name to build a Xen VM. When I used a MAC address for the system it worked just fine. I should have yielded to the advice in the man pages and just used a MAC. Also, thanks for the heads up on 0.6.0. Along with the bug fixes it looks to have some neat features including an increased number of builtin variables, along with the kickstart snippets that I can use for kickstart templates. When will 0.6.0 be moved over to the standard EPEL repo? Regards, Jason From mdehaan at redhat.com Tue Aug 28 14:24:41 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 28 Aug 2007 10:24:41 -0400 Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <49304.192.168.0.10.1188258741.squirrel@www.3dogs.us> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> <46D2DFB2.5000501@redhat.com> <49304.192.168.0.10.1188258741.squirrel@www.3dogs.us> Message-ID: <46D43029.4050701@redhat.com> Jason Hartley wrote: > On Mon, August 27, 2007 10:29 am, Michael DeHaan wrote: > >> Stephen John Smoogen wrote: >> >>> I would first get the cobbler && koan from EPEL testing. They are >>> updated and fix a lot of bugs from the EPEL-stable version. >>> >>> >>> >>> >>> >>> >> Exactly what I was going to say :) >> >> 0.6.0 is out now, and I'll be releasing 0.6.1 (a few more refinements) >> later this week. >> 0.4.x is pretty old. >> >> --Michael >> >> > > Thanks, Michael. Through trail and error I discovered that Cobbler/Koan > did not like me using a DNS name for the system name to build a Xen VM. > When I used a MAC address for the system it worked just fine. I should > have yielded to the advice in the man pages and just used a MAC. > You don't have to do that in 0.6.x :) > Also, thanks for the heads up on 0.6.0. Along with the bug fixes it > looks to have some neat features including an increased number of > builtin variables, along with the kickstart snippets that I can use for > kickstart templates. When will 0.6.0 be moved over to the standard EPEL > repo? > Everyone just pull the RPM's out of EPEL testing. For what it's worth, I'm using EPEL for "stable" releases only, and "testing" builds of cobbler (like the 0.5.x series) will never be pushed there. From recent discussions on #epel (IRC), epel is probably going to push quarterly, which means I can't update periodically for bugfixes and improvements as I would like. So, as a result, what is in EPEL stable is currently very old and I have no way of updating it. If you don't want the rest of your system to get packages out of EPEL "testing" for whatever reason, you can just enable the repo, install cobbler, and disable the repo. As long as you are on the list you'll get the release announcements so that may be workable. > Regards, > Jason > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From krammer at aim-ag.com Tue Aug 28 15:09:45 2007 From: krammer at aim-ag.com (Krammer Oliver) Date: Tue, 28 Aug 2007 17:09:45 +0200 Subject: [et-mgmt-tools] Can't clone a DomU Message-ID: Hi Everybody, I hobe i am at the right list. I used virt-clone for cloning a domu i started the clone process with: ./virt-clone -o defaultsuse10 -n suseclone01 -f /var/lib/xen/images/suseclone01/disk0 -d This worked after that i got servxen01:/etc/xen/vm # xm list Name ID Mem VCPUs State Time(s) Domain-0 0 15093 8 r----- 4509.4 defaultsuse10 13 1024 2 -b---- 24.6 suseclone01 1024 1 0.0 When i start the dom with xm start suseclone01 Xen will hang up and dosn't respond any more The log file give me following statemantes: [2007-08-28 19:08:35 xend.XendDomainInfo 31046] DEBUG (XendDomainInfo:1335) XendDomainInfo.constructDomain [2007-08-28 19:08:35 xend 31046] DEBUG (balloon:146) Balloon: 16724 KiB free; need 16384; done. [2007-08-28 19:08:35 xend.XendDomainInfo 31046] DEBUG (XendDomainInfo:1382) XendDomainInfo.initDomain: 17 256 [2007-08-28 19:08:35 xend 31113] DEBUG (XendBootloader:69) Launching bootloader as ['/usr/lib/xen/boot/domUloader.py', '-q', '--args=TERM=xterm ', '--output=/var/run/xend/boot/xenbl.23654', '/var/lib/xen/images/suseclone01/disk0']. Some body could help me please? Thanx Mit freundlichen Gr??en, Best Regards. DI(FH) Oliver Krammer aim FINANCIAL SYSTEMS GmbH net www.aim-ag.com e-mail krammer at aim-ag.com tel (+43-1) 533 29 96 - 1390 fax (+43-1) 533 29 96 - 1500 smail aim FINANCIAL SYSTEMS GmbH Donau-City-Strasse 1 A-1220 Wien, Austria Diese E-Mail enth?lt vertrauliche und/oder rechtlich gesch?tzte Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail irrt?mlich erhalten haben, informieren Sie bitte sofort den Absender und vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte Weitergabe dieser Mail sind nicht gestattet. This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is strictly forbidden. -------------- next part -------------- An HTML attachment was scrubbed... URL: From lfarkas at bppiac.hu Tue Aug 28 15:22:09 2007 From: lfarkas at bppiac.hu (Farkas Levente) Date: Tue, 28 Aug 2007 17:22:09 +0200 Subject: [et-mgmt-tools] bridged network in kvm with virt-manager In-Reply-To: <1188232650.25884.100.camel@localhost.localdomain> References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> Message-ID: <46D43DA1.1010407@bppiac.hu> David Lutterkort wrote: > On Mon, 2007-08-27 at 17:21 +0200, Farkas Levente wrote: >> hi, >> is it possible to setup bridged network for kvm guest with virt-manager >> as described in: http://libvirt.org/format.html#KVM1? >> or should i 've to tune config files manualy? >> thanks in advance. > > Absolutely .. all you need to do is set things up so that your physical > NIC is shared over a bridge. See [1] for details > > David > > [1] > http://watzmann.net/blog/index.php/2007/04/27/networking_with_kvm_and_libvirt in virt-manager i'm not able to select shared physical devices since there is no device under it. so that's why i ask it. or how can i setup my nic to shared over a bridge? -- Levente "Si vis pacem para bellum!" From berrange at redhat.com Tue Aug 28 15:27:42 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 28 Aug 2007 16:27:42 +0100 Subject: [et-mgmt-tools] bridged network in kvm with virt-manager In-Reply-To: <46D43DA1.1010407@bppiac.hu> References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> <46D43DA1.1010407@bppiac.hu> Message-ID: <20070828152742.GD9503@redhat.com> On Tue, Aug 28, 2007 at 05:22:09PM +0200, Farkas Levente wrote: > David Lutterkort wrote: > > On Mon, 2007-08-27 at 17:21 +0200, Farkas Levente wrote: > >> hi, > >> is it possible to setup bridged network for kvm guest with virt-manager > >> as described in: http://libvirt.org/format.html#KVM1? > >> or should i 've to tune config files manualy? > >> thanks in advance. > > > > Absolutely .. all you need to do is set things up so that your physical > > NIC is shared over a bridge. See [1] for details > > > > David > > > > [1] > > http://watzmann.net/blog/index.php/2007/04/27/networking_with_kvm_and_libvirt > > in virt-manager i'm not able to select shared physical devices since > there is no device under it. > so that's why i ask it. > or how can i setup my nic to shared over a bridge? Read the link above '2. Xen-like bridging' tells you exactly how to config the networking in /etc/sysconfig/network-scripts to enable shared physical devices Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From smooge at gmail.com Tue Aug 28 18:14:23 2007 From: smooge at gmail.com (Stephen John Smoogen) Date: Tue, 28 Aug 2007 12:14:23 -0600 Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <46D43029.4050701@redhat.com> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> <46D2DFB2.5000501@redhat.com> <49304.192.168.0.10.1188258741.squirrel@www.3dogs.us> <46D43029.4050701@redhat.com> Message-ID: <80d7e4090708281114m61e101eeh59dfa1c469918ced@mail.gmail.com> On 8/28/07, Michael DeHaan wrote: > Jason Hartley wrote: > > On Mon, August 27, 2007 10:29 am, Michael DeHaan wrote: > > > >> Stephen John Smoogen wrote: > >> > >>> I would first get the cobbler && koan from EPEL testing. They are > >>> updated and fix a lot of bugs from the EPEL-stable version. > >>> > >>> > >>> > >>> > >>> > >>> > >> Exactly what I was going to say :) > >> > >> 0.6.0 is out now, and I'll be releasing 0.6.1 (a few more refinements) > >> later this week. > >> 0.4.x is pretty old. > >> > >> --Michael > >> > >> > > > > Thanks, Michael. Through trail and error I discovered that Cobbler/Koan > > did not like me using a DNS name for the system name to build a Xen VM. > > When I used a MAC address for the system it worked just fine. I should > > have yielded to the advice in the man pages and just used a MAC. > > > > You don't have to do that in 0.6.x :) > > > Also, thanks for the heads up on 0.6.0. Along with the bug fixes it > > looks to have some neat features including an increased number of > > builtin variables, along with the kickstart snippets that I can use for > > kickstart templates. When will 0.6.0 be moved over to the standard EPEL > > repo? > > > > Everyone just pull the RPM's out of EPEL testing. For what it's worth, > I'm using EPEL for "stable" releases > only, and "testing" builds of cobbler (like the 0.5.x series) will never > be pushed there. > > From recent discussions on #epel (IRC), epel is probably going to push > quarterly, which means I can't update > periodically for bugfixes and improvements as I would like. So, as a > result, what is in EPEL stable is currently very old and I have > no way of updating it. > Also to let people know... EPEL people are looking at how to handle this better in the future. There are lots of different potential customers to EPEL (people who want old stable stuff forever) and people who want to be able to get newest stable updates. The original idea was to try and solve both by locking down a release for a 'quarter' so that 5.0.0 tree would be stable for people wanting old stuff and 5.1.0 tree would be stable on the next release, etc.. growing pains. > If you don't want the rest of your system to get packages out of EPEL > "testing" for whatever reason, you can just enable the repo, > install cobbler, and disable the repo. As long as you are on the list > you'll get the release announcements so that may be workable. > -- Stephen J Smoogen. -- CSIRT/Linux System Administrator How far that little candle throws his beams! So shines a good deed in a naughty world. = Shakespeare. "The Merchant of Venice" From pnixon at gmail.com Tue Aug 28 19:06:04 2007 From: pnixon at gmail.com (Patrick Nixon) Date: Tue, 28 Aug 2007 15:06:04 -0400 Subject: [et-mgmt-tools] distro not updating on an IP Address change Message-ID: Hey all, I just downloaded and configured my laptop to be a cobbler server. One of the requirements is to be able to take this laptop, plug it into the appropriate network switch, update the IP Address, and have everything work. I've worked out updating /var/lib/cobbler/settings and /etc/cobbler/dhcp.template and everything is updated with the new IP Address once I run cobbler sync. Except... /var/lib/cobbler/distros. It's retaining the old IP Address in ks_meta tree item and the source_repos item. Any suggestions on the best method to manage these two items in the file? Thanks, --Pat From mdehaan at redhat.com Tue Aug 28 19:09:46 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 28 Aug 2007 15:09:46 -0400 Subject: [et-mgmt-tools] distro not updating on an IP Address change In-Reply-To: References: Message-ID: <46D472FA.3080105@redhat.com> Patrick Nixon wrote: > Hey all, > I just downloaded and configured my laptop to be a cobbler server. > > One of the requirements is to be able to take this laptop, plug it > into the appropriate network switch, update the IP Address, and have > everything work. > > I've worked out updating /var/lib/cobbler/settings and > /etc/cobbler/dhcp.template and everything is updated with the new IP > Address once I run cobbler sync. > > Except... /var/lib/cobbler/distros. > > It's retaining the old IP Address in ks_meta tree item and the > source_repos item. > > Any suggestions on the best method to manage these two items in the file? > You could write a script that read the value of /var/lib/cobbler/settings ($server parameter) and then updated the contents of /var/lib/cobbler/distros appropriately. It is a perfectly reasonable request to be able to have a value like "$server" inside the "tree" parameter and have it parse out though. I'll see about adding that for 0.6.1 -- we could only use this behavior for distros we imported with "cobbler import" and didn't use the "--available-as" parameter, but that would cover a very large amount of distros. If implemented as mentioned above, all you would have to do is change /var/lib/cobbler/settings and re-sync. > Thanks, > --Pat > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Tue Aug 28 19:14:38 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 28 Aug 2007 15:14:38 -0400 Subject: [et-mgmt-tools] distro not updating on an IP Address change In-Reply-To: <46D472FA.3080105@redhat.com> References: <46D472FA.3080105@redhat.com> Message-ID: <46D4741E.1090101@redhat.com> Michael DeHaan wrote: > Patrick Nixon wrote: >> >> > > You could write a script that read the value of > /var/lib/cobbler/settings ($server parameter) > and then updated the contents of /var/lib/cobbler/distros appropriately. > > It is a perfectly reasonable request to be able to have a value like > "$server" inside the "tree" parameter and have > it parse out though. > I'll see about adding that for 0.6.1 -- we could only use this > behavior for distros > we imported with "cobbler import" and didn't use the "--available-as" > parameter, but that would > cover a very large amount of distros. > > If implemented as mentioned above, all you would have to do is change > /var/lib/cobbler/settings and re-sync. In case it wasn't obvious, I'm shooting for 0.6.1 to release Friday, so it won't be that long if you want to wait. Short term, just hack /var/lib/cobbler/distros, re-run cobbler sync, and things will work. From pnixon at gmail.com Tue Aug 28 19:20:56 2007 From: pnixon at gmail.com (Patrick Nixon) Date: Tue, 28 Aug 2007 15:20:56 -0400 Subject: [et-mgmt-tools] distro not updating on an IP Address change In-Reply-To: <46D4741E.1090101@redhat.com> References: <46D472FA.3080105@redhat.com> <46D4741E.1090101@redhat.com> Message-ID: Thanks for the answer and the follow up. --Pat On 8/28/07, Michael DeHaan wrote: > Michael DeHaan wrote: > > Patrick Nixon wrote: > >> > >> > > > > You could write a script that read the value of > > /var/lib/cobbler/settings ($server parameter) > > and then updated the contents of /var/lib/cobbler/distros appropriately. > > > > It is a perfectly reasonable request to be able to have a value like > > "$server" inside the "tree" parameter and have > > it parse out though. > > I'll see about adding that for 0.6.1 -- we could only use this > > behavior for distros > > we imported with "cobbler import" and didn't use the "--available-as" > > parameter, but that would > > cover a very large amount of distros. > > > > If implemented as mentioned above, all you would have to do is change > > /var/lib/cobbler/settings and re-sync. > > In case it wasn't obvious, I'm shooting for 0.6.1 to release Friday, so > it won't be that long if you want to wait. > Short term, just hack /var/lib/cobbler/distros, re-run cobbler sync, and > things will work. > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Tue Aug 28 19:46:10 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Tue, 28 Aug 2007 15:46:10 -0400 Subject: [et-mgmt-tools] distro not updating on an IP Address change In-Reply-To: References: <46D472FA.3080105@redhat.com> <46D4741E.1090101@redhat.com> Message-ID: <46D47B82.2080106@redhat.com> Patrick Nixon wrote: > Thanks for the answer and the follow up. > > --Pat > I can do better :) The feature is checked into git now. Basically new cobbler import commands will "just work". For older ones, go into your /var/lib/cobbler/distros and replace the server addresses with "@@server@@" How this works, for templating enthusiasts, is that for every string variable accessible in what we normally feed to Cheetah, we replace @@variable@@ with it's value /after/ we complete our first pass of Cheetah rendering. This second pass does not use Cheetah to avoid the problem of having to escape anything differently -- it just uses a basic string.replace() loop. The additional level of templating is very basic and I would encourage people to pretend it doesn't exist rather than making anything depend on it. Here's the diff: http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h=e29737d38da69887368fc0acd3666fd78e842a80 > On 8/28/07, Michael DeHaan wrote: > >> Michael DeHaan wrote: >> >>> Patrick Nixon wrote: >>> >>>> >>> You could write a script that read the value of >>> /var/lib/cobbler/settings ($server parameter) >>> and then updated the contents of /var/lib/cobbler/distros appropriately. >>> >>> It is a perfectly reasonable request to be able to have a value like >>> "$server" inside the "tree" parameter and have >>> it parse out though. >>> I'll see about adding that for 0.6.1 -- we could only use this >>> behavior for distros >>> we imported with "cobbler import" and didn't use the "--available-as" >>> parameter, but that would >>> cover a very large amount of distros. >>> >>> If implemented as mentioned above, all you would have to do is change >>> /var/lib/cobbler/settings and re-sync. >>> >> In case it wasn't obvious, I'm shooting for 0.6.1 to release Friday, so >> it won't be that long if you want to wait. >> Short term, just hack /var/lib/cobbler/distros, re-run cobbler sync, and >> things will work. >> >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> >> > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From jason at 3dogs.us Wed Aug 29 03:01:11 2007 From: jason at 3dogs.us (Jason Hartley) Date: Tue, 28 Aug 2007 23:01:11 -0400 (EDT) Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <80d7e4090708281114m61e101eeh59dfa1c469918ced@mail.gmail.com> References: <37675.192.168.0.10.1188003146.squirrel@www.3dogs.us> <80d7e4090708251705q723d8daeu9c5bb0143e439a5e@mail.gmail.com> <46D2DFB2.5000501@redhat.com> <49304.192.168.0.10.1188258741.squirrel@www.3dogs.us> <46D43029.4050701@redhat.com> <80d7e4090708281114m61e101eeh59dfa1c469918ced@mail.gmail.com> Message-ID: <33385.192.168.0.10.1188356471.squirrel@www.3dogs.us> On Tue, August 28, 2007 2:14 pm, Stephen John Smoogen wrote: > On 8/28/07, Michael DeHaan wrote: >> Jason Hartley wrote: >> > On Mon, August 27, 2007 10:29 am, Michael DeHaan wrote: >> > >> >> Stephen John Smoogen wrote: >> >> >> >>> I would first get the cobbler && koan from EPEL testing. They are >> >>> updated and fix a lot of bugs from the EPEL-stable version. >> >>> >> >>> >> >>> >> >>> >> >>> >> >>> >> >> Exactly what I was going to say :) >> >> >> >> 0.6.0 is out now, and I'll be releasing 0.6.1 (a few more >> refinements) >> >> later this week. >> >> 0.4.x is pretty old. >> >> >> >> --Michael >> >> >> >> >> > >> > Thanks, Michael. Through trail and error I discovered that >> Cobbler/Koan >> > did not like me using a DNS name for the system name to build a Xen >> VM. >> > When I used a MAC address for the system it worked just fine. I >> should >> > have yielded to the advice in the man pages and just used a MAC. >> > >> >> You don't have to do that in 0.6.x :) >> >> > Also, thanks for the heads up on 0.6.0. Along with the bug fixes it >> > looks to have some neat features including an increased number of >> > builtin variables, along with the kickstart snippets that I can use >> for >> > kickstart templates. When will 0.6.0 be moved over to the standard >> EPEL >> > repo? >> > >> >> Everyone just pull the RPM's out of EPEL testing. For what it's worth, >> I'm using EPEL for "stable" releases >> only, and "testing" builds of cobbler (like the 0.5.x series) will never >> be pushed there. >> >> From recent discussions on #epel (IRC), epel is probably going to push >> quarterly, which means I can't update >> periodically for bugfixes and improvements as I would like. So, as a >> result, what is in EPEL stable is currently very old and I have >> no way of updating it. >> > > Also to let people know... EPEL people are looking at how to handle > this better in the future. There are lots of different potential > customers to EPEL (people who want old stable stuff forever) and > people who want to be able to get newest stable updates. The original > idea was to try and solve both by locking down a release for a > 'quarter' so that 5.0.0 tree would be stable for people wanting old > stuff and 5.1.0 tree would be stable on the next release, etc.. > growing pains. > > >> If you don't want the rest of your system to get packages out of EPEL >> "testing" for whatever reason, you can just enable the repo, >> install cobbler, and disable the repo. As long as you are on the list >> you'll get the release announcements so that may be workable. >> > > > -- > Stephen J Smoogen. -- CSIRT/Linux System Administrator > How far that little candle throws his beams! So shines a good deed > in a naughty world. = Shakespeare. "The Merchant of Venice" > Thanks, for all the great information and help. I will checking out version 6 tonight! Regards, Jason From zxvdr.au at gmail.com Wed Aug 29 03:20:14 2007 From: zxvdr.au at gmail.com (David Robinson) Date: Wed, 29 Aug 2007 13:20:14 +1000 Subject: [et-mgmt-tools] more gui improvements In-Reply-To: <46CE010D.4080908@redhat.com> References: <46CE010D.4080908@redhat.com> Message-ID: <46D4E5EE.9090909@gmail.com> Hugh Brock wrote: > Good evening everyone. I have just pushed a change to virt-manager that > adds "action" icons to the list of vms and connections in the manager > window. At the moment there are icons for "new" and "delete" for each > connection, and a "delete" icon for each inactive VM. Be careful where > you click; I still need to add an "are you sure" dialog for the delete > icon! Hi Hugh, I attached a patch to BZ 253144 that adds the "are you sure" dialog to the delete button. The patch applies against virt-manager in RHEL5.1 beta. Cheers, Dave From Kevin.Maute at WPAFB.AF.MIL Wed Aug 29 12:11:11 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Wed, 29 Aug 2007 08:11:11 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Folks, Be kind to a newbie. Michael has been a big help but now I need to expand my net. I am moving at a pretty good pace now that I found cobbler but I am a bit stuck. I am trying to get kickstart to recognize the RPMs in the Cluster dir of RHEL5 Sever. It shows up correctly in my local repo but everytime I add it to my ks package list I get the dreaded Abort/Continue dialog. I realize this might be better asked on the kickstart list BUT I thought I would start here since I see a lot of virtualization activity which from a kickstart perspective is probably similar... Thanks, Kevin -------------------------------------------- Kevin Maute Sr. Linux/Unix System Administrator RCF Information Systems AFRL/VAOO Contractor 937.255.1810 kevin.maute at wpafb.af.mil -------------- next part -------------- An HTML attachment was scrubbed... URL: From oehmes at de.ibm.com Wed Aug 29 12:44:16 2007 From: oehmes at de.ibm.com (Sven Oehme) Date: Wed, 29 Aug 2007 14:44:16 +0200 Subject: [Qemu-devel] Re: [et-mgmt-tools] Image Corruption Possible with qemu and qemu-kvm In-Reply-To: <46D384AE.9080509@qumranet.com> References: <46D2EC13.8020005@bppiac.hu> <1188232650.25884.100.camel@localhost.localdomain> <20070827202748.GK9043@redhat.com> <1188248558.21696.1.camel@squirrel> <46D384AE.9080509@qumranet.com> Message-ID: qemu-devel-bounces+oehmes=de.ibm.com at nongnu.org wrote on 08/28/2007 04:13:02 AM: > Anthony Liguori wrote: > >> In the scenario you mention, libvirt should probably do a sanity check for > >> this before letting you start the guest. libvirt already supports the idea > >> of 'shared' disk images where two or more guests can be > optionally configured > >> to have write access - basically assumes the admin requesting sharing knows > >> what they're doing. > >> > > > > I think this is the right level myself. Advisory locks work okay but > > not all filesystems support them. It's particularly nasty when you have > > a clustered filesystem in the host. I think it would do more harm than > > good to have a feature like that was supposed to provide a safe-guard > > but then frequently didn't work. > > > > There's still the unmanaged use case to worry about. I think qemu can > default to advisory locking, and management tools can do their own > locking and always override qemu. > > It's too easy to kill an image by starting up another instance right now. > i agree default should be advisory locking and a switch to disable it .. would that be hard to implement ? thanks. Sven -------------- next part -------------- An HTML attachment was scrubbed... URL: From hbrock at redhat.com Wed Aug 29 13:03:37 2007 From: hbrock at redhat.com (Hugh Brock) Date: Wed, 29 Aug 2007 09:03:37 -0400 Subject: [et-mgmt-tools] more gui improvements In-Reply-To: <46D4E5EE.9090909@gmail.com> References: <46CE010D.4080908@redhat.com> <46D4E5EE.9090909@gmail.com> Message-ID: <46D56EA9.4010407@redhat.com> David Robinson wrote: > Hugh Brock wrote: >> Good evening everyone. I have just pushed a change to virt-manager >> that adds "action" icons to the list of vms and connections in the >> manager window. At the moment there are icons for "new" and "delete" >> for each connection, and a "delete" icon for each inactive VM. Be >> careful where you click; I still need to add an "are you sure" dialog >> for the delete icon! > > Hi Hugh, > > I attached a patch to BZ 253144 that adds the "are you sure" dialog to > the delete button. The patch applies against virt-manager in RHEL5.1 beta. > Heh... well, I'm not sure we really need it in the 5.1 version, because the delete button there is fairly obvious and located separately at the bottom of the window. I would consider it for 5.2 however (patch deadline for 5.1 is way past now). Friday I committed an "are you sure" dialog to the tip, so unfortunately that is now too late! However we do need one that pops up if you click the "create" button on a remote connection and says "Sorry the create feature is not yet available for remote connections..." You can check if a connection is remote by calling the vmmConnection.is_remote() method. Thanks for the effort, hopefully I can get something of yours in... --Hugh -- Red Hat Virtualization Group http://redhat.com/virtualization Hugh Brock | virt-manager http://virt-manager.org hbrock at redhat.com | virtualization library http://libvirt.org From berrange at redhat.com Wed Aug 29 13:55:54 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Wed, 29 Aug 2007 14:55:54 +0100 Subject: [et-mgmt-tools] more gui improvements In-Reply-To: <46D56EA9.4010407@redhat.com> References: <46CE010D.4080908@redhat.com> <46D4E5EE.9090909@gmail.com> <46D56EA9.4010407@redhat.com> Message-ID: <20070829135554.GA4330@redhat.com> On Wed, Aug 29, 2007 at 09:03:37AM -0400, Hugh Brock wrote: > David Robinson wrote: > >Hugh Brock wrote: > >>Good evening everyone. I have just pushed a change to virt-manager > >>that adds "action" icons to the list of vms and connections in the > >>manager window. At the moment there are icons for "new" and "delete" > >>for each connection, and a "delete" icon for each inactive VM. Be > >>careful where you click; I still need to add an "are you sure" dialog > >>for the delete icon! > > > >Hi Hugh, > > > >I attached a patch to BZ 253144 that adds the "are you sure" dialog to > >the delete button. The patch applies against virt-manager in RHEL5.1 beta. > > > Heh... well, I'm not sure we really need it in the 5.1 version, because > the delete button there is fairly obvious and located separately at the > bottom of the window. I would consider it for 5.2 however (patch > deadline for 5.1 is way past now). > > Friday I committed an "are you sure" dialog to the tip, so unfortunately > that is now too late! However we do need one that pops up if you click > the "create" button on a remote connection and says "Sorry the create > feature is not yet available for remote connections..." You can check if > a connection is remote by calling the vmmConnection.is_remote() method. I added the latter at about 11pm last night :-) Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From mdehaan at redhat.com Wed Aug 29 14:49:50 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 10:49:50 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <46D5878E.4050002@redhat.com> Maute, Kevin P CTR AFRL/VAOO wrote: > > Folks, > > Be kind to a newbie. Michael has been a big help but now I need to > expand my net. > > I am moving at a pretty good pace now that I found cobbler but I am a > bit stuck. I am trying to get kickstart to recognize the RPMs in the > Cluster dir of RHEL5 Sever. It shows up correctly in my local repo but > everytime I add it to my ks package list I get the dreaded > Abort/Continue dialog. I realize this might be better asked on the > kickstart list BUT I thought I would start here since I see a lot of > virtualization activity which from a kickstart perspective is probably > similar? > This sounds like a good question for this list to me, especially if it involves setting up RHEL 5 with your Cobbler boot server. Quick question -- Did you start your configuration with an DVD import using "cobbler import" ? If so, cobbler should be keeping track of the individual "split" repos for RHEL5, and making them available in the kickstart file. It should have also ran createrepo on each split repository such that it has correct yum metadata. Basically if you look at the rendered kickstart file in /var/www/cobbler/kickstarts/$name (this is the output of cobbler, not the input), you should see a couple of extra "repo" lines that represent the additional repos that have been attached. If you are not using a kickstart template based off one of the ones that shipped in cobbler (/etc/cobbler/*.ks) you may be missing the directive ($yum_repo_stanza) that allows for attaching of these associated repos automatically. To see what kickstart template you are using, run "cobbler profile report $profilename". --Michael > Thanks, > Kevin > > > *--------------------------------------------* > *Kevin Maute* > Sr. Linux/Unix System Administrator > RCF Information Systems > AFRL/VAOO Contractor > 937.255.1810 > kevin.maute at wpafb.af.mil > > ------------------------------------------------------------------------ > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From Daniel.Hennessey at aah.co.uk Wed Aug 29 16:18:17 2007 From: Daniel.Hennessey at aah.co.uk (Hennessey Daniel) Date: Wed, 29 Aug 2007 17:18:17 +0100 Subject: [et-mgmt-tools] Xen System builds using Koan Message-ID: <03BB5BD4CF43594B9542DCCD4E7989900B046CB2@GBW607SC0054.GB-WS.net> In my uneducated opinion it looks like you are unable to make a connection to your cobbler server ("hydra") on port 25151. HTH Cheers Dan -----Original Message----- From: et-mgmt-tools-bounces at redhat.com [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Jason Hartley Sent: 25 August 2007 01:52 To: et-mgmt-tools at redhat.com Subject: [et-mgmt-tools] Xen System builds using Koan First off, I really like the design of Cobbler and Koan. I especially like the kickstart templates and the flexibility that the templates give you. The only problem is that I am having trouble with the issue below and would appreciate any kind of help. I have a Centos 5 system that I have setup as a Xen host. I have setup the EPEL repository and used it to install Cobbler version 0.4.8-1,along with koan 0.4.0-1. I have successfully used the tutorial at http://cobbler.et.redhat.com/cobbler-import.php with a few modifications to build a Xen guest using the default profile for Centos 5 that is created when you import from the Centos 5 DVD. When I attempt to build from a system definition that I have built from the default profile I get the below error. I have started to dig through the code, but would like some direction in debugging the problem and maybe a push in the right direction on what might be wrong. [root at hydra ~]# koan --server=hydra --system=sirrush --virt - fetching configuration for system: (sirrush) Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/koan/app.py", line 468, in get_profile_xmlrpc return self.xmlrpc_server.get_profile_for_koan(profile_name) File "/usr/lib/python2.4/xmlrpclib.py", line 1096, in __call__ return self.__send(self.__name, args) File "/usr/lib/python2.4/xmlrpclib.py", line 1383, in __request verbose=self.__verbose File "/usr/lib/python2.4/xmlrpclib.py", line 1147, in request return self._parse_response(h.getfile(), sock) File "/usr/lib/python2.4/xmlrpclib.py", line 1286, in _parse_response return u.close() File "/usr/lib/python2.4/xmlrpclib.py", line 744, in close raise Fault(**self._stack[0]) Fault: File "/usr/bin/koan", line 38, in ? sys.exit(app.main() or 0) File "/usr/lib/python2.4/site-packages/koan/app.py", line 108, in main k.run() File "/usr/lib/python2.4/site-packages/koan/app.py", line 175, in run self.do_virt() File "/usr/lib/python2.4/site-packages/koan/app.py", line 306, in do_virt return self.do_net_install("/var/lib/xen",after_download) File "/usr/lib/python2.4/site-packages/koan/app.py", line 266, in do_net_install profile_data = self.get_system_xmlrpc(self.system) File "/usr/lib/python2.4/site-packages/koan/app.py", line 538, in get_system_xmlrpc profile_data = self.get_profile_xmlrpc(self.safe_load(system_data,'profile')) File "/usr/lib/python2.4/site-packages/koan/app.py", line 471, in get_profile_xmlrpc self.connect_fail() File "/usr/lib/python2.4/site-packages/koan/app.py", line 454, in connect_fail print self.trace_me() File "/usr/lib/python2.4/site-packages/koan/app.py", line 449, in trace_me x = traceback.extract_stack() Could not communicate with hydra:25151 [root at hydra ~]# Thanks, Jason _______________________________________________ et-mgmt-tools mailing list et-mgmt-tools at redhat.com https://www.redhat.com/mailman/listinfo/et-mgmt-tools ************************************************************************ DISCLAIMER The information contained in this e-mail is confidential and is intended for the recipient only. If you have received it in error, please notify us immediately by reply e-mail and then delete it from your system. Please do not copy it or use it for any other purposes, or disclose the content of the e-mail to any other person or store or copy the information in any medium. The views contained in this e-mail are those of the author and not necessarily those of Admenta UK Group. Admenta UK plc is a company incorporated in England and Wales under company number 3011757 and whose registered office is at Sapphire Court, Walsgrave Triangle, Coventry CV2 2TX ************************************************************************ From mdehaan at redhat.com Wed Aug 29 16:31:57 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 12:31:57 -0400 Subject: [et-mgmt-tools] Xen System builds using Koan In-Reply-To: <03BB5BD4CF43594B9542DCCD4E7989900B046CB2@GBW607SC0054.GB-WS.net> References: <03BB5BD4CF43594B9542DCCD4E7989900B046CB2@GBW607SC0054.GB-WS.net> Message-ID: <46D59F7D.1020607@redhat.com> Hennessey Daniel wrote: > In my uneducated opinion it looks like you are unable to make a > connection to your cobbler server ("hydra") on port 25151. > > HTH > > Cheers > > Dan Yep. Upgrading would be a good idea regardless. cobblerd is much more reliable in new versions. The init scripts also are improved in 0.6.1 (out tomorrow) which keeps it from being slain by logrotate. From rjones at redhat.com Wed Aug 29 17:03:35 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Wed, 29 Aug 2007 18:03:35 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-top 0.3.2.5 released Message-ID: <46D5A6E7.2070608@redhat.com> I'm pleased to announce a new release of virt-top, which is a utility like 'top' that works for virtual machines. Many keys and command line options are the same as for ordinary top. Thanks to all the people who have been playing with it and sending me feedback. This version adds block device and network statistics (if supported by the hypervisor and libvirt). This version also fixes a bug where it would fail silently if unable to contact the hypervisor. Main page: http://et.redhat.com/~rjones/virt-top/ Source and binaries for several architectures available from: http://et.redhat.com/~rjones/virt-top/download.html If you have Fedora 7, this will probably work: yum --enablerepo=development install virt-top Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From drew.einhorn at gmail.com Wed Aug 29 17:17:26 2007 From: drew.einhorn at gmail.com (drew einhorn) Date: Wed, 29 Aug 2007 11:17:26 -0600 Subject: [et-mgmt-tools] Snippets Message-ID: It would be convenient if snippets could reference other snippets. And if snippets could set templating variables as well as reference them. But it's not necessary. No matter what you do we will always want more. -- Drew Einhorn -------------- next part -------------- An HTML attachment was scrubbed... URL: From Kevin.Maute at WPAFB.AF.MIL Wed Aug 29 17:46:49 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Wed, 29 Aug 2007 13:46:49 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5878E.4050002@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Michael, When you say DVD import do you literally mean DVD? I created a HD filesystem which was the standard copy of all the Server CDs DLed from RHN. I actually DID get a DVD from the RHCT class I took last month :-). And, yes I did do a cobbler import as we discussed previously. Also, I removed all the distros and profiles that cobbler created (using cobbler {distro,profile} remove!) and recreated them. I just noticed the dates in /var/www/cobbler/ks_mirror has old dates for the repodata & repoview dirs. Is this copying the repo metadata from my local repo or creating new dirs? Lastly, I did retain the variables in the original ks template from /etc/cobbler. In the "output" ks file I see the comment about the add'l repos but the $yum_repo_stanza appears to be empty as there is no value on the next line. BTW, $yum_config_stanza does expand to "wget ..." so it is expanding "some" vars. I assume $kickstart_done is a sentinel variable as it also does not generate any strings. I tried again (@clustering) and still get grief. Stepping back, what I would like is not only the clustering and later vt support but also EPEL as well. Interestingly, everytime I do a distro and profile I DO get the correspondnig xen one as well. Currently, my local repo is: /home/repos | -- EPEL5 | --RHEL5 (all Server discs which include Clustering* and VT directories/RPMs) What is the recommended way to (recreate) my repo(s) and get cobbler/kickstart to play well with others? Thanks, Kevin > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Wednesday, August 29, 2007 10:50 AM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Maute, Kevin P CTR AFRL/VAOO wrote: > > > > Folks, > > > > Be kind to a newbie. Michael has been a big help but now I need to > > expand my net. > > > > I am moving at a pretty good pace now that I found cobbler > but I am a > > bit stuck. I am trying to get kickstart to recognize the > RPMs in the > > Cluster dir of RHEL5 Sever. It shows up correctly in my > local repo but > > everytime I add it to my ks package list I get the dreaded > > Abort/Continue dialog. I realize this might be better asked on the > > kickstart list BUT I thought I would start here since I see > a lot of > > virtualization activity which from a kickstart perspective > is probably > > similar... > > > > This sounds like a good question for this list to me, > especially if it > involves setting up RHEL 5 with your Cobbler boot server. > > Quick question -- Did you start your configuration with an DVD import > using "cobbler import" ? If so, cobbler should be keeping > track of the > individual "split" repos for RHEL5, and making them available in the > kickstart file. It should have also ran createrepo on each split > repository such that it has correct yum metadata. > > Basically if you look at the rendered kickstart file in > /var/www/cobbler/kickstarts/$name (this is the output of cobbler, not > the input), you should see a couple of extra "repo" > lines that represent the additional repos that have been attached. > > If you are not using a kickstart template based off one of > the ones that > shipped in cobbler (/etc/cobbler/*.ks) you may be missing the > directive > ($yum_repo_stanza) that allows > for attaching of these associated repos automatically. To see what > kickstart template you are using, run "cobbler profile report > $profilename". > > --Michael > > > > Thanks, > > Kevin > > > > > > *--------------------------------------------* > > *Kevin Maute* > > Sr. Linux/Unix System Administrator > > RCF Information Systems > > AFRL/VAOO Contractor > > 937.255.1810 > > kevin.maute at wpafb.af.mil > > > > > -------------------------------------------------------------- > ---------- > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Wed Aug 29 17:54:06 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 13:54:06 -0400 Subject: [et-mgmt-tools] Snippets In-Reply-To: References: Message-ID: <46D5B2BE.4070707@redhat.com> drew einhorn wrote: > It would be convenient if snippets could reference other snippets. > And if snippets could set templating variables as well as reference them. > > But it's not necessary. Ah, but they can :) Cheetah has a "#include" directive. It turns out that cobbler snippets are really little more than Cheetah #include's, so if you use #include instead of SNIPPET:: inside your individual snippets, that all works. > > No matter what you do we will always want more. This is a good thing :) > > -- > Drew Einhorn > ------------------------------------------------------------------------ > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools From mdehaan at redhat.com Wed Aug 29 18:20:19 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 14:20:19 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <46D5B8E3.1090800@redhat.com> Maute, Kevin P CTR AFRL/VAOO wrote: > Michael, > > When you say DVD import do you literally mean DVD? I created a HD > filesystem which was the standard copy of all the Server CDs DLed from > RHN. I actually DID get a DVD from the RHCT class I took last month > :-). And, yes I did do a cobbler import as we discussed previously. > Either one works as long as the directory layout is basically the same. > Also, I removed all the distros and profiles that cobbler created (using > cobbler {distro,profile} remove!) and recreated them. I just noticed > the dates in /var/www/cobbler/ks_mirror has old dates for the repodata & > repoview dirs. Is this copying the repo metadata from my local repo or > creating new dirs? > Those might be the creation dates of the files on the CD media. Rsync is being invoked with -a. > Lastly, I did retain the variables in the original ks template from > /etc/cobbler. In the "output" ks file I see the comment about the add'l > repos but the $yum_repo_stanza appears to be empty as there is no value > on the next line. BTW, $yum_config_stanza does expand to "wget ..." so > it is expanding "some" vars. I assume $kickstart_done is a sentinel > variable as it also does not generate any strings. > > It seems like the import process did not find the cluster repos from the DVD import. If that is not working, one way to do get the same content would be to just add them as regular yum repositories in cobbler. cobbler repo add --name=rhel5i386clustering --mirror=/path/to/DVD/source/Clustering # repeat for other source repos cobbler repo add --name=epel5i386testing --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/i386/ cobbler reposync cobbler repo edit --name=rhel5i386clustering --keep-updated=0 cobbler profile edit --name=rhel-5-i386 --repos="rhel5i386clustering rhel5i386epeltesting" cobbler sync Notice though that there are some fixes to repo management in 0.6.1 (especially related to non-http/ftp mirrors), so that's something to be aware of. > I tried again (@clustering) and still get grief. > > Stepping back, what I would like is not only the clustering and later vt > support but also EPEL as well. Interestingly, everytime I do a distro > and profile I DO get the correspondnig xen one as well. > See above on the epel bits. Those work via "cobbler repo add" as do any arbitrary yum repo. Admitedly I would have expected your RHEL 5 clustering repo to be added automatically in the kickstarts and I'll do some testing on my end to make sure that feature still works. > >> -----Original Message----- >> From: et-mgmt-tools-bounces at redhat.com >> [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan >> Sent: Wednesday, August 29, 2007 10:50 AM >> To: Fedora/Linux Management Tools >> Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question >> >> Maute, Kevin P CTR AFRL/VAOO wrote: >> >>> Folks, >>> >>> Be kind to a newbie. Michael has been a big help but now I need to >>> expand my net. >>> >>> I am moving at a pretty good pace now that I found cobbler >>> >> but I am a >> >>> bit stuck. I am trying to get kickstart to recognize the >>> >> RPMs in the >> >>> Cluster dir of RHEL5 Sever. It shows up correctly in my >>> >> local repo but >> >>> everytime I add it to my ks package list I get the dreaded >>> Abort/Continue dialog. I realize this might be better asked on the >>> kickstart list BUT I thought I would start here since I see >>> >> a lot of >> >>> virtualization activity which from a kickstart perspective >>> >> is probably >> >>> similar... >>> >>> >> This sounds like a good question for this list to me, >> especially if it >> involves setting up RHEL 5 with your Cobbler boot server. >> >> Quick question -- Did you start your configuration with an DVD import >> using "cobbler import" ? If so, cobbler should be keeping >> track of the >> individual "split" repos for RHEL5, and making them available in the >> kickstart file. It should have also ran createrepo on each split >> repository such that it has correct yum metadata. >> >> Basically if you look at the rendered kickstart file in >> /var/www/cobbler/kickstarts/$name (this is the output of cobbler, not >> the input), you should see a couple of extra "repo" >> lines that represent the additional repos that have been attached. >> >> If you are not using a kickstart template based off one of >> the ones that >> shipped in cobbler (/etc/cobbler/*.ks) you may be missing the >> directive >> ($yum_repo_stanza) that allows >> for attaching of these associated repos automatically. To see what >> kickstart template you are using, run "cobbler profile report >> $profilename". >> >> --Michael >> >> >> >>> Thanks, >>> Kevin >>> >>> >>> *--------------------------------------------* >>> *Kevin Maute* >>> Sr. Linux/Unix System Administrator >>> RCF Information Systems >>> AFRL/VAOO Contractor >>> 937.255.1810 >>> kevin.maute at wpafb.af.mil >>> >>> >>> >> -------------------------------------------------------------- >> ---------- >> >>> _______________________________________________ >>> et-mgmt-tools mailing list >>> et-mgmt-tools at redhat.com >>> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >>> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> >> > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From tomg at mcclatchyinteractive.com Wed Aug 29 18:45:43 2007 From: tomg at mcclatchyinteractive.com (Tom Georgoulias) Date: Wed, 29 Aug 2007 14:45:43 -0400 Subject: [et-mgmt-tools] ANNOUNCE: virt-top 0.3.2.5 released In-Reply-To: <46D5A6E7.2070608@redhat.com> References: <46D5A6E7.2070608@redhat.com> Message-ID: <46D5BED7.4000100@mcclatchyinteractive.com> Richard W.M. Jones wrote: > I'm pleased to announce a new release of virt-top, which is a utility > like 'top' that works for virtual machines. Many keys and command line > options are the same as for ordinary top. > Source and binaries for several architectures available from: > http://et.redhat.com/~rjones/virt-top/download.html The RHEL5 RPM on the download page is for the RHEL5.1 beta, right? -- Tom Georgoulias Sr. Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com From mdehaan at redhat.com Wed Aug 29 18:57:08 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 14:57:08 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5B8E3.1090800@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5B8E3.1090800@redhat.com> Message-ID: <46D5C184.3060508@redhat.com> Michael DeHaan wrote: > Maute, Kevin P CTR AFRL/VAOO wrote: [snip] (RHEL5-repo-question) [/snip] Ok, so I found the problem, and here's the fix (which will be in 0.6.1): http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h=ada8263d700431bc6c731cf7faf374f24dd0e683 If you re-run the import with the changes above, you should see the following repos auto-attached. repo --name=ClusterStorage --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/ClusterStorage repo --name=VT --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/VT repo --name=Cluster --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/Cluster repo --name=Server --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/Server This works for me and I'll do some more testing with other distros to see if they continue to work as advertised. If you turn on the additional "mirror" setting in the settings file, you'll also see the associated wgets to set them up on the provisioned box. EPEL still needs to be added with "cobbler repo add" (see manpage and previous email). Keep in mind that, for now, you'll probably want to use EPEL testing so you can install a compatible koan on target boxes. --Michael From Kevin.Maute at WPAFB.AF.MIL Wed Aug 29 19:29:53 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Wed, 29 Aug 2007 15:29:53 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5C184.3060508@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021BABCA@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Excellent, I haven't been a beta tester (well, typing this from Outlook on XP, I guess I AM a beta tester) in a while. I am proceeding with your suggestion in the last post. Hopefully, 0.6.1 will still be coming out tomorrow even with the "noise" :) > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Wednesday, August 29, 2007 2:57 PM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Michael DeHaan wrote: > > Maute, Kevin P CTR AFRL/VAOO wrote: > [snip] (RHEL5-repo-question) [/snip] > > Ok, so I found the problem, and here's the fix (which will be > in 0.6.1): > > http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h= > ada8263d700431bc6c731cf7faf374f24dd0e683 > > If you re-run the import with the changes above, you should see the > following repos auto-attached. > > repo --name=ClusterStorage > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/o > s/ClusterStorage > repo --name=VT > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/VT > repo --name=Cluster > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/o > s/Cluster > repo --name=Server > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/Server > > This works for me and I'll do some more testing with other distros to > see if they continue to work as advertised. > > If you turn on the additional "mirror" setting in the settings file, > you'll also see the associated wgets to set them up on the > provisioned box. > > EPEL still needs to be added with "cobbler repo add" (see manpage and > previous email). Keep in mind that, for now, you'll probably > want to use EPEL testing so you can install a compatible koan > on target > boxes. > > --Michael > > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From Kevin.Maute at WPAFB.AF.MIL Wed Aug 29 19:58:07 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Wed, 29 Aug 2007 15:58:07 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5C184.3060508@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Success!!! One issue though, epel5*testing, it appeared to download everything fine from the fedora repo but seemed to choke with a traceback when it was populating the local repo. Not knowing the internals or python, I can't give any more input. I backed out that repo and am golden. Thanks, Kevin > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Wednesday, August 29, 2007 2:57 PM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Michael DeHaan wrote: > > Maute, Kevin P CTR AFRL/VAOO wrote: > [snip] (RHEL5-repo-question) [/snip] > > Ok, so I found the problem, and here's the fix (which will be > in 0.6.1): > > http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h= > ada8263d700431bc6c731cf7faf374f24dd0e683 > > If you re-run the import with the changes above, you should see the > following repos auto-attached. > > repo --name=ClusterStorage > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/o > s/ClusterStorage > repo --name=VT > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/VT > repo --name=Cluster > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/o > s/Cluster > repo --name=Server > --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/Server > > This works for me and I'll do some more testing with other distros to > see if they continue to work as advertised. > > If you turn on the additional "mirror" setting in the settings file, > you'll also see the associated wgets to set them up on the > provisioned box. > > EPEL still needs to be added with "cobbler repo add" (see manpage and > previous email). Keep in mind that, for now, you'll probably > want to use EPEL testing so you can install a compatible koan > on target > boxes. > > --Michael > > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Wed Aug 29 20:00:33 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 16:00:33 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <46D5D061.5090505@redhat.com> Maute, Kevin P CTR AFRL/VAOO wrote: > Success!!! One issue though, epel5*testing, it appeared to download > everything fine from the fedora repo but seemed to choke with a > traceback when it was populating the local repo. Not knowing the > internals or python, I can't give any more input. I backed out that > repo and am golden. > > Thanks, > Kevin > Good deal. If you can email me the traceback (cobbler output) and the command(s) that caused it that would be helpful. --Michael > >> -----Original Message----- >> From: et-mgmt-tools-bounces at redhat.com >> [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan >> Sent: Wednesday, August 29, 2007 2:57 PM >> To: Fedora/Linux Management Tools >> Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question >> >> Michael DeHaan wrote: >> >>> Maute, Kevin P CTR AFRL/VAOO wrote: >>> >> [snip] (RHEL5-repo-question) [/snip] >> >> Ok, so I found the problem, and here's the fix (which will be >> in 0.6.1): >> >> http://git.fedoraproject.org/?p=hosted/cobbler;a=commitdiff;h= >> ada8263d700431bc6c731cf7faf374f24dd0e683 >> >> If you re-run the import with the changes above, you should see the >> following repos auto-attached. >> >> repo --name=ClusterStorage >> --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/o >> s/ClusterStorage >> repo --name=VT >> --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/VT >> repo --name=Cluster >> --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/o >> s/Cluster >> repo --name=Server >> --baseurl=http://mdehaan.rdu.redhat.com/cobbler/ks_mirror/r5/os/Server >> >> This works for me and I'll do some more testing with other distros to >> see if they continue to work as advertised. >> >> If you turn on the additional "mirror" setting in the settings file, >> you'll also see the associated wgets to set them up on the >> provisioned box. >> >> EPEL still needs to be added with "cobbler repo add" (see manpage and >> previous email). Keep in mind that, for now, you'll probably >> want to use EPEL testing so you can install a compatible koan >> on target >> boxes. >> >> --Michael >> >> >> >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> >> > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Wed Aug 29 20:06:00 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 16:06:00 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5D061.5090505@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5D061.5090505@redhat.com> Message-ID: <46D5D1A8.9080900@redhat.com> Michael DeHaan wrote: > Maute, Kevin P CTR AFRL/VAOO wrote: >> Success!!! One issue though, epel5*testing, it appeared to download >> everything fine from the fedora repo but seemed to choke with a >> traceback when it was populating the local repo. Not knowing the >> internals or python, I can't give any more input. I backed out that >> repo and am golden. >> >> Thanks, >> Kevin > > Good deal. > > If you can email me the traceback (cobbler output) and the command(s) > that caused it that would be helpful. > > --Michael I just ran the following for EPEL testing and it worked for me (using 0.6.1): From mdehaan at redhat.com Wed Aug 29 20:08:30 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 29 Aug 2007 16:08:30 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5D1A8.9080900@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com> Message-ID: <46D5D23E.2050408@redhat.com> Michael DeHaan wrote: > Michael DeHaan wrote: >> Maute, Kevin P CTR AFRL/VAOO wrote: >>> Success!!! One issue though, epel5*testing, it appeared to download >>> everything fine from the fedora repo but seemed to choke with a >>> traceback when it was populating the local repo. Not knowing the >>> internals or python, I can't give any more input. I backed out that >>> repo and am golden. >>> >>> Thanks, >>> Kevin >> >> Good deal. >> >> If you can email me the traceback (cobbler output) and the command(s) >> that caused it that would be helpful. >> >> --Michael > > I just ran the following for EPEL testing and it worked for me (using > 0.6.1): > > Grr, thunderbird :) What I meant to say was ... I just ran the following for EPEL testing and it worked for me (using 0.6.1): [root at mdehaan cobbler]# cobbler repo add --name=epel5i386 --mirror=http://linux.nssl.noaa.gov/epel/testing/5Server/i386 [root at mdehaan cobbler]# cobbler repo sync You then of course would do: cobbler profile edit --name=nameofprofile --repos="epel5i386 otherrepo1 otherrepo2" etc. You would not have to repeat the repos that Cobbler pulled in from the import but would want to mirror your package updates. From berrange at redhat.com Wed Aug 29 22:57:29 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Wed, 29 Aug 2007 23:57:29 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-viewer 0.0.2 release Message-ID: <20070829225729.GG4330@redhat.com> I'm happy to announce an incremetal release of virt-viewer, version 0.0.2 is available for download from http://virt-manager.org/download/sources/virt-viewer/virt-viewer-0.0.2.tar.gz This release adds the ability to directly connect to remote consoles, to query a remote hypervisor using the libvirt daemon and to tunnel the console over SSH. It also adds support for providing the x509 certificate credentials for a VNC console using the TLS authentication extensions. Further information about this and other virtualization management tools is available from the project website http://virt-manager.org/ Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Wed Aug 29 23:02:18 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 30 Aug 2007 00:02:18 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virtinst release 0.300.0 Message-ID: <20070829230218.GH4330@redhat.com> I'm happy to announce that a new release of virtinst, version 0.300.0, is available for download from http://virt-manager.org/download/sources/virtinst/virtinst-0.300.0.tar.gz This release introduces the virt-image tool which provides the ability to create guests based on a pre-define image. The image provides metadata describing the required host virtualization capabilities, the recommend resource allocations and one or more pre-installed disk templates. From this is can instantiate new guests on any suitable hypervisor supported by libvirt. With the virt-install tool, all new fullyvirtualized guests now get a USB tablet device configured providing an easier to use pointer. Further improvements are made to input validation to catch user errors earlier. Use of the 'xm console' and 'vncviewer' external programs has been replaced with 'virsh console' and 'virt-viewer' Thanks to David Lutterkort for creating & documenting the virt-image tool, and to everyone who has contributing to this release through submitting bugs, writing patches & generally testing the code. Further information on this and other virtualization management tools is available from the project website" http://virt-manager.org/ Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Wed Aug 29 23:05:12 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 30 Aug 2007 00:05:12 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-manager 0.5.0 release Message-ID: <20070829230512.GI4330@redhat.com> I'm happy to announce the release of virt-manager version 0.5.0, available for download from http://virt-manager.org/download/sources/virt-manager/virt-manager-0.5.0.tar.gz This release introduces the ability to manage multiple remote machines, using either SSH+public keys, or TLS+x509 certificates to connect and authenticate. The main user interface is re-worked to show multiple hosts in a tree view, remebering connections across restarts. It is not currently possible to create new guests with a remote host connection. This capability will be added in a future release. The guest VNC console implementation has been replaced with the GTK-VNC widget for greatly improved performance and increased feature set. Other miscellaneous bug fixes and feature enhancements are also included. Thanks to Hugh Brock for doing the major refactoring of the code/UI to let us support multiple remote connections, and to everyone else involved in contributing to this release through submitting bug reports, translating the UI, writing patches & generally testing the code! Further information on this and other virtualization management tools is available from the project website http://virt-manager.org/ Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From Kevin.Maute at WPAFB.AF.MIL Thu Aug 30 12:46:13 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Thu, 30 Aug 2007 08:46:13 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D5D23E.2050408@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com> <46D5D23E.2050408@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> As requested: Script started on Thu 30 Aug 2007 08:35:37 AM EDT ]0;root at coalescence2:/var/www/cobbler/kickstarts_sys/00:E0:81:31:7C:D4 [root at coalescence2 00:E0:81:31:7C:D4]# cobbler repo add --name=epel5testing --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_ 64/ ]0;root at coalescence2:/var/www/cobbler/kickstarts_sys/00:E0:81:31:7C:D4 [root at coalescence2 00:E0:81:31:7C:D4]# cobbler reposync - rsync -av --delete --delete-excluded --exclude-from=/etc/cobbler/rsync.exclude /home/repos/RHEL5/os/x86_64/Cluster/ /var/www/cobbler/repo_mirror/rhel5clustering building file list ... done deleting config.repo ./ sent 1420 bytes received 26 bytes 2892.00 bytes/sec total size is 73819934 speedup is 51051.13 - walking: /var/www/cobbler/repo_mirror/rhel5clustering - creating: /var/www/cobbler/repo_mirror/rhel5clustering/config.repo - rsync -av --delete --delete-excluded --exclude-from=/etc/cobbler/rsync.exclude /home/repos/EPEL5/x86_64/ /var/www/cobbler/repo_mirror/epel5 building file list ... done deleting config.repo ./ sent 62467 bytes received 26 bytes 41662.00 bytes/sec total size is 965800920 speedup is 15454.55 - walking: /var/www/cobbler/repo_mirror/epel5 - creating: /var/www/cobbler/repo_mirror/epel5/config.repo - creating: /var/www/cobbler/repo_mirror/.origin/epel5testing.repo - /usr/bin/reposync --config=/var/www/cobbler/repo_mirror/.origin/epel5testing.repo --repoid=epel5testing --download_path=/var/www/cobbler/repo_mirror Traceback (most recent call last): File "/usr/bin/reposync", line 196, in ? main() File "/usr/bin/reposync", line 140, in main my.doRepoSetup() File "/usr/lib/python2.4/site-packages/yum/__init__.py", line 299, in doRepoSetup repo.setup(self.conf.cache) File "/usr/lib/python2.4/site-packages/yum/yumRepo.py", line 505, in setup raise Errors.RepoError, ('Cannot open/read repomd.xml file for repository: %s' % self) yum.Errors.RepoError: Cannot open/read repomd.xml file for repository: epel5testing cobbler reposync failed ]0;root at coalescence2:/var/www/cobbler/kickstarts_sys/00:E0:81:31:7C:D4 [root at coalescence2 00:E0:81:31:7C:D4]# exit Script done on Thu 30 Aug 2007 08:39:12 AM EDT > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Wednesday, August 29, 2007 4:08 PM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Michael DeHaan wrote: > > Michael DeHaan wrote: > >> Maute, Kevin P CTR AFRL/VAOO wrote: > >>> Success!!! One issue though, epel5*testing, it appeared > to download > >>> everything fine from the fedora repo but seemed to choke with a > >>> traceback when it was populating the local repo. Not knowing the > >>> internals or python, I can't give any more input. I > backed out that > >>> repo and am golden. > >>> > >>> Thanks, > >>> Kevin > >> > >> Good deal. > >> > >> If you can email me the traceback (cobbler output) and the > command(s) > >> that caused it that would be helpful. > >> > >> --Michael > > > > I just ran the following for EPEL testing and it worked for > me (using > > 0.6.1): > > > > > > Grr, thunderbird :) > > What I meant to say was ... > > I just ran the following for EPEL testing and it worked for me (using > 0.6.1): > > [root at mdehaan cobbler]# cobbler repo add --name=epel5i386 > --mirror=http://linux.nssl.noaa.gov/epel/testing/5Server/i386 > [root at mdehaan cobbler]# cobbler repo sync > > You then of course would do: > > cobbler profile edit --name=nameofprofile --repos="epel5i386 > otherrepo1 > otherrepo2" etc. > > You would not have to repeat the repos that Cobbler pulled in > from the > import but would want to mirror your package updates. > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From rjones at redhat.com Thu Aug 30 15:02:03 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 30 Aug 2007 16:02:03 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-top 0.3.2.5 released In-Reply-To: <46D5BED7.4000100@mcclatchyinteractive.com> References: <46D5A6E7.2070608@redhat.com> <46D5BED7.4000100@mcclatchyinteractive.com> Message-ID: <46D6DBEB.4060100@redhat.com> Tom Georgoulias wrote: > Richard W.M. Jones wrote: >> I'm pleased to announce a new release of virt-top, which is a utility >> like 'top' that works for virtual machines. Many keys and command >> line options are the same as for ordinary top. > >> Source and binaries for several architectures available from: >> http://et.redhat.com/~rjones/virt-top/download.html > > The RHEL5 RPM on the download page is for the RHEL5.1 beta, right? That RPM should (about 99% certain) work on RHEL 5 GA. Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From mdehaan at redhat.com Thu Aug 30 15:04:19 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 30 Aug 2007 11:04:19 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com> <46D5D23E.2050408@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <46D6DC73.9050003@redhat.com> Maute, Kevin P CTR AFRL/VAOO wrote: > yum.Errors.RepoError: Cannot open/read repomd.xml file for repository: > epel5testing Hmm, yum's unhappy. I just tried this using the 0.6.1 version (releasing soon) and it works for me, so I believe we're ok. Anyhow, I believe you said this was fixed with the earlier patch from the "Success!" email? There are a few things that can cause an error with yum. Connectivity problems with the repo or having the URL wrong (which appears fine) are the most likely things to go wrong. Occasionally if you kill the reposync process while it's running you'll leave it in an indeterminate state -- though I don't really recall seeing that recently. In the worst case, you may want to clean out /var/www/cobbler/repo_mirror/epel5testing and then re-sync. Though, again, I think you said it worked earlier with the last email? --Michael From Kevin.Maute at WPAFB.AF.MIL Thu Aug 30 15:10:22 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Thu, 30 Aug 2007 11:10:22 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D6DC73.9050003@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com><46D5D23E.2050408@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DC73.9050003@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil> I backed out epel5 testing for now, epel5 works fine. > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Thursday, August 30, 2007 11:04 AM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Maute, Kevin P CTR AFRL/VAOO wrote: > > yum.Errors.RepoError: Cannot open/read repomd.xml file for > repository: > > epel5testing > > Hmm, yum's unhappy. I just tried this using the 0.6.1 version > (releasing > soon) and it works for me, so I believe > we're ok. Anyhow, I believe you said this was fixed with the earlier > patch from the "Success!" email? > > There are a few things that can cause an error with yum. Connectivity > problems with the repo > or having the URL wrong (which appears fine) are the most > likely things > to go wrong. Occasionally if > you kill the reposync process while it's running you'll leave > it in an > indeterminate state -- though I don't > really recall seeing that recently. > > In the worst case, you may want to clean out > /var/www/cobbler/repo_mirror/epel5testing and > then re-sync. > > Though, again, I think you said it worked earlier with the last email? > > --Michael > > > > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From mdehaan at redhat.com Thu Aug 30 15:16:05 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 30 Aug 2007 11:16:05 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com><46D5D23E.2050408@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DC73.9050003@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <46D6DF35.3090707@redhat.com> Maute, Kevin P CTR AFRL/VAOO wrote: > I backed out epel5 testing for now, epel5 works fine. > Weird. Works for me, and if normal EPEL works, testing is the same. Any weird proxies/filtering in the way? Here's the commands I used ... # cobbler repo add --name=e5t64 --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_64/ # cobbler reposync As I've said, in the worst case you can clean out /var/www/cobbler/repo_mirror/$name-of-repo, delete the repo definition, and retry... though I'm not entirely sure that would help. Good luck! --Michael From Kevin.Maute at WPAFB.AF.MIL Thu Aug 30 17:02:45 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Thu, 30 Aug 2007 13:02:45 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D6DF35.3090707@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com><46D5D23E.2050408@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DC73.9050003@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DF35.3090707@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021EBBC8@VFOHMLAO13.Enterprise.afmc.ds.af.mil> I built the EPEL local repo before I started messing with cobbler so it was local. The first time I did the epel testing I used your suggestion as documented. The first time it appeared to pull them down as expected and then choked when it was populating the local repo. All the previous speaking as an end user who is still getting familiar with th RH way of package mgmt... > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Thursday, August 30, 2007 11:16 AM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Maute, Kevin P CTR AFRL/VAOO wrote: > > I backed out epel5 testing for now, epel5 works fine. > > > > Weird. Works for me, and if normal EPEL works, testing is the > same. Any > weird proxies/filtering in the way? > > Here's the commands I used ... > > # cobbler repo add --name=e5t64 > --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_64/ > # cobbler reposync > > As I've said, in the worst case you can clean out > /var/www/cobbler/repo_mirror/$name-of-repo, delete the repo > definition, > and retry... though I'm not entirely sure that would help. Good luck! > > --Michael > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From tomg at mcclatchyinteractive.com Thu Aug 30 17:29:31 2007 From: tomg at mcclatchyinteractive.com (Tom Georgoulias) Date: Thu, 30 Aug 2007 13:29:31 -0400 Subject: [et-mgmt-tools] ANNOUNCE: virt-top 0.3.2.5 released In-Reply-To: <46D6DBEB.4060100@redhat.com> References: <46D5A6E7.2070608@redhat.com> <46D5BED7.4000100@mcclatchyinteractive.com> <46D6DBEB.4060100@redhat.com> Message-ID: <46D6FE7B.3020409@mcclatchyinteractive.com> Richard W.M. Jones wrote: > Tom Georgoulias wrote: >> Richard W.M. Jones wrote: >>> I'm pleased to announce a new release of virt-top, which is a utility >>> like 'top' that works for virtual machines. Many keys and command >>> line options are the same as for ordinary top. >> >>> Source and binaries for several architectures available from: >>> http://et.redhat.com/~rjones/virt-top/download.html >> >> The RHEL5 RPM on the download page is for the RHEL5.1 beta, right? > > That RPM should (about 99% certain) work on RHEL 5 GA. I asked that because tried installing it on one of my RHEL5 servers and it didn't work at all. When I run virt-top, the terminal kinda "flashes" and returns a prompt. [root at rvrt001d ~]# cat /etc/redhat-release Red Hat Enterprise Linux Server release 5 (Tikanga) [root at rvrt001d ~]# rpm -q virt-top virt-top-0.3.2.5-1 [root at rvrt001d ~]# rpm -q libvirt libvirt-0.1.8-15.el5 libvirt for RHEL 5.0 is older than the 0.2.1 you mentioned on the download page, and the RHEL5 beta on RHN has 0.2.3-4. -- Tom Georgoulias Sr. Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com From mdehaan at redhat.com Thu Aug 30 19:05:00 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 30 Aug 2007 15:05:00 -0400 Subject: [et-mgmt-tools] ANNOUNCE: Cobbler 0.6.1 and Koan 0.6.1 released Message-ID: <46D714DC.2040607@redhat.com> Hi folks, It's Cobbler release time again. The main point of this stable release is to include bugfixes, primarily in some of the advanced cobbler features like profile inheritance, repository mirroring, and so on. Serious thanks go out to all those folks on #cobbler that helped with testing, ideas, patches, and bug reports. You folks are great! This release also includes early zeroconf support [1], improved init scripts that restart cleanly (yay!), the "search arbitrary fields" API patch, and modifications to keep the cobblerd daemon from printing all over your console :) I've included the full changelogs [2] below. Another interesting recent patch is the ability to relocate your cobbler install (say, if you keep it on a laptop) by just editing the server parameter in /var/lib/cobbler/settings and re-syncing. Before, you'd have to do a bit more editing -- particularly if you used "cobbler import". Koan is basically the same but now has the optional --virt-bridge parameter, which defaults to an "autodetect" mode. If you are using Xen, this will set up bridged networking out of the box, so that you can dial in and out of your virtual guests. If you're using KVM, you'll probably want to create a bridge first (you won't have "xenbr0"), and you can see instructions on how to do that here: https://hosted.fedoraproject.org/projects/cobbler/wiki/VirtNetworkingSetupForUseWithKoan . Koan also contains one fix that is needed to use KVM happily on x86_64. None of this effects the "--replace-self" usage, this is just virt related. Both cobbler and koan 0.6.1-2 have been pushed out to FC-6, F-7, rawhide (F8), EPEL-4 testing, and EPEL-5 testing -- they should be available shortly. The source RPMs for each are available at http://cobbler.et.redhat.com/download/. Questions/comments welcome! --Michael --- Footnotes: [1] to enable zeroconf, you'll need a Fedora OS at this point -- I'll be working on other options. Install avahi-tools on the server, and restart cobblerd -- your cobbler server will begin publishing. Any koan client that also has avahi-tools installed can then use "--server=DISCOVER" to find the cobbler server. This would be very useful for the koan live CD as you could build one CD, carry it around, and use it at multiple sites. That live CD would be built from a koan source checkout using "build.py --server=DISCOVER --koan=""". [2.1] Changelogs for Cobbler * Thu Aug 30 2007 - 0.6.1 - re enable --resolve in yumdownloader (cobbler repo mgmt feature) - fix get_distros_for_koan API function in cobblerd (not used by koan) - allow find API to search by arbitrary fields - status and logging now shows system names - upgraded init scripts - zeroconf/avahi publishing for cobblerd service - logRequests = 0 for XMLRPC. Make it be quiet. - ignore subdirectories of /var/lib/cobbler/snippets - fixed bug in graph rendering that allowed for upward property propogation in some cases - fixed bug that did not correctly evaluate repository settings of inherited sub-profiles/objects - tweaked domU sample kickstart to include wget - added some more unit tests - fix typo down one error path in cobbler sync. - fix reposync handling when using rsync protocol and directory paths do not contain arch - allow basic usage of Cheetah variables in config files @@server@@, etc. - fix auto-repo attachment for distros with split trees (i.e. RHEL5) [2.2] Changelogs for koan * Thu Aug 30 2007 Michael DeHaan - 0.6.1-1 - Bridged networking control options (--virt-bridge) - avahi support (--server=DISCOVER) - do not allow autodetection of mac with --virt - allow "xen" as alias for "xenpv" for --virt-type - correct arch name when passing to qemu hypervisor - don't complain when using default vcpu setting From levon at movementarian.org Thu Aug 30 20:17:54 2007 From: levon at movementarian.org (John Levon) Date: Thu, 30 Aug 2007 21:17:54 +0100 Subject: [et-mgmt-tools] Admin request: separate virt list Message-ID: <20070830201754.GA3654@totally.trollied.org.uk> Hi, would it be possible to get a separate list for things such as virt-manager, virt-top, virt-viewer etc.? They have an application much wider than just Fedora/RedHat systems and I'm seeing quite a lot of unrelated traffic. There's already a libvirt list, of course. regards, john From mdehaan at redhat.com Thu Aug 30 20:20:44 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 30 Aug 2007 16:20:44 -0400 Subject: [et-mgmt-tools] Admin request: separate virt list In-Reply-To: <20070830201754.GA3654@totally.trollied.org.uk> References: <20070830201754.GA3654@totally.trollied.org.uk> Message-ID: <46D7269C.1090605@redhat.com> John Levon wrote: > Hi, would it be possible to get a separate list for things such as > virt-manager, virt-top, virt-viewer etc.? They have an application much > wider than just Fedora/RedHat systems and I'm seeing quite a lot of > unrelated traffic. > > There's already a libvirt list, of course. > > regards, > john > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > What's "unrelated" traffic with respect to "Fedora/Linux Management Tools"? From levon at movementarian.org Thu Aug 30 22:32:59 2007 From: levon at movementarian.org (John Levon) Date: Thu, 30 Aug 2007 23:32:59 +0100 Subject: [et-mgmt-tools] Admin request: separate virt list In-Reply-To: <46D7269C.1090605@redhat.com> References: <20070830201754.GA3654@totally.trollied.org.uk> <46D7269C.1090605@redhat.com> Message-ID: <20070830223259.GD3654@totally.trollied.org.uk> On Thu, Aug 30, 2007 at 04:20:44PM -0400, Michael DeHaan wrote: > >Hi, would it be possible to get a separate list for things such as > >virt-manager, virt-top, virt-viewer etc.? They have an application much > >wider than just Fedora/RedHat systems and I'm seeing quite a lot of > >unrelated traffic. > > > >There's already a libvirt list, of course. > > What's "unrelated" traffic with respect to "Fedora/Linux Management Tools"? Unrelated to stuff layered on top of libvirt I mean. john From zxvdr.au at gmail.com Fri Aug 31 01:13:36 2007 From: zxvdr.au at gmail.com (David Robinson) Date: Fri, 31 Aug 2007 11:13:36 +1000 Subject: [et-mgmt-tools] more gui improvements In-Reply-To: <46D56EA9.4010407@redhat.com> References: <46CE010D.4080908@redhat.com> <46D4E5EE.9090909@gmail.com> <46D56EA9.4010407@redhat.com> Message-ID: <46D76B40.4080707@gmail.com> >> I attached a patch to BZ 253144 that adds the "are you sure" dialog to >> the delete button. The patch applies against virt-manager in RHEL5.1 >> beta. >> > Heh... well, I'm not sure we really need it in the 5.1 version, because > the delete button there is fairly obvious and located separately at the > bottom of the window. I would consider it for 5.2 however (patch > deadline for 5.1 is way past now). No problem. I was thinking more along the lines of "yes, I did mean to press Delete, but opps I selected the wrong guest", and "opps, I meant to press New, not Delete" :-) --Dave From mdehaan at redhat.com Fri Aug 31 14:47:13 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 31 Aug 2007 10:47:13 -0400 Subject: [et-mgmt-tools] Admin request: separate virt list In-Reply-To: <20070830223259.GD3654@totally.trollied.org.uk> References: <20070830201754.GA3654@totally.trollied.org.uk> <46D7269C.1090605@redhat.com> <20070830223259.GD3654@totally.trollied.org.uk> Message-ID: <46D829F1.2010503@redhat.com> John Levon wrote: > On Thu, Aug 30, 2007 at 04:20:44PM -0400, Michael DeHaan wrote: > > >>> Hi, would it be possible to get a separate list for things such as >>> virt-manager, virt-top, virt-viewer etc.? They have an application much >>> wider than just Fedora/RedHat systems and I'm seeing quite a lot of >>> unrelated traffic. >>> >>> There's already a libvirt list, of course. >>> >> What's "unrelated" traffic with respect to "Fedora/Linux Management Tools"? >> > > Unrelated to stuff layered on top of libvirt I mean. > > john > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools This is still infinitely lower traffic than things like fedora-devel. You'll survive :) While this isn't a libvirt exclusive list, though it should be pointed out that koan /does/ use libvirt. It's a list for community projects coming out of Red Hat -- i.e. http://et.redhat.com/ --Michael From berrange at redhat.com Fri Aug 31 15:07:51 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 31 Aug 2007 16:07:51 +0100 Subject: [et-mgmt-tools] Admin request: separate virt list In-Reply-To: <20070830223259.GD3654@totally.trollied.org.uk> References: <20070830201754.GA3654@totally.trollied.org.uk> <46D7269C.1090605@redhat.com> <20070830223259.GD3654@totally.trollied.org.uk> Message-ID: <20070831150751.GC10940@redhat.com> On Thu, Aug 30, 2007 at 11:32:59PM +0100, John Levon wrote: > On Thu, Aug 30, 2007 at 04:20:44PM -0400, Michael DeHaan wrote: > > > >Hi, would it be possible to get a separate list for things such as > > >virt-manager, virt-top, virt-viewer etc.? They have an application much > > >wider than just Fedora/RedHat systems and I'm seeing quite a lot of > > >unrelated traffic. > > > > > >There's already a libvirt list, of course. > > > > What's "unrelated" traffic with respect to "Fedora/Linux Management Tools"? > > Unrelated to stuff layered on top of libvirt I mean. There has been rather alot of end user debugging / stuff relating to Cobbler Koan recently, not directly relating to development discussion, but not sure whether there's enough traffic to justify separate lists just yet. Perhaps lets just see how traffic develops - may become desirable to have separate lists as more people get involved. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From mdehaan at redhat.com Fri Aug 31 15:12:17 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 31 Aug 2007 11:12:17 -0400 Subject: [et-mgmt-tools] Admin request: separate virt list In-Reply-To: <20070831150751.GC10940@redhat.com> References: <20070830201754.GA3654@totally.trollied.org.uk> <46D7269C.1090605@redhat.com> <20070830223259.GD3654@totally.trollied.org.uk> <20070831150751.GC10940@redhat.com> Message-ID: <46D82FD1.9040701@redhat.com> Daniel P. Berrange wrote: > On Thu, Aug 30, 2007 at 11:32:59PM +0100, John Levon wrote: > >> On Thu, Aug 30, 2007 at 04:20:44PM -0400, Michael DeHaan wrote: >> >> >>>> Hi, would it be possible to get a separate list for things such as >>>> virt-manager, virt-top, virt-viewer etc.? They have an application much >>>> wider than just Fedora/RedHat systems and I'm seeing quite a lot of >>>> unrelated traffic. >>>> >>>> There's already a libvirt list, of course. >>>> >>> What's "unrelated" traffic with respect to "Fedora/Linux Management Tools"? >>> >> Unrelated to stuff layered on top of libvirt I mean. >> > > There has been rather alot of end user debugging / stuff relating to Cobbler > Koan recently, not directly relating to development discussion, but not > sure whether there's enough traffic to justify separate lists just yet. > Perhaps lets just see how traffic develops - may become desirable to have > separate lists as more people get involved. > > Dan. > FWIW, this was originally formed as an end-user list. --Michael From rjones at redhat.com Fri Aug 31 15:21:06 2007 From: rjones at redhat.com (Richard W.M. Jones) Date: Fri, 31 Aug 2007 16:21:06 +0100 Subject: [et-mgmt-tools] ANNOUNCE: virt-top 0.3.2.5 released In-Reply-To: <46D6FE7B.3020409@mcclatchyinteractive.com> References: <46D5A6E7.2070608@redhat.com> <46D5BED7.4000100@mcclatchyinteractive.com> <46D6DBEB.4060100@redhat.com> <46D6FE7B.3020409@mcclatchyinteractive.com> Message-ID: <46D831E2.8030801@redhat.com> Tom Georgoulias wrote: > Richard W.M. Jones wrote: >> Tom Georgoulias wrote: >>> Richard W.M. Jones wrote: >>>> I'm pleased to announce a new release of virt-top, which is a >>>> utility like 'top' that works for virtual machines. Many keys and >>>> command line options are the same as for ordinary top. >>> >>>> Source and binaries for several architectures available from: >>>> http://et.redhat.com/~rjones/virt-top/download.html >>> >>> The RHEL5 RPM on the download page is for the RHEL5.1 beta, right? >> >> That RPM should (about 99% certain) work on RHEL 5 GA. > > I asked that because tried installing it on one of my RHEL5 servers and > it didn't work at all. When I run virt-top, the terminal kinda > "flashes" and returns a prompt. > > [root at rvrt001d ~]# cat /etc/redhat-release > Red Hat Enterprise Linux Server release 5 (Tikanga) > [root at rvrt001d ~]# rpm -q virt-top > virt-top-0.3.2.5-1 > [root at rvrt001d ~]# rpm -q libvirt > libvirt-0.1.8-15.el5 > > libvirt for RHEL 5.0 is older than the 0.2.1 you mentioned on the > download page, and the RHEL5 beta on RHN has 0.2.3-4. Ah right, I see. libvirt 0.1.8 is too old for virt-top. You need minimum libvirt 0.2.1. The reason is because we need the vcpu info stuff which was only added in libvirt 0.2.1 (http://libvirt.org/hvsupport.html). One simple solution is to compile a local copy of libvirt and point the binary at it, something like this: LD_LIBRARY_PATH=/my/local/copy/lib virt-top By the way, you can narrow down these problems by doing: virt-top --debug somefile which will write debugging info to the file (basically, everything which would normally go to stderr, but in a curses program stderr isn't really available). Rich. -- Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/ Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 03798903 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 3237 bytes Desc: S/MIME Cryptographic Signature URL: From Kevin.Maute at WPAFB.AF.MIL Fri Aug 31 16:21:46 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Fri, 31 Aug 2007 12:21:46 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D6DF35.3090707@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com><46D5D23E.2050408@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DC73.9050003@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DF35.3090707@redhat.com> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Michael, Good news/bad news. 0.6.1 seems to have resolved the below issue. Now, however, dhcpd does not restart (Warning: dhcpd restart failed) on a cobbler sync. My settings file looks good and (I swear!) nothing changed other than me installing 0.6.1. doing a ps on dhcpd shows it running since yesterday. Has anyone else noticed this? Kevin > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > Sent: Thursday, August 30, 2007 11:16 AM > To: Fedora/Linux Management Tools > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > Maute, Kevin P CTR AFRL/VAOO wrote: > > I backed out epel5 testing for now, epel5 works fine. > > > > Weird. Works for me, and if normal EPEL works, testing is the > same. Any > weird proxies/filtering in the way? > > Here's the commands I used ... > > # cobbler repo add --name=e5t64 > --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_64/ > # cobbler reposync > > As I've said, in the worst case you can clean out > /var/www/cobbler/repo_mirror/$name-of-repo, delete the repo > definition, > and retry... though I'm not entirely sure that would help. Good luck! > > --Michael > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From Kevin.Maute at WPAFB.AF.MIL Fri Aug 31 16:40:43 2007 From: Kevin.Maute at WPAFB.AF.MIL (Maute, Kevin P CTR AFRL/VAOO) Date: Fri, 31 Aug 2007 12:40:43 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAC34@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5D061.5090505@redhat.com> <46D5D1A8.9080900@redhat.com><46D5D23E.2050408@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BAE08@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D6DC73.9050003@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D6DF35.3090707@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <29A7C5F65D54744CBFD97CD62A33DEE9021EC195@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Never mind. Friday gremlins. I cleaned up my system names and it worked :-) > -----Original Message----- > From: et-mgmt-tools-bounces at redhat.com > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Maute, > Kevin P CTR AFRL/VAOO > Sent: Friday, August 31, 2007 12:22 PM > To: Fedora/Linux Management Tools > Subject: RE: [et-mgmt-tools] Cobbler install/kickstart question > > Michael, > > Good news/bad news. 0.6.1 seems to have resolved the below > issue. Now, > however, dhcpd does not restart (Warning: dhcpd restart failed) on a > cobbler sync. My settings file looks good and (I swear!) nothing > changed other than me installing 0.6.1. doing a ps on dhcpd shows it > running since yesterday. Has anyone else noticed this? > > Kevin > > > -----Original Message----- > > From: et-mgmt-tools-bounces at redhat.com > > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of > Michael DeHaan > > Sent: Thursday, August 30, 2007 11:16 AM > > To: Fedora/Linux Management Tools > > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > > > Maute, Kevin P CTR AFRL/VAOO wrote: > > > I backed out epel5 testing for now, epel5 works fine. > > > > > > > Weird. Works for me, and if normal EPEL works, testing is the > > same. Any > > weird proxies/filtering in the way? > > > > Here's the commands I used ... > > > > # cobbler repo add --name=e5t64 > > > --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_64/ > > # cobbler reposync > > > > As I've said, in the worst case you can clean out > > /var/www/cobbler/repo_mirror/$name-of-repo, delete the repo > > definition, > > and retry... though I'm not entirely sure that would help. > Good luck! > > > > --Michael > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > From bk at suse.de Fri Aug 31 16:44:18 2007 From: bk at suse.de (Bernhard Kaindl) Date: Fri, 31 Aug 2007 18:44:18 +0200 (CEST) Subject: [et-mgmt-tools] [PATCH] use localhost when connecting local VNC guest console Message-ID: Hi, I tried virt-manager-0.5.0 and one of the new features of 0.5.0, the support for managing guest on other hosts has resulted in a regession where in 0.4.0, the destination port for the VNC connect was 127.0.0.1:5900 for the local host, while it is now :5900: --- virt-manager-0.4.0/src/virtManager/domain.py 2007-04-16 +++ virt-manager-0.5.0/src/virtManager/domain.py 2007-08-29 ... @@ -430,16 +446,18 @@ return os.access(tty, os.R_OK | os.W_OK) def get_graphics_console(self): + self.xml = None type = self.get_xml_string("/domain/devices/graphics/@type") port = None if type == "vnc": port = self.get_xml_string("/domain/devices/graphics[@type='vnc']/@port") - if port == None: - port = 5900 + self.get_id() - else: + if port is not None: port = int(port) - return [type, "127.0.0.1", port] - return [type, None, None] + + transport, username = self.connection.get_transport() + + return [type, self.connection.get_hostname(), port, transport] + def get_disk_devices(self): xml = self.get_xml() ... What resulted in the regession for me was that I was using libvirtd to manage QEMU guests and this setup didn't bind it's VNC port to :5900, but to 127.0.0.1:5900, where it was expected to be with virt-manager-0.4.0. I'd usually not recommend to make the VNC console my guests list : unless it's neccessary and the machine is inside an secure environment because is normally defined to resolve to the LAN address of the machine and that would make the VNC console accessible from the LAN in case no firewall is active. How it securely connect to the VNC console of other hypervisor hosts is over an insecure network is another issue, but it seems easy to make the console widget of virt-manager-0.5.0 working with the current VNC console setup when checking by using 127.0.0.1 for the hostname: virtManager's console.py uses get_graphics_console(self) to get the directions for connecting to the graphics_console of the guest, so as far as I tested yet, that is fixed by checking if self.connection.get_hostname() equals self.connection.get_local_hostname(): and if that is so (both are the hostname of the local machine) then the VNC connection can be directed to 127.0.0.1 as before: --- virt-manager--devel/src/virtManager/domain.py +++ virt-manager--devel/src/virtManager/domain.py @@ -456,7 +456,12 @@ def get_graphics_console(self): transport, username = self.connection.get_transport() - return [type, self.connection.get_hostname(), port, transport] + hostname = self.connection.get_hostname() + + if hostname == self.connection.get_local_hostname(): + hostname = "127.0.0.1" + + return [type, hostname, port, transport] def get_disk_devices(self): Comments? -- Bernhard Kaindl From wright at imageworks.com Fri Aug 31 16:29:00 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 31 Aug 2007 09:29:00 -0700 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat6D6DC73.9050003@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D6DF35.3090707@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil> Message-ID: <46D841CC.5050306@imageworks.com> I've run into this issue before. Check your /etc/cobbler/dhcp.template. I've had my local version get moved to dhcp.template.rpmsave, and cobbler will install the default one. If you define a custom subnet for example, dhcp will refuse to start (cobbler defaults to 192.168.1/24). /var/log/messages should show you why dhcpd is failing to start as well. HTH -pete Maute, Kevin P CTR AFRL/VAOO wrote: > > Michael, > > Good news/bad news. 0.6.1 seems to have resolved the below issue. Now, > however, dhcpd does not restart (Warning: dhcpd restart failed) on a > cobbler sync. My settings file looks good and (I swear!) nothing > changed other than me installing 0.6.1. doing a ps on dhcpd shows it > running since yesterday. Has anyone else noticed this? > > Kevin > > > -----Original Message----- > > From: et-mgmt-tools-bounces at redhat.com > > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan > > Sent: Thursday, August 30, 2007 11:16 AM > > To: Fedora/Linux Management Tools > > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question > > > > Maute, Kevin P CTR AFRL/VAOO wrote: > > > I backed out epel5 testing for now, epel5 works fine. > > > > > > > Weird. Works for me, and if normal EPEL works, testing is the > > same. Any > > weird proxies/filtering in the way? > > > > Here's the commands I used ... > > > > # cobbler repo add --name=e5t64 > > --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_64/ > > # cobbler reposync > > > > As I've said, in the worst case you can clean out > > /var/www/cobbler/repo_mirror/$name-of-repo, delete the repo > > definition, > > and retry... though I'm not entirely sure that would help. Good luck! > > > > --Michael > > > > _______________________________________________ > > et-mgmt-tools mailing list > > et-mgmt-tools at redhat.com > > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > > > > _______________________________________________ > et-mgmt-tools mailing list > et-mgmt-tools at redhat.com > https://www.redhat.com/mailman/listinfo/et-mgmt-tools > -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From bk at suse.de Fri Aug 31 17:05:43 2007 From: bk at suse.de (Bernhard Kaindl) Date: Fri, 31 Aug 2007 19:05:43 +0200 (CEST) Subject: [et-mgmt-tools] Re: [PATCH] use localhost when connecting local VNC guest console In-Reply-To: References: Message-ID: On Fri, 31 Aug 2007, Bernhard Kaindl wrote: > and if that is so (both are the hostname of the local machine) then > the VNC connection can be directed to 127.0.0.1 as before: > > --- virt-manager--devel/src/virtManager/domain.py > +++ virt-manager--devel/src/virtManager/domain.py > @@ -456,7 +456,12 @@ def get_graphics_console(self): > > transport, username = self.connection.get_transport() > > - return [type, self.connection.get_hostname(), port, transport] > + hostname = self.connection.get_hostname() > + > + if hostname == self.connection.get_local_hostname(): > + hostname = "127.0.0.1" > + > + return [type, hostname, port, transport] > > > def get_disk_devices(self): Note: The message which was displayed when the gtk-vnc widget got the ECONNREFUSED (Connection refused) error when connecting to :5900 sounds a bit misleading to me. socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 4 connect(4, {sa_family=AF_INET, sin_port=htons(5900), sin_addr=inet_addr("")}, 16) = -1 ECONNREFUSED (Connection refused) The message resulting from this was: "Console was disconnected from guest" It's shown by: virt-manager-0.5.0/src/virtManager/console.py:234 def _vnc_disconnected(self, src): logging.debug("VNC disconnected") self.vncViewerFailures = self.vncViewerFailures + 1 self.activate_unavailable_page(_("Console was disconnected from guest")) which is connected with: virt-manager-0.5.0/src/virtManager/console.py:58 self.vncViewer.connect("vnc-disconnected", self._vnc_disconnected) which itself is connected with: /usr/share/doc/gtk-vnc-python-0.1.0/gvncviewer.py:173 vnc.connect("vnc-disconnected", vnc_disconnected) originating from gtk-vnc-0.1.0/src/vncdisplay.c: signals[VNC_DISCONNECTED] = g_signal_new ("vnc-disconnected", G_OBJECT_CLASS_TYPE (object_class), G_SIGNAL_RUN_FIRST, G_STRUCT_OFFSET (VncDisplayClass, vnc_disconnected), NULL, NULL, g_cclosure_marshal_VOID__VOID, G_TYPE_NONE, 0); Unfortunately, gtk-vnc does not seem ot have a signal for VNC_CONNECTION_REFUSED yet, it seems. I think that is something to push for to get accurate signal info in this case to be able to possibly handle that in a more specific way. Bernhard From mdehaan at redhat.com Fri Aug 31 17:41:28 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 31 Aug 2007 13:41:28 -0400 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D841CC.5050306@imageworks.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat6D6DC73.9050003@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D6DF35.3090707@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D841CC.5050306@imageworks.com> Message-ID: <46D852C8.9000405@redhat.com> Peter Wright wrote: > I've run into this issue before. Check your > /etc/cobbler/dhcp.template. I've had my local version get moved to > dhcp.template.rpmsave, and cobbler will install the default one. If > you define a custom subnet for example, dhcp will refuse to start > (cobbler defaults to 192.168.1/24). > > /var/log/messages should show you why dhcpd is failing to start as well. > > HTH > -pete According to git, the template file has been marked as "config(noreplace)" in the RPM since 10/25/2006. So upgrades should make a ".rpmnew" but they should not install the existing one and move your old one to where it is not used. Only files just marked "config" should have the ".rpmsave" behavior. I'm not sure why you're seeing that exactly? Of course /var/log/messages is a great place to look regardless. --Michael > > Maute, Kevin P CTR AFRL/VAOO wrote: >> >> Michael, >> >> Good news/bad news. 0.6.1 seems to have resolved the below issue. Now, >> however, dhcpd does not restart (Warning: dhcpd restart failed) on a >> cobbler sync. My settings file looks good and (I swear!) nothing >> changed other than me installing 0.6.1. doing a ps on dhcpd shows it >> running since yesterday. Has anyone else noticed this? >> >> Kevin >> >> > -----Original Message----- >> > From: et-mgmt-tools-bounces at redhat.com >> > [mailto:et-mgmt-tools-bounces at redhat.com] On Behalf Of Michael DeHaan >> > Sent: Thursday, August 30, 2007 11:16 AM >> > To: Fedora/Linux Management Tools >> > Subject: Re: [et-mgmt-tools] Cobbler install/kickstart question >> > >> > Maute, Kevin P CTR AFRL/VAOO wrote: >> > > I backed out epel5 testing for now, epel5 works fine. >> > > > >> > Weird. Works for me, and if normal EPEL works, testing is the >> > same. Any >> > weird proxies/filtering in the way? >> > >> > Here's the commands I used ... >> > >> > # cobbler repo add --name=e5t64 >> > --mirror=http://download.fedora.redhat.com/pub/epel/testing/5/x86_64/ >> > # cobbler reposync >> > >> > As I've said, in the worst case you can clean out >> > /var/www/cobbler/repo_mirror/$name-of-repo, delete the repo >> > definition, >> > and retry... though I'm not entirely sure that would help. Good luck! >> > >> > --Michael >> > >> > _______________________________________________ >> > et-mgmt-tools mailing list >> > et-mgmt-tools at redhat.com >> > https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> > >> >> _______________________________________________ >> et-mgmt-tools mailing list >> et-mgmt-tools at redhat.com >> https://www.redhat.com/mailman/listinfo/et-mgmt-tools >> > > From berrange at redhat.com Fri Aug 31 17:54:20 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 31 Aug 2007 18:54:20 +0100 Subject: [et-mgmt-tools] Re: [PATCH] use localhost when connecting local VNC guest console In-Reply-To: References: Message-ID: <20070831175420.GA7881@redhat.com> On Fri, Aug 31, 2007 at 07:05:43PM +0200, Bernhard Kaindl wrote: > On Fri, 31 Aug 2007, Bernhard Kaindl wrote: > > >and if that is so (both are the hostname of the local machine) then > >the VNC connection can be directed to 127.0.0.1 as before: > > > >--- virt-manager--devel/src/virtManager/domain.py > >+++ virt-manager--devel/src/virtManager/domain.py > >@@ -456,7 +456,12 @@ def get_graphics_console(self): > > > > transport, username = self.connection.get_transport() > > > >- return [type, self.connection.get_hostname(), port, transport] > >+ hostname = self.connection.get_hostname() > >+ > >+ if hostname == self.connection.get_local_hostname(): > >+ hostname = "127.0.0.1" > >+ > >+ return [type, hostname, port, transport] > > > > > > def get_disk_devices(self): > > Note: The message which was displayed when the gtk-vnc widget got the > ECONNREFUSED (Connection refused) error when connecting to :5900 > sounds a bit misleading to me. > > socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 4 > connect(4, {sa_family=AF_INET, sin_port=htons(5900), > sin_addr=inet_addr("")}, 16) = -1 > ECONNREFUSED (Connection refused) > > The message resulting from this was: > > "Console was disconnected from guest" Yes, I still need to get proper error reports being propagated out from GTK-VNC back to the app. It'll addressed in next GTK-VNC i hope. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From bk at suse.de Fri Aug 31 17:57:17 2007 From: bk at suse.de (Bernhard Kaindl) Date: Fri, 31 Aug 2007 19:57:17 +0200 (CEST) Subject: [et-mgmt-tools] [PATCH] workarounds and fixes for new gtk-vnc integration in console.py In-Reply-To: References: Message-ID: On Fri, 31 Aug 2007, Bernhard Kaindl wrote: > > Note: The message which was displayed when the gtk-vnc widget got the > ECONNREFUSED (Connection refused) error when connecting to :5900 > sounds a bit misleading to me. > > socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 4 > connect(4, {sa_family=AF_INET, sin_port=htons(5900), sin_addr=inet_addr(" host's LAN IP address here>")}, 16) = -1 ECONNREFUSED (Connection refused) > > The message resulting from this was: > > "Console was disconnected from guest" Another message like this still apppeared when starting a new QEMU / KVM guest from virt-manager, and the changes which I did in that regard to fix this are found in the cumulated patch below. I added some comments, so I also attached it as a MIME attachment: --- virt-manager--devel/src/virtManager/console.py 2007/08/30 23:11:35 1.1 +++ virt-manager--devel/src/virtManager/console.py 2007/08/31 00:53:58 @@ -74,8 +74,9 @@ self.window.get_widget("console-pages").append_page(self.vncViewer) self.vncViewer.realize() self.vncViewer.show() - self.vncViewerFailures = 0 + self.vncViewerRetriesScheduled = 0 # Simple rename to be more exact to what I'm using it now. self.vncViewerRetryDelay = 125 + self.connected = 0 # New: to keep track if connected or not (stop retries when connected) self.notifyID = None try: @@ -231,44 +232,61 @@ return 1 return 0 + def view_vm_status(self): + status = self.vm.status() + if status == libvirt.VIR_DOMAIN_SHUTOFF: + self.activate_unavailable_page(_("Guest not running")) + else: + if status == libvirt.VIR_DOMAIN_CRASHED: + self.activate_unavailable_page(_("Guest has crashed")) # Helper for displaying more accurate status + def _vnc_disconnected(self, src): + self.connected = 0 logging.debug("VNC disconnected") - self.vncViewerFailures = self.vncViewerFailures + 1 - self.activate_unavailable_page(_("Console was disconnected from guest")) # separately moved away to other places, see futher and it becomes clear if not self.is_visible(): return - if self.vncViewerFailures < 10: - self.schedule_retry() - else: - logging.error("Too many connection failures, not retrying again") + if self.vm.status() in [ libvirt.VIR_DOMAIN_SHUTOFF, libvirt.VIR_DOMAIN_CRASHED ]: + self.view_vm_status() + return # Well, there is no point in scheduling login retries to the gues when it is down # make use of self.view_vm_status() to view more explicit status information + + self.activate_unavailable_page(_("TCP/IP error: VNC connection to hypervisor host got refused or disconnected!")) # Updated version of "Console was disconnected from guest" until gtk-vnc supports the signal vnc-connection-refused + self.schedule_retry() # retry couting moved to schedule_retry, so no counting done here (and elsewhere) def _vnc_initialized(self, src): + self.connected = 1 logging.debug("VNC initialized") self.activate_viewer_page() # Had a succesfull connect, so reset counters now - self.vncViewerFailures = 0 + self.vncViewerRetriesScheduled = 0 # Simple rename to be more exact to what I'm using it now. self.vncViewerRetryDelay = 125 def schedule_retry(self): + self.vncViewerRetriesScheduled = self.vncViewerRetriesScheduled + 1 + if self.vncViewerRetriesScheduled >= 10: + logging.error("Too many connection failures, not retrying again") + return # retry couting moved to schedule_retry, retriy counting for all users of this function logging.warn("Retrying connection in %d ms", self.vncViewerRetryDelay) gobject.timeout_add(self.vncViewerRetryDelay, self.retry_login) if self.vncViewerRetryDelay < 2000: self.vncViewerRetryDelay = self.vncViewerRetryDelay * 2 def retry_login(self): + if self.connected: + return # retrying login if we are already logged in results in strange errors gtk.gdk.threads_enter() try: logging.debug("Got timed retry") self.try_login() - return False + return # return code not used AFAICS, can as well just return without status finally: gtk.gdk.threads_leave() def try_login(self, src=None): if self.vm.get_id() < 0: - self.activate_unavailable_page(_("Console not available for inactive guest")) # self.vm.get_id() < 0 means that the guest is not running, but "Console not available for inactive guest" # reads like a error meesage so make it more simple: + self.activate_unavailable_page(_("Guest not running")) # and schedule a retry in case the guest was just started but takes ~300ms to start in case of QEMU/KVM: + self.schedule_retry() return logging.debug("Trying console login") @@ -320,7 +338,7 @@ self.vncViewer.set_credential(credList[i], "libvirt") else: # Force it to stop re-trying - self.vncViewerFailures = 10 + self.vncViewerRetriesScheduled = 10 # Simple rename to be more exact to what I'm using it now. self.vncViewer.close() self.activate_unavailable_page(_("Unsupported console authentication type")) @@ -503,7 +521,7 @@ if self.window.get_widget("console-pages").get_current_page() != PAGE_UNAVAILABLE: self.vncViewer.close() self.window.get_widget("console-pages").set_current_page(PAGE_UNAVAILABLE) - self.activate_unavailable_page(_("Console not available for inactive guest")) + self.view_vm_status() # display more exact message else: if status == libvirt.VIR_DOMAIN_PAUSED: if self.window.get_widget("console-pages").get_current_page() == PAGE_VNCVIEWER: @@ -543,7 +561,7 @@ if self.vncViewer.is_open(): self.activate_viewer_page() else: - self.vncViewerFailures = 0 + self.vncViewerRetriesScheduled = 0 # Simple rename to be more exact to what I'm using it now. self.vncViewerRetryDelay = 125 self.try_login() self.ignorePause = False Without this patch, a the console of a newly started KVM / QEMU guest is only shown after toggling pause if the console window was open while starting it, and also some intermittent messages which could cause unneccesary confusion are displayed. I can split it up, but it should be rather clear what it does. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: virt-manager-0.5.0-gtkvnc-fixes.diff Type: text/x-patch Size: 4615 bytes Desc: URL: From berrange at redhat.com Fri Aug 31 18:00:10 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 31 Aug 2007 19:00:10 +0100 Subject: [et-mgmt-tools] [PATCH] use localhost when connecting local VNC guest console In-Reply-To: References: Message-ID: <20070831180010.GB7881@redhat.com> On Fri, Aug 31, 2007 at 06:44:18PM +0200, Bernhard Kaindl wrote: > Hi, > I tried virt-manager-0.5.0 and one of the new features of 0.5.0, > the support for managing guest on other hosts has resulted in a regession > where in 0.4.0, the destination port for the VNC connect was 127.0.0.1:5900 > for the local host, while it is now :5900: > > What resulted in the regession for me was that I was using libvirtd > to manage QEMU guests and this setup didn't bind it's VNC port to > :5900, but to 127.0.0.1:5900, where it was expected to be > with virt-manager-0.4.0. Ok, I think I understand the situation you're seeing, but just to be sure -you are running virt-manager directly on the Dom0 box being managed & its using a 'local' libvirt connect URI (ie one without any hostname in the URI). In Fedora /etc/hosts always has an entry for the qualfied hostname listed against 127.0.0.1 which'd be why i didn't notice this regression. Obviously this isn't something we can rely on, so need to explicitly connect to the localhost > How it securely connect to the VNC console of other hypervisor hosts > is over an insecure network is another issue, but it seems easy to > make the console widget of virt-manager-0.5.0 working with the current > VNC console setup when checking by using 127.0.0.1 for the hostname: QEMU now supports TLS + x509 certs in its VNC server. In Fedora I have also patched the Xen & KVM forks of QEMU to have the same capability, so it is secure for virt-manager to talk directly to the VNC server across an untrusted LAN. I will be submitting the KVM & Xen backports of this code upstream in the near future. I also intend to make virt-manager able to tunnel VNC over SSH for scenarios where the VNC server isn't listening publically. > > virtManager's console.py uses get_graphics_console(self) to get the > directions for connecting to the graphics_console of the guest, so > as far as I tested yet, that is fixed by checking if > > self.connection.get_hostname() > > equals > > self.connection.get_local_hostname(): > > and if that is so (both are the hostname of the local machine) then > the VNC connection can be directed to 127.0.0.1 as before: Hmm, I think I'll change the get_hostname() method so you can pass a flag indicating whether it should just return '127.0.0.1' for local connections, which should have same effect as the patch you provide. Regards, -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From bk at suse.de Fri Aug 31 18:04:38 2007 From: bk at suse.de (Bernhard Kaindl) Date: Fri, 31 Aug 2007 20:04:38 +0200 (CEST) Subject: [et-mgmt-tools] [PATCH] use localhost when connecting local VNC guest console In-Reply-To: <20070831180010.GB7881@redhat.com> References: <20070831180010.GB7881@redhat.com> Message-ID: On Fri, 31 Aug 2007, Daniel P. Berrange wrote: >> self.connection.get_hostname() >> >> equals >> >> self.connection.get_local_hostname(): >> >> and if that is so (both are the hostname of the local machine) then >> the VNC connection can be directed to 127.0.0.1 as before: > > Hmm, I think I'll change the get_hostname() method so you can pass a > flag indicating whether it should just return '127.0.0.1' for local > connections, which should have same effect as the patch you provide. Thanks, yes of course. Bernhard From wright at imageworks.com Fri Aug 31 18:25:31 2007 From: wright at imageworks.com (Peter Wright) Date: Fri, 31 Aug 2007 11:25:31 -0700 Subject: [et-mgmt-tools] Cobbler install/kickstart question In-Reply-To: <46D852C8.9000405@redhat.com> References: <29A7C5F65D54744CBFD97CD62A33DEE9021BA597@VFOHMLAO13.Enterprise.afmc.ds.af.mil> <46D5878E.4050002@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021BAA07@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D5B8E3.1090800@redhat.com> <46D5C184.3060508@redhat6D6DC73.9050003@redhat.com><29A7C5F65D54744CBFD97CD62A33DEE9021BB0AF@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D6DF35.3090707@redhat.com> <29A7C5F65D54744CBFD97CD62A33DEE9021EC180@VFOHMLAO13.Enterprise.afmc.ds.af.mil><46D841CC.5050306@imageworks.com> <46D852C8.9000405@redhat.com> Message-ID: <46D85D1B.3010009@imageworks.com> Michael DeHaan wrote: > > Peter Wright wrote: > > I've run into this issue before. Check your > > /etc/cobbler/dhcp.template. I've had my local version get moved to > > dhcp.template.rpmsave, and cobbler will install the default one. If > > you define a custom subnet for example, dhcp will refuse to start > > (cobbler defaults to 192.168.1/24). > > > > /var/log/messages should show you why dhcpd is failing to start as well. > > > > HTH > > -pete > > According to git, the template file has been marked as > "config(noreplace)" in the RPM since 10/25/2006. So upgrades should > make a ".rpmnew" but they should not install the existing one and move > your old one to where it is not used. Only files just marked "config" > should have the ".rpmsave" behavior. I'm not sure why you're seeing > that exactly? > > Of course /var/log/messages is a great place to look regardless. > > --Michael > > > d'oh! i must have seen this behavior because did remove then install. sorry for the noise! -pete -- Peter Wright Systems Administrator Sony Pictures Imageworks wright at imageworks.com www.imageworks.com From berrange at redhat.com Fri Aug 31 21:32:09 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 31 Aug 2007 22:32:09 +0100 Subject: [et-mgmt-tools] [PATCH] use localhost when connecting local VNC guest console In-Reply-To: <20070831180010.GB7881@redhat.com> References: <20070831180010.GB7881@redhat.com> Message-ID: <20070831213209.GF7881@redhat.com> On Fri, Aug 31, 2007 at 07:00:10PM +0100, Daniel P. Berrange wrote: > On Fri, Aug 31, 2007 at 06:44:18PM +0200, Bernhard Kaindl wrote: > > Hi, > > I tried virt-manager-0.5.0 and one of the new features of 0.5.0, > > the support for managing guest on other hosts has resulted in a regession > > where in 0.4.0, the destination port for the VNC connect was 127.0.0.1:5900 > > for the local host, while it is now :5900: > > > > What resulted in the regession for me was that I was using libvirtd > > to manage QEMU guests and this setup didn't bind it's VNC port to > > :5900, but to 127.0.0.1:5900, where it was expected to be > > with virt-manager-0.4.0. > > Ok, I think I understand the situation you're seeing, but just to be sure -you > are running virt-manager directly on the Dom0 box being managed & its using a > 'local' libvirt connect URI (ie one without any hostname in the URI). > > In Fedora /etc/hosts always has an entry for the qualfied hostname listed > against 127.0.0.1 which'd be why i didn't notice this regression. Obviously > this isn't something we can rely on, so need to explicitly connect to the > localhost > > > How it securely connect to the VNC console of other hypervisor hosts > > is over an insecure network is another issue, but it seems easy to > > make the console widget of virt-manager-0.5.0 working with the current > > VNC console setup when checking by using 127.0.0.1 for the hostname: > > QEMU now supports TLS + x509 certs in its VNC server. In Fedora I have also > patched the Xen & KVM forks of QEMU to have the same capability, so it is > secure for virt-manager to talk directly to the VNC server across an untrusted > LAN. I will be submitting the KVM & Xen backports of this code upstream in the > near future. I also intend to make virt-manager able to tunnel VNC over SSH > for scenarios where the VNC server isn't listening publically. > > > > > virtManager's console.py uses get_graphics_console(self) to get the > > directions for connecting to the graphics_console of the guest, so > > as far as I tested yet, that is fixed by checking if > > > > self.connection.get_hostname() > > > > equals > > > > self.connection.get_local_hostname(): > > > > and if that is so (both are the hostname of the local machine) then > > the VNC connection can be directed to 127.0.0.1 as before: > > Hmm, I think I'll change the get_hostname() method so you can pass a > flag indicating whether it should just return '127.0.0.1' for local > connections, which should have same effect as the patch you provide. It should be fixed with this changeset - let me know if there's still issues changeset: 586:4ee6526c32cd tag: tip user: "Daniel P. Berrange " date: Fri Aug 31 17:31:00 2007 -0400 summary: Ensure VNC widget always trys to connect to localhost for local connections Regards, Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From berrange at redhat.com Fri Aug 31 21:46:25 2007 From: berrange at redhat.com (Daniel P. Berrange) Date: Fri, 31 Aug 2007 22:46:25 +0100 Subject: [et-mgmt-tools] [PATCH] workarounds and fixes for new gtk-vnc integration in console.py In-Reply-To: References: Message-ID: <20070831214625.GG7881@redhat.com> On Fri, Aug 31, 2007 at 07:57:17PM +0200, Bernhard Kaindl wrote: > On Fri, 31 Aug 2007, Bernhard Kaindl wrote: > > > >Note: The message which was displayed when the gtk-vnc widget got the > >ECONNREFUSED (Connection refused) error when connecting to :5900 > >sounds a bit misleading to me. > > > >socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 4 > >connect(4, {sa_family=AF_INET, sin_port=htons(5900), > >sin_addr=inet_addr("")}, 16) = -1 > >ECONNREFUSED (Connection refused) > > > >The message resulting from this was: > > > >"Console was disconnected from guest" > > Another message like this still apppeared when starting a new QEMU / KVM > guest > from virt-manager, and the changes which I did in that regard to fix this > are found in the cumulated patch below. [snip] > Without this patch, a the console of a newly started KVM / QEMU guest is > only shown after > toggling pause if the console window was open while starting it, and also > some intermittent > messages which could cause unneccesary confusion are displayed. I can split > it up, but > it should be rather clear what it does. That's fine - it works as expected, so I've committed it. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=| From mdehaan at redhat.com Fri Aug 31 21:44:17 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 31 Aug 2007 17:44:17 -0400 Subject: [et-mgmt-tools] Cobbler auto-discovery and configuration of yum repos to mirror Message-ID: <46D88BB1.4040308@redhat.com> Ok, Don't get exited yet. This needs yum 3.2.4 which is available in F7-testing and probably will be in EL5.2. This means you probably don't have it. If you don't have it, cobbler will figure that out and will not run the command. Example (on my F7 laptop) [root at mdehaan cobbler]# python cobbler.py repo auto-add auto adding: Fedora7-i386 (http://download.fedora.devel.redhat.com/pub/fedora/linux/releases/7/Everything/i386/os) auto adding: Fedora7-i386-TestUpdates (http://download.fedora.devel.redhat.com/pub/fedora/linux/updates/testing/7/i386) auto adding: Fedora7-i386-Updates (http://download.fedora.devel.redhat.com/pub/fedora/linux/updates/7/i386) run cobbler reposync to apply changes [root at mdehaan cobbler]# What it does is look to see what yum repos you have configured on your boot server. It then creates a cobbler "repo" record for each one of them. This is a quick and easy way to make, say, a Fedora 7 boot server into a package mirror for the same content -- with just one command (ok, two if you count "cobbler reposync"). This probably does not work with yum-rhn-plugin just yet though I'll be testing that shortly. Checked in upstream and will be in 0.6.2. --Michael From mdehaan at redhat.com Fri Aug 31 21:56:21 2007 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 31 Aug 2007 17:56:21 -0400 Subject: [et-mgmt-tools] Alternative storage backends for Cobbler / other features Message-ID: <46D88E85.3050904@redhat.com> So there's been a lot of interest around having Cobbler read LDAP recently, and possibly I'm guessing it would be useful to have it /write/ to LDAP. I know several folks already have their own scripts to interface between the two. I'm thinking about abstracting the serializer code to allow for configs in arbitrary formats, though the existing YAML will be the default and will not require any configuration for existing installs -- or new ones. I will not taking your non-XML human-readable config files away from you :) Anyhow, the serializer stuff is already somewhat modular so I don't expect this to be terribly complicated. The hard part will be engineering things to not need to worry about schema upgrades. The DB options are mainly to keep queries fast as we scale up into thousands of system records. Future scaling work may also (probably) imply looking more towards OMAPI when dealing with ISC's dhcp.conf versus having to template out the file. If someone things that is needed (or even better, would like to work on that), please speak up. Most likely what would happen is I'll implement the framework for allowing arbitrary formats with a sqlite prototype, and if someone else wants to add in LDAP later that would be pretty easy to do by following the sqlite module's lead. The other thing on the radar is finally making the XMLRPC API bi-directional (by adding an additional secure version on another port) to make the life of webapps using the Cobbler API easier. I've been meaning to do that for a while. Until then apps that need write access to cobbler configs can go through the python API and/or the YAML tree. There was also a great suggestion about giving koan a very basic --register function, that would add an entry in the cobbler DB that tells the admin that he needs to set the system up. It would fill in the MAC, IP, and the hostname -- but that's it. This is probably going to be a bit further down the pipe than the above but it would be useful for cases where taking manual inventory of all the MACs in a datacenter would be a bit painful. We already have most of the code to do this from virt-factory, in fact, and we can add this into the existing Cobbler API. Comments? Questions? Ideas? --Michael