From skvidal at phy.duke.edu Wed Jun 1 05:15:14 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Wed, 01 Jun 2005 01:15:14 -0400 Subject: [Yum] yum in kickstart %post and HP Insight Mgr In-Reply-To: <20050531215549.GB23011@iwsp.com> References: <1117568770.3743.32.camel@localhost.localdomain> <20050531215550.GA14291@jadzia.bu.edu> <20050531215549.GB23011@iwsp.com> Message-ID: <1117602914.24314.29.camel@cutter> On Tue, 2005-05-31 at 17:55 -0400, Joshua Jensen wrote: > On Tue, May 31, 2005 at 05:55:50PM -0400, Matthew Miller wrote: > > > I know there's the "tsflags=noscripts" option in the config file -- is there > > a way to do that from the command line? Then, you could add any other bits > > of the %post script to your kickstart file. > > Funny you should ask: > > https://devel.linux.duke.edu/bugzilla/show_bug.cgi?id=462 > > > "RESOLVED WONTFIX" is the answer :-( I coulda resolved it as fixed in cvs. B/c it ultimately is - the plugin command line option adder lets you add whatever cli option(s) you want. but it's not in yum 2.2.X, only in yum 2.3.X and cvs HEAD. on the plus side, nothing in yum 2.3.X precludes it from running on systems that can run yum 2.2.X. You'll need to add a python-urlgrabber package and python-elementree. you don't even require sqlite if you don't want it. -sv From James_Martin at ao.uscourts.gov Wed Jun 1 13:23:31 2005 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Wed, 1 Jun 2005 09:23:31 -0400 Subject: [Yum] yum in kickstart %post and HP Insight Mgr In-Reply-To: <1117576742.18404.60.camel@cutter> Message-ID: HP has made a lot of crappy packages. And they still are. Red Hat might do us all a favor if they send them a copy of the "Red Hat RPM Guide". In a meeting with HP in March, I expressed this to the VP of Linux Engineering. What I've had to do is re-roll a lot of HP packages.. In my spec file I dissassemble the rpm using rpm2cpio and take the crap out of %post %pre that I don't want and put it back together. Seth is right, don't encourage them. Make a working package and email it back to them and tell them "this is how it should be done." The worst package I've ever seen was from Netvault-- Their rpm of their backup software basically dropped some files in /tmp and then there was %post script that ran the installer. Ugh, what a complete mess and waste. James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 05/31/2005 05:59:02 PM: > On Tue, 2005-05-31 at 17:55 -0400, Matthew Miller wrote: > > On Tue, May 31, 2005 at 03:46:10PM -0400, Brian Long wrote: > > > written RPMs, this works great. However, I've come across the lovely HP > > > Insight Manager RPMS (hpasm, hprsm, etc) which seem to run "service > > > hpasm start" during the RPM postinstall (yuck). This wouldn't be too > > [...] > > > Any ideas would be appreciated. Thanks. > > > > I know there's the "tsflags=noscripts" option in the config file -- is there > > a way to do that from the command line? Then, you could add any other bits > > of the %post script to your kickstart file. > > > > Don't encourage them. > > It does not make sense to add this to the cli in yum in order to work > around a broken package. > > -sv > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From gobbledegeek at gmail.com Wed Jun 1 14:03:43 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Wed, 1 Jun 2005 19:33:43 +0530 Subject: Automatic kickstart install Message-ID: <463aea57050601070360a7ddfb@mail.gmail.com> Hello I am creating a custom linux cd. I want the ks.cfg file on cdrom to be read automatically at boot without having to enter any command manually a boot prompt. How ? Duh ..! -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From Chris.McKenzie at entrust.com Wed Jun 1 17:38:56 2005 From: Chris.McKenzie at entrust.com (Chris Mckenzie) Date: Wed, 1 Jun 2005 13:38:56 -0400 Subject: Deparate - KickStart USB DVD device mount name? Message-ID: Hi all. I'm sort of desperate. I've got someone trying to install a kickstart ISO install I put together that has the linuxiso.cfg loading the kickstart from /mnt/cdrom, but from a USB DVD drive. It's a phone conversation, so you can understand that I can't see the kernel log. Does anyone know what mount location Anaconda mounts a USB DVD (or CD) drive when installing from it? Please and thank you! - Chris -------------- next part -------------- An HTML attachment was scrubbed... URL: From steve.mah at oracle.com Wed Jun 1 18:30:29 2005 From: steve.mah at oracle.com (Stephen Mah) Date: Wed, 01 Jun 2005 11:30:29 -0700 Subject: updating a distro tree Message-ID: <429DFEC5.2060007@oracle.com> Is there an easy way to update our existing kickstart distro tree to U5? Our current directory structure has u4 rpms and some third party rpms. I've extracted the U5 CDs already, but it's tedious finding duplicate packages. any suggestions. -regards steve From phr at doc.ic.ac.uk Wed Jun 1 18:53:32 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 1 Jun 2005 19:53:32 +0100 (BST) Subject: Automatic kickstart install In-Reply-To: <463aea57050601070360a7ddfb@mail.gmail.com> References: <463aea57050601070360a7ddfb@mail.gmail.com> Message-ID: On Wed, 1 Jun 2005, Gobbledegeek wrote: > I am creating a custom linux cd. I want the ks.cfg file on cdrom >to be read automatically at boot without having to enter any command >manually a boot prompt. Tell the bootloader (ISOLINUX) to include ks=cdrom:/path/to/kickstart_file in the kernel command line. (Example would go here if today weren't so busy.) Cheers, Phil From info at hostinthebox.net Wed Jun 1 18:54:45 2005 From: info at hostinthebox.net (dan) Date: Wed, 01 Jun 2005 11:54:45 -0700 Subject: Deparate - KickStart USB DVD device mount name? In-Reply-To: References: Message-ID: <429E0475.6090001@hostinthebox.net> Chris Mckenzie wrote: > Hi all. > > I'm sort of desperate. I've got someone trying to install a kickstart > ISO install I put together that has the linuxiso.cfg loading the > kickstart from /mnt/cdrom, but from a USB DVD drive. > > It's a phone conversation, so you can understand that I can't see the > kernel log. > > Does anyone know what mount location Anaconda mounts a USB DVD (or CD) > drive when installing from it? > > Please and thank you! > > - Chris > I haven't much experience with USB in Linux, but if I recall correctly, USB storage devices are mounted as psuedo-SCSI devices. So your device names would be /dev/sdXX (/dev/sda{1}, /dev/sdb{3}, etc etc). That will iether help you or hurt you. I only wish that it helps. haha. Thanks! -dant From phr at doc.ic.ac.uk Wed Jun 1 18:58:38 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 1 Jun 2005 19:58:38 +0100 (BST) Subject: Deparate - KickStart USB DVD device mount name? In-Reply-To: References: Message-ID: Please don't post HTML to mailing lists. On Wed, 1 Jun 2005, Chris Mckenzie wrote: >I'm sort of desperate. I've got someone trying to install a kickstart >ISO install I put together that has the linuxiso.cfg loading the >kickstart from /mnt/cdrom, but from a USB DVD drive. > >It's a phone conversation, so you can understand that I can't see the >kernel log. > >Does anyone know what mount location Anaconda mounts a USB DVD (or CD) >drive when installing from it? Two question about this in 5 minutes. Spooky... Does ks=cdrom:/path/to/kickstart_file do what you want? Cheers, Phil From bdoctor at ps-ax.com Wed Jun 1 19:18:22 2005 From: bdoctor at ps-ax.com (Brad Doctor) Date: Wed, 1 Jun 2005 13:18:22 -0600 (MDT) Subject: Automatic kickstart install In-Reply-To: Message-ID: <20050601191822.33210802E@bigbox.ps-ax.com> Here is an example from isolinux.cfg: default linux prompt 1 display general.msg label install kernel vmlinuz append ks=cdrom:/myKS.cfg initrd=initrd.img lang= devfs=nomount ramdisk_size=8192 > On Wed, 1 Jun 2005, Gobbledegeek wrote: > > > I am creating a custom linux cd. I want the ks.cfg file on cdrom > >to be read automatically at boot without having to enter any command > >manually a boot prompt. > > Tell the bootloader (ISOLINUX) to include > ks=cdrom:/path/to/kickstart_file in the kernel command line. > > (Example would go here if today weren't so busy.) > > > Cheers, > Phil > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Brad Doctor, CISSP From Chris.McKenzie at entrust.com Wed Jun 1 19:07:32 2005 From: Chris.McKenzie at entrust.com (Chris Mckenzie) Date: Wed, 1 Jun 2005 15:07:32 -0400 Subject: Deparate - KickStart USB DVD device mount name? Message-ID: Well it's not so much USB in Linux as it is Anaconda installing from a USB drive. With an IDE/SCSI drive, Anaconda mounts /mnt/cdrom. It has to be different with a USB drive, because I can't find my kickstart from /mnt/cdrom. I think it's /media/usbdisk, so I've created a new kickstart and isolinux.cfg for them to try. If anyone can confirm /media/usbdisk, it would be re-assuring. Thanks! - Chris -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of dan Sent: June 1, 2005 2:55 PM To: Discussion list about Kickstart Subject: Re: Deparate - KickStart USB DVD device mount name? Chris Mckenzie wrote: > Hi all. > > I'm sort of desperate. I've got someone trying to install a kickstart > ISO install I put together that has the linuxiso.cfg loading the > kickstart from /mnt/cdrom, but from a USB DVD drive. > > It's a phone conversation, so you can understand that I can't see the > kernel log. > > Does anyone know what mount location Anaconda mounts a USB DVD (or CD) > drive when installing from it? > > Please and thank you! > > - Chris > I haven't much experience with USB in Linux, but if I recall correctly, USB storage devices are mounted as psuedo-SCSI devices. So your device names would be /dev/sdXX (/dev/sda{1}, /dev/sdb{3}, etc etc). That will iether help you or hurt you. I only wish that it helps. haha. Thanks! -dant _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From tibbs at math.uh.edu Wed Jun 1 19:11:32 2005 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Wed, 01 Jun 2005 14:11:32 -0500 Subject: updating a distro tree In-Reply-To: <429DFEC5.2060007@oracle.com> (Stephen Mah's message of "Wed, 01 Jun 2005 11:30:29 -0700") References: <429DFEC5.2060007@oracle.com> Message-ID: >>>>> "SM" == Stephen Mah writes: SM> Is there an easy way to update our existing kickstart distro tree SM> to U5? Here's a link to the script I use; it's been hacked a bit and I only use Fedora but you may find it useful. - J< From rspaulding at mail.arc.nasa.gov Wed Jun 1 19:59:00 2005 From: rspaulding at mail.arc.nasa.gov (Ryan C. Spaulding) Date: Wed, 1 Jun 2005 12:59:00 -0700 Subject: 3.0 to 4.0 preserving existing raid and LVM volumes Message-ID: Hi, I am currently trying to upgrade some AS 3.x systems to WS 4.x. Since most of the systems I am rebuilding have a 200+ Gigs on some of the volumes I would like to just re-kickstart the machines and preserve the existing raid and LVM partitions. Below is the kickstart configuration of the machine (3.0 AS) currently: clearpart --all --initlabel #Disk partitioning information part raid.00 --size 128 --ondisk sda part raid.01 --size 128 --ondisk sdb part raid.02 --size 15000 --grow --ondisk sda part raid.03 --size 15000 --grow --ondisk sdb raid /boot --level 1 --device md0 raid.00 raid.01 raid pv.00 --level 1 --device md1 raid.02 raid.03 volgroup vgsys pv.00 logvol / --vgname vgsys --name lvroot --size 10240 logvol /usr --vgname vgsys --name lvusr --size 8192 logvol /var --vgname vgsys --name lvvar --size 4096 logvol swap --vgname vgsys --name lvswap --size 1024 --fstype swap logvol /web --vgname vgsys --name lvweb --size 122880 Now this works with a clean install of 3.0 AS or 4.0 WS. When I try using the --noformat option anywhere the install either crashes or errors out with "could not create requested partitions" when I try and install on top of the AS installation. Thank you for any and all help. Ryan From tibbs at math.uh.edu Wed Jun 1 20:07:29 2005 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Wed, 01 Jun 2005 15:07:29 -0500 Subject: updating a distro tree In-Reply-To: (Jason L. Tibbitts, III's message of "Wed, 01 Jun 2005 14:11:32 -0500") References: <429DFEC5.2060007@oracle.com> Message-ID: >>>>> "JLT" == Jason L Tibbitts, writes: JLT> Here's a link to the script I use; it's been hacked a bit and I JLT> only use Fedora but you may find it useful. Bah, how did I forget to paste that in? http://www.math.uh.edu/~tibbs/do_update - J< From klaus.steden at thomson.net Wed Jun 1 21:00:23 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 1 Jun 2005 17:00:23 -0400 Subject: updating a distro tree In-Reply-To: References: <429DFEC5.2060007@oracle.com> Message-ID: <20050601210023.GN16989@thomson.net> > >>>>> "JLT" == Jason L Tibbitts, writes: > > JLT> Here's a link to the script I use; it's been hacked a bit and I > JLT> only use Fedora but you may find it useful. > > Bah, how did I forget to paste that in? > > http://www.math.uh.edu/~tibbs/do_update > Jason - I'm not much for python (yet) - could this script be used to incorporate third-party apps into a distribution? I am currently sorting out how to go about this process, since I need some other apps like ttcp, cfengine, and apt-get to be installed on my RH systems. You can reply off list if it's easier. cheers, Klaus From tibbs at math.uh.edu Wed Jun 1 23:42:02 2005 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Wed, 01 Jun 2005 18:42:02 -0500 Subject: updating a distro tree In-Reply-To: <20050601210023.GN16989@thomson.net> (Klaus Steden's message of "Wed, 1 Jun 2005 17:00:23 -0400") References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> Message-ID: >>>>> "KS" == Klaus Steden writes: KS> I'm not much for python (yet) Frankly, neither am I, but I guess that's neither here nor there. KS> could this script be used to incorporate third-party apps into a KS> distribution? Assuming those apps are in the form of RPMS, you just need to put the packages into the RPMS directory and then run genhdlist and maybe pkgorder. do_update normally does that if it updated any packages but you can force it with -G (or just look near the end of the script to see how to do it yourself). But generally I prefer not to mess with adding extra applications to the install tree and instead put them in a yum repo to be installed in %post or on the first boot. - J< From brilong at cisco.com Thu Jun 2 13:24:10 2005 From: brilong at cisco.com (Brian Long) Date: Thu, 02 Jun 2005 09:24:10 -0400 Subject: VMWare vmxnet integration? Message-ID: <1117718650.29766.27.camel@localhost.localdomain> Hello, I'm wondering if anyone has been successful in integrating VMWare's vmxnet driver into the install images for RHEL 3 or RHEL 4. VMWare ESX has two NIC emulations: vlance (pcnet32 low performance) and vmxnet (requires a VMWare-provided driver). vmxnet is much faster and is the preferred NIC when deploying virtual machines into production. We'd like to be able to kickstart VM's using the vmxnet driver. Otherwise, you have to configure the VM for vlance, kickstart, install VMWare's vmxnet driver, configure the VM for vmxnet, etc. We're trying to slipstream the vmxnet driver into the kickstart process. We're familiar with running buildinstall to remake the boot images, initrd, etc. but have been unsuccessful integrating the vmxnet driver. Thanks for any pointers. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From kickstart at lordcow.org Thu Jun 2 15:51:19 2005 From: kickstart at lordcow.org (gareth) Date: Thu, 2 Jun 2005 17:51:19 +0200 Subject: death Message-ID: <20050602155119.GA25662@lordcow.org> yo, i just installed fedora core 3, and then used the anaconda.cfg created as the kickstart file (and uncommenting the partions etc.) for reinstalling on the same machine. the install dies tho with: Traceback (most recent call last): File "/usr/bin/anaconda", line 1165, in ? intf.run(id, dispatch, configFileData) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 705, in run self.icw.run (self.runres, configFileData) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1449, in run self.setup_window(runres) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1421, in setup_window self.setScreen () File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1086, in setScreen (step, args) = self.dispatch.currentStep() File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 276, in currentStep self.gotoNext() File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 171, in gotoNext self.moveStep() File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 239, in moveStep rc = apply(func, self.bindArgs(args)) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1416, in doAutoPartition doPartitioning(diskset, partitions, doRefresh = 0) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1003, in doPartitioning (ret, msg) = processPartitioning(diskset, requests, newParts) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 953, in processPartitioning ret = fitSized(diskset, requests, 0, newParts) File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 361, in fitSized requestSectors = long((request.requestSize * 1024L * 1024L) / part.disk.dev.sector_size) - 1 TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' any clues? From ajaymulwani at gmail.com Thu Jun 2 16:35:20 2005 From: ajaymulwani at gmail.com (Ajay Mulwani) Date: Thu, 2 Jun 2005 22:05:20 +0530 Subject: death In-Reply-To: <20050602155119.GA25662@lordcow.org> References: <20050602155119.GA25662@lordcow.org> Message-ID: <5520b27905060209355d6f839c@mail.gmail.com> > (and uncommenting the partions etc.) ---> ?? Please post the anaconda.cfg file which was used for renstalling.. Ajay On 6/2/05, gareth wrote: > yo, i just installed fedora core 3, and then used the anaconda.cfg > created as the kickstart file (and uncommenting the partions etc.) > for reinstalling on the same machine. the install dies tho with: > > Traceback (most recent call last): > File "/usr/bin/anaconda", line 1165, in ? > intf.run(id, dispatch, configFileData) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 705, in run > self.icw.run (self.runres, configFileData) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1449, in run > self.setup_window(runres) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1421, in setup_window > self.setScreen () > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1086, in setScreen > (step, args) = self.dispatch.currentStep() > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 276, in currentStep > self.gotoNext() > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 171, in gotoNext > self.moveStep() > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 239, in moveStep > rc = apply(func, self.bindArgs(args)) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1416, in doAutoPartition > doPartitioning(diskset, partitions, doRefresh = 0) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1003, in doPartitioning > (ret, msg) = processPartitioning(diskset, requests, newParts) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 953, in processPartitioning > ret = fitSized(diskset, requests, 0, newParts) > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 361, in fitSized > requestSectors = long((request.requestSize * 1024L * 1024L) / part.disk.dev.sector_size) - 1 > TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' > > any clues? > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From kickstart at lordcow.org Thu Jun 2 16:42:25 2005 From: kickstart at lordcow.org (gareth) Date: Thu, 2 Jun 2005 18:42:25 +0200 Subject: death In-Reply-To: <5520b27905060209355d6f839c@mail.gmail.com> References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> Message-ID: <20050602164225.GA27011@lordcow.org> sure On Thu 2005-06-02 (22:05), Ajay Mulwani wrote: > Please post the anaconda.cfg file which was used for renstalling.. > > Ajay > > > On 6/2/05, gareth wrote: > > yo, i just installed fedora core 3, and then used the anaconda.cfg > > created as the kickstart file (and uncommenting the partions etc.) > > for reinstalling on the same machine. the install dies tho with: > > > > Traceback (most recent call last): > > File "/usr/bin/anaconda", line 1165, in ? > > intf.run(id, dispatch, configFileData) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 705, in run > > self.icw.run (self.runres, configFileData) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1449, in run > > self.setup_window(runres) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1421, in setup_window > > self.setScreen () > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1086, in setScreen > > (step, args) = self.dispatch.currentStep() > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 276, in currentStep > > self.gotoNext() > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 171, in gotoNext > > self.moveStep() > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 239, in moveStep > > rc = apply(func, self.bindArgs(args)) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1416, in doAutoPartition > > doPartitioning(diskset, partitions, doRefresh = 0) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1003, in doPartitioning > > (ret, msg) = processPartitioning(diskset, requests, newParts) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 953, in processPartitioning > > ret = fitSized(diskset, requests, 0, newParts) > > File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 361, in fitSized > > requestSectors = long((request.requestSize * 1024L * 1024L) / part.disk.dev.sector_size) - 1 > > TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' > > > > any clues? > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- install cdrom lang en_US.UTF-8 langsupport --default=en_ZA.UTF-8 en_ZA.UTF-8 keyboard us xconfig --card "Trident Blade3D (generic)" --videoram 8192 --hsync 31.5-37.9 --vsync 50-70 --resolution 800x600 --depth 16 --startxonboot --defaultdesktop kde network --device eth0 --bootproto dhcp network --device eth1 --bootproto static --ip 172.16.1.1 --netmask 255.255.255.0 rootpw --iscrypted $1$etc firewall --disabled selinux --enforcing authconfig --enableshadow --enablemd5 timezone Africa/Johannesburg bootloader --location=mbr --append="rhgb quiet" --md5pass=$1$etc clearpart --all --drives=sda part /boot --fstype ext3 --size=64 part /home --fstype ext3 --size=32768 part pv.2 --size=0 --grow --ondisk=sda part /data --fstype ext3 --size=8192 part /var --fstype ext3 --size=384 part /tmp --fstype ext3 --size=256 volgroup VolGroup00 pv.2 logvol / --fstype ext3 --name=LogVol00 --vgname=VolGroup00 --size=26240 logvol swap --fstype swap --name=LogVol01 --vgname=VolGroup00 --size=2048 %packages @ editors @ system-tools @ text-internet @ dialup @ emacs @ base-x @ web-server @ kde-desktop @ development-tools -kdepim joe -kdemultimedia subversion -ckermit -gdm -samba-client -tux -authconfig-gtk -ethereal -kdeartwork -up2date-gnome -webalizer -zsh -vnc-server -autorun -squid -usermode-gtk -system-config-network -system-config-printer-gui mod_auth_mysql -bluez-pin -slrn kernel -system-config-display -desktop-backgrounds-extra -elinks php-mysql -php-ldap -xdelta e2fsprogs -system-config-date grub -open -cadaver -fetchmail -openldap-clients -system-config-packages -system-config-users lvm2 -system-logviewer -system-config-soundcard %post From SSeremeth at anacomp.com Thu Jun 2 17:43:19 2005 From: SSeremeth at anacomp.com (Seremeth, Stephen) Date: Thu, 2 Jun 2005 13:43:19 -0400 Subject: Kickstart dies after post-install Message-ID: <66911C8E1F2DA24EB29DE289297A135E4856E0@usrd104> So, chalk this up to another "I spent way too long trying to figure out why anaconda acts a certain way" type issue, but I just saw my first congratulations screen since kickstarting my own custom ES 4 installers. In previous versions (and maybe it was just AS 2.1), this worked fine in the %post section: /sbin/service network start /sbin/service portmap start ifconfig $ETH_ACTIVE $ip up /usr/sbin/ntpdate $NTPSERVER In ES 4, these commands somehow (haven't figured out how yet) cause the install to crash (throws a traceback referenced in previous email) after the post section has been run and before the congrats screen. I'm hoping this email will hopefully save someone else the agony I have just lived through trying to debug this. Btw - I was getting the same problems on two types of hardware doing PXE nfs installs, and the hardware checked out fine. Regards, Steve From info at hostinthebox.net Thu Jun 2 18:18:54 2005 From: info at hostinthebox.net (dan) Date: Thu, 02 Jun 2005 11:18:54 -0700 Subject: Kickstart dies after post-install In-Reply-To: <66911C8E1F2DA24EB29DE289297A135E4856E0@usrd104> References: <66911C8E1F2DA24EB29DE289297A135E4856E0@usrd104> Message-ID: <429F4D8E.3000101@hostinthebox.net> Seremeth, Stephen wrote: > So, chalk this up to another "I spent way too long trying to figure out > why anaconda acts a certain way" type issue, but I just saw my first > congratulations screen since kickstarting my own custom ES 4 installers. > > In previous versions (and maybe it was just AS 2.1), this worked fine in > the %post section: > > /sbin/service network start > /sbin/service portmap start > ifconfig $ETH_ACTIVE $ip up > /usr/sbin/ntpdate $NTPSERVER > > In ES 4, these commands somehow (haven't figured out how yet) cause the > install to crash (throws a traceback referenced in previous email) after > the post section has been run and before the congrats screen. > > I'm hoping this email will hopefully save someone else the agony I have > just lived through trying to debug this. Btw - I was getting the same > problems on two types of hardware doing PXE nfs installs, and the > hardware checked out fine. > > Regards, > > Steve > Steve - Sorry for not keeping current on this issue, so please forgive my ignorance, but I have a few questions that might yield solutions. Are $ETH_ACTIVE and $ip defined variables? What about $NTPSERVER? These may very well be set by Anaconda at install-time, but if they're not, and you don't set them, this would cause a problem, no? Thanks -dant From SSeremeth at anacomp.com Thu Jun 2 20:11:50 2005 From: SSeremeth at anacomp.com (Seremeth, Stephen) Date: Thu, 2 Jun 2005 16:11:50 -0400 Subject: Kickstart dies after post-install Message-ID: <66911C8E1F2DA24EB29DE289297A135E4856E1@usrd104> > > Steve - > > Sorry for not keeping current on this issue, so please forgive my > ignorance, but I have a few questions that might yield solutions. > > Are $ETH_ACTIVE and $ip defined variables? What about $NTPSERVER? > These may very well be set by Anaconda at install-time, but > if they're > not, and you don't set them, this would cause a problem, no? > > Thanks > -dant I define them in a previous part of %post. And further, I see no errors from those command being run. From error27 at gmail.com Fri Jun 3 00:08:40 2005 From: error27 at gmail.com (Dan Carpenter) Date: Thu, 2 Jun 2005 17:08:40 -0700 Subject: death In-Reply-To: <20050602164225.GA27011@lordcow.org> References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> <20050602164225.GA27011@lordcow.org> Message-ID: This is completely wild guess. What happens if you change old: part pv.2 --size=0 --grow --ondisk=sda new: part pv.2 --size=1 --grow --ondisk=sda Anyway it has something to do with partitioning... Perhaps the existing partitions are screwing it up. I've seen that happen quite often with software RAID for example. regards, dan carpenter From dky at utcc.utoronto.ca Fri Jun 3 00:24:45 2005 From: dky at utcc.utoronto.ca (Derek Yeung) Date: Thu, 2 Jun 2005 20:24:45 -0400 (EDT) Subject: death In-Reply-To: <20050602164225.GA27011@lordcow.org> References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> <20050602164225.GA27011@lordcow.org> Message-ID: Looks like you're trying to use a kickstart configuration file and trying to apply a partition spec for a disk with a different geometery/size. Your partitioning scheme looks simple enough -- blank out all the automatic partitioning lines and then try the configuration file again, but manually configuring the partitions yourself. /dky On Thu, 2 Jun 2005, gareth wrote: > sure > > On Thu 2005-06-02 (22:05), Ajay Mulwani wrote: >> Please post the anaconda.cfg file which was used for renstalling.. >> >> Ajay >> >> >> On 6/2/05, gareth wrote: >>> yo, i just installed fedora core 3, and then used the anaconda.cfg >>> created as the kickstart file (and uncommenting the partions etc.) >>> for reinstalling on the same machine. the install dies tho with: >>> >>> Traceback (most recent call last): >>> File "/usr/bin/anaconda", line 1165, in ? >>> intf.run(id, dispatch, configFileData) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 705, in run >>> self.icw.run (self.runres, configFileData) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1449, in run >>> self.setup_window(runres) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1421, in setup_window >>> self.setScreen () >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", line 1086, in setScreen >>> (step, args) = self.dispatch.currentStep() >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 276, in currentStep >>> self.gotoNext() >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 171, in gotoNext >>> self.moveStep() >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", line 239, in moveStep >>> rc = apply(func, self.bindArgs(args)) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1416, in doAutoPartition >>> doPartitioning(diskset, partitions, doRefresh = 0) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 1003, in doPartitioning >>> (ret, msg) = processPartitioning(diskset, requests, newParts) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 953, in processPartitioning >>> ret = fitSized(diskset, requests, 0, newParts) >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", line 361, in fitSized >>> requestSectors = long((request.requestSize * 1024L * 1024L) / part.disk.dev.sector_size) - 1 >>> TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' >>> >>> any clues? >>> >>> _______________________________________________ >>> Kickstart-list mailing list >>> Kickstart-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/kickstart-list >>> >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list > From kickstart at lordcow.org Fri Jun 3 09:50:12 2005 From: kickstart at lordcow.org (gareth) Date: Fri, 3 Jun 2005 11:50:12 +0200 Subject: death In-Reply-To: References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> <20050602164225.GA27011@lordcow.org> Message-ID: <20050603095012.GA7689@lordcow.org> it's the same machine & disk i got the anaconda.cfg file off. ok i commented out the partition steps & the installation now runs fine, except i haveta key in the partitions. so how do i go about getting the partitioning right in the ks.cfg? On Thu 2005-06-02 (20:24), Derek Yeung wrote: > Looks like you're trying to use a kickstart configuration file and trying > to apply a partition spec for a disk with a different geometery/size. > Your partitioning scheme looks simple enough -- blank out all the > automatic partitioning lines and then try the configuration file again, > but manually configuring the partitions yourself. > > > /dky > > > > On Thu, 2 Jun 2005, gareth wrote: > > >sure > > > >On Thu 2005-06-02 (22:05), Ajay Mulwani wrote: > >>Please post the anaconda.cfg file which was used for renstalling.. > >> > >>Ajay > >> > >> > >>On 6/2/05, gareth wrote: > >>>yo, i just installed fedora core 3, and then used the anaconda.cfg > >>>created as the kickstart file (and uncommenting the partions etc.) > >>>for reinstalling on the same machine. the install dies tho with: > >>> > >>>Traceback (most recent call last): > >>> File "/usr/bin/anaconda", line 1165, in ? > >>> intf.run(id, dispatch, configFileData) > >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>> line 705, in run > >>> self.icw.run (self.runres, configFileData) > >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>> line 1449, in run > >>> self.setup_window(runres) > >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>> line 1421, in setup_window > >>> self.setScreen () > >>> File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>> line 1086, in setScreen > >>> (step, args) = self.dispatch.currentStep() > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", > >>> line 276, in currentStep > >>> self.gotoNext() > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", > >>> line 171, in gotoNext > >>> self.moveStep() > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", > >>> line 239, in moveStep > >>> rc = apply(func, self.bindArgs(args)) > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>> line 1416, in doAutoPartition > >>> doPartitioning(diskset, partitions, doRefresh = 0) > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>> line 1003, in doPartitioning > >>> (ret, msg) = processPartitioning(diskset, requests, newParts) > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>> line 953, in processPartitioning > >>> ret = fitSized(diskset, requests, 0, newParts) > >>> File > >>> "/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>> line 361, in fitSized > >>> requestSectors = long((request.requestSize * 1024L * 1024L) / > >>> part.disk.dev.sector_size) - 1 > >>>TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' > >>> > >>>any clues? > >>> > >>>_______________________________________________ > >>>Kickstart-list mailing list > >>>Kickstart-list at redhat.com > >>>https://www.redhat.com/mailman/listinfo/kickstart-list > >>> > >> > >>_______________________________________________ > >>Kickstart-list mailing list > >>Kickstart-list at redhat.com > >>https://www.redhat.com/mailman/listinfo/kickstart-list > > From dan at half-asleep.com Fri Jun 3 13:15:24 2005 From: dan at half-asleep.com (Daniel Segall) Date: Fri, 03 Jun 2005 09:15:24 -0400 Subject: death In-Reply-To: <20050603095012.GA7689@lordcow.org> References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> <20050602164225.GA27011@lordcow.org> <20050603095012.GA7689@lordcow.org> Message-ID: <42A057EC.1080007@half-asleep.com> I usually just comment out the zero mbr and clearpart options in my ks.cfg, then just use something like: / --fstype ext3 --onpart=hda1 swap --fstype swap --onpart=hda2 That is from memory, so I'd verify the format in the kickstart manual. -Dan gareth wrote: > it's the same machine & disk i got the anaconda.cfg file off. > ok i commented out the partition steps & the installation > now runs fine, except i haveta key in the partitions. > so how do i go about getting the partitioning right in the > ks.cfg? > > On Thu 2005-06-02 (20:24), Derek Yeung wrote: > >>Looks like you're trying to use a kickstart configuration file and trying >>to apply a partition spec for a disk with a different geometery/size. >>Your partitioning scheme looks simple enough -- blank out all the >>automatic partitioning lines and then try the configuration file again, >>but manually configuring the partitions yourself. >> >> >>/dky >> >> >> >>On Thu, 2 Jun 2005, gareth wrote: >> >> >>>sure >>> >>>On Thu 2005-06-02 (22:05), Ajay Mulwani wrote: >>> >>>>Please post the anaconda.cfg file which was used for renstalling.. >>>> >>>>Ajay >>>> >>>> >>>>On 6/2/05, gareth wrote: >>>> >>>>>yo, i just installed fedora core 3, and then used the anaconda.cfg >>>>>created as the kickstart file (and uncommenting the partions etc.) >>>>>for reinstalling on the same machine. the install dies tho with: >>>>> >>>>>Traceback (most recent call last): >>>>>File "/usr/bin/anaconda", line 1165, in ? >>>>> intf.run(id, dispatch, configFileData) >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >>>>>line 705, in run >>>>> self.icw.run (self.runres, configFileData) >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >>>>>line 1449, in run >>>>> self.setup_window(runres) >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >>>>>line 1421, in setup_window >>>>> self.setScreen () >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >>>>>line 1086, in setScreen >>>>> (step, args) = self.dispatch.currentStep() >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", >>>>>line 276, in currentStep >>>>> self.gotoNext() >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", >>>>>line 171, in gotoNext >>>>> self.moveStep() >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", >>>>>line 239, in moveStep >>>>> rc = apply(func, self.bindArgs(args)) >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >>>>>line 1416, in doAutoPartition >>>>> doPartitioning(diskset, partitions, doRefresh = 0) >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >>>>>line 1003, in doPartitioning >>>>> (ret, msg) = processPartitioning(diskset, requests, newParts) >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >>>>>line 953, in processPartitioning >>>>> ret = fitSized(diskset, requests, 0, newParts) >>>>>File >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >>>>>line 361, in fitSized >>>>> requestSectors = long((request.requestSize * 1024L * 1024L) / >>>>> part.disk.dev.sector_size) - 1 >>>>>TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' >>>>> >>>>>any clues? >>>>> >>>>>_______________________________________________ >>>>>Kickstart-list mailing list >>>>>Kickstart-list at redhat.com >>>>>https://www.redhat.com/mailman/listinfo/kickstart-list >>>>> >>>> >>>>_______________________________________________ >>>>Kickstart-list mailing list >>>>Kickstart-list at redhat.com >>>>https://www.redhat.com/mailman/listinfo/kickstart-list >>> > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From kickstart at lordcow.org Fri Jun 3 13:36:47 2005 From: kickstart at lordcow.org (gareth) Date: Fri, 3 Jun 2005 15:36:47 +0200 Subject: death In-Reply-To: <42A057EC.1080007@half-asleep.com> References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> <20050602164225.GA27011@lordcow.org> <20050603095012.GA7689@lordcow.org> <42A057EC.1080007@half-asleep.com> Message-ID: <20050603133647.GA3250@lordcow.org> don't i need clearpart if i wanna override the existing partitions? On Fri 2005-06-03 (09:15), Daniel Segall wrote: > I usually just comment out the zero mbr and clearpart options in my > ks.cfg, then just use something like: > / --fstype ext3 --onpart=hda1 > swap --fstype swap --onpart=hda2 > > That is from memory, so I'd verify the format in the kickstart manual. > > -Dan > > gareth wrote: > >it's the same machine & disk i got the anaconda.cfg file off. > >ok i commented out the partition steps & the installation > >now runs fine, except i haveta key in the partitions. > >so how do i go about getting the partitioning right in the > >ks.cfg? > > > >On Thu 2005-06-02 (20:24), Derek Yeung wrote: > > > >>Looks like you're trying to use a kickstart configuration file and trying > >>to apply a partition spec for a disk with a different geometery/size. > >>Your partitioning scheme looks simple enough -- blank out all the > >>automatic partitioning lines and then try the configuration file again, > >>but manually configuring the partitions yourself. > >> > >> > >>/dky > >> > >> > >> > >>On Thu, 2 Jun 2005, gareth wrote: > >> > >> > >>>sure > >>> > >>>On Thu 2005-06-02 (22:05), Ajay Mulwani wrote: > >>> > >>>>Please post the anaconda.cfg file which was used for renstalling.. > >>>> > >>>>Ajay > >>>> > >>>> > >>>>On 6/2/05, gareth wrote: > >>>> > >>>>>yo, i just installed fedora core 3, and then used the anaconda.cfg > >>>>>created as the kickstart file (and uncommenting the partions etc.) > >>>>>for reinstalling on the same machine. the install dies tho with: > >>>>> > >>>>>Traceback (most recent call last): > >>>>>File "/usr/bin/anaconda", line 1165, in ? > >>>>> intf.run(id, dispatch, configFileData) > >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>>>>line 705, in run > >>>>> self.icw.run (self.runres, configFileData) > >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>>>>line 1449, in run > >>>>> self.setup_window(runres) > >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>>>>line 1421, in setup_window > >>>>> self.setScreen () > >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", > >>>>>line 1086, in setScreen > >>>>> (step, args) = self.dispatch.currentStep() > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", > >>>>>line 276, in currentStep > >>>>> self.gotoNext() > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", > >>>>>line 171, in gotoNext > >>>>> self.moveStep() > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", > >>>>>line 239, in moveStep > >>>>> rc = apply(func, self.bindArgs(args)) > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>>>>line 1416, in doAutoPartition > >>>>> doPartitioning(diskset, partitions, doRefresh = 0) > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>>>>line 1003, in doPartitioning > >>>>> (ret, msg) = processPartitioning(diskset, requests, newParts) > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>>>>line 953, in processPartitioning > >>>>> ret = fitSized(diskset, requests, 0, newParts) > >>>>>File > >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", > >>>>>line 361, in fitSized > >>>>> requestSectors = long((request.requestSize * 1024L * 1024L) / > >>>>> part.disk.dev.sector_size) - 1 > >>>>>TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' > >>>>> > >>>>>any clues? > >>>>> > >>>>>_______________________________________________ > >>>>>Kickstart-list mailing list > >>>>>Kickstart-list at redhat.com > >>>>>https://www.redhat.com/mailman/listinfo/kickstart-list > >>>>> > >>>> > >>>>_______________________________________________ > >>>>Kickstart-list mailing list > >>>>Kickstart-list at redhat.com > >>>>https://www.redhat.com/mailman/listinfo/kickstart-list > >>> > > > >_______________________________________________ > >Kickstart-list mailing list > >Kickstart-list at redhat.com > >https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From dan at half-asleep.com Fri Jun 3 16:36:16 2005 From: dan at half-asleep.com (Daniel Segall) Date: Fri, 3 Jun 2005 12:36:16 -0400 (EDT) Subject: death In-Reply-To: <20050603133647.GA3250@lordcow.org> References: <20050602155119.GA25662@lordcow.org> <5520b27905060209355d6f839c@mail.gmail.com> <20050602164225.GA27011@lordcow.org> <20050603095012.GA7689@lordcow.org> <42A057EC.1080007@half-asleep.com> <20050603133647.GA3250@lordcow.org> Message-ID: <50758.192.80.55.73.1117816576.squirrel@webmail.half-asleep.com> No! You do not want that. Clearpart REMOVES the partitions. The following is from the manual: clearpart (optional) Removes partitions from the system, prior to creation of new partitions. By default, no partitions are removed. Note If the clearpart command is used, then the --onpart command cannot be used on a logical partition -Dan > don't i need clearpart if i wanna override the existing partitions? > > On Fri 2005-06-03 (09:15), Daniel Segall wrote: >> I usually just comment out the zero mbr and clearpart options in my >> ks.cfg, then just use something like: >> / --fstype ext3 --onpart=hda1 >> swap --fstype swap --onpart=hda2 >> >> That is from memory, so I'd verify the format in the kickstart manual. >> >> -Dan >> >> gareth wrote: >> >it's the same machine & disk i got the anaconda.cfg file off. >> >ok i commented out the partition steps & the installation >> >now runs fine, except i haveta key in the partitions. >> >so how do i go about getting the partitioning right in the >> >ks.cfg? >> > >> >On Thu 2005-06-02 (20:24), Derek Yeung wrote: >> > >> >>Looks like you're trying to use a kickstart configuration file and >> trying >> >>to apply a partition spec for a disk with a different geometery/size. >> >>Your partitioning scheme looks simple enough -- blank out all the >> >>automatic partitioning lines and then try the configuration file >> again, >> >>but manually configuring the partitions yourself. >> >> >> >> >> >>/dky >> >> >> >> >> >> >> >>On Thu, 2 Jun 2005, gareth wrote: >> >> >> >> >> >>>sure >> >>> >> >>>On Thu 2005-06-02 (22:05), Ajay Mulwani wrote: >> >>> >> >>>>Please post the anaconda.cfg file which was used for renstalling.. >> >>>> >> >>>>Ajay >> >>>> >> >>>> >> >>>>On 6/2/05, gareth wrote: >> >>>> >> >>>>>yo, i just installed fedora core 3, and then used the anaconda.cfg >> >>>>>created as the kickstart file (and uncommenting the partions etc.) >> >>>>>for reinstalling on the same machine. the install dies tho with: >> >>>>> >> >>>>>Traceback (most recent call last): >> >>>>>File "/usr/bin/anaconda", line 1165, in ? >> >>>>> intf.run(id, dispatch, configFileData) >> >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >> >>>>>line 705, in run >> >>>>> self.icw.run (self.runres, configFileData) >> >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >> >>>>>line 1449, in run >> >>>>> self.setup_window(runres) >> >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >> >>>>>line 1421, in setup_window >> >>>>> self.setScreen () >> >>>>>File "/usr/src/build/475969-i386/install//usr/lib/anaconda/gui.py", >> >>>>>line 1086, in setScreen >> >>>>> (step, args) = self.dispatch.currentStep() >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", >> >>>>>line 276, in currentStep >> >>>>> self.gotoNext() >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", >> >>>>>line 171, in gotoNext >> >>>>> self.moveStep() >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/dispatch.py", >> >>>>>line 239, in moveStep >> >>>>> rc = apply(func, self.bindArgs(args)) >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >> >>>>>line 1416, in doAutoPartition >> >>>>> doPartitioning(diskset, partitions, doRefresh = 0) >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >> >>>>>line 1003, in doPartitioning >> >>>>> (ret, msg) = processPartitioning(diskset, requests, newParts) >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >> >>>>>line 953, in processPartitioning >> >>>>> ret = fitSized(diskset, requests, 0, newParts) >> >>>>>File >> >>>>>"/usr/src/build/475969-i386/install//usr/lib/anaconda/autopart.py", >> >>>>>line 361, in fitSized >> >>>>> requestSectors = long((request.requestSize * 1024L * 1024L) / >> >>>>> part.disk.dev.sector_size) - 1 >> >>>>>TypeError: unsupported operand type(s) for *: 'NoneType' and 'long' >> >>>>> >> >>>>>any clues? >> >>>>> >> >>>>>_______________________________________________ >> >>>>>Kickstart-list mailing list >> >>>>>Kickstart-list at redhat.com >> >>>>>https://www.redhat.com/mailman/listinfo/kickstart-list >> >>>>> >> >>>> >> >>>>_______________________________________________ >> >>>>Kickstart-list mailing list >> >>>>Kickstart-list at redhat.com >> >>>>https://www.redhat.com/mailman/listinfo/kickstart-list >> >>> >> > >> >_______________________________________________ >> >Kickstart-list mailing list >> >Kickstart-list at redhat.com >> >https://www.redhat.com/mailman/listinfo/kickstart-list >> > >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > From mschillinger at wellinx.com Fri Jun 3 17:56:37 2005 From: mschillinger at wellinx.com (Matt Schillinger) Date: Fri, 03 Jun 2005 12:56:37 -0500 Subject: Centos and eject Message-ID: <1117821396.3859.30.camel@dhcp66.wellinx.com> Is there no way to eject a cd in Centos? We have configs that did eject on /dev/hd[a-f] (in %post) from rh9 and fc2, and they worked fine. But in centos, it completely ignores it.. Am I missing something? M a t t S c h i l l i n g e r mschillinger at wellinx.com From klaus.steden at thomson.net Fri Jun 3 22:36:32 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 3 Jun 2005 18:36:32 -0400 Subject: genhdlist? (was: Re: updating a distro tree) In-Reply-To: References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> Message-ID: <20050603223632.GF16989@thomson.net> > KS> could this script be used to incorporate third-party apps into a > KS> distribution? > > Assuming those apps are in the form of RPMS, you just need to put the > packages into the RPMS directory and then run genhdlist and maybe > pkgorder. do_update normally does that if it updated any packages but > you can force it with -G (or just look near the end of the script to > see how to do it yourself). > > But generally I prefer not to mess with adding extra applications to > the install tree and instead put them in a yum repo to be installed in > %post or on the first boot. > I'm being a brave (foolish?) soul and am doing just that ... there's no yum in EL4 (at least, I can't find one on the ISOs), so I'm bundling in APT instead. Could someone out there possibly explain to me a little more about what genhdlist does, and what its various options mean? It's got a usage message, no man page, no info page, and the only docs I've found online relate to RH7.2 and earlier. Can someone shed some light on this mysterious-yet-essential application? thanks, Klaus From phr at doc.ic.ac.uk Fri Jun 3 22:56:00 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Fri, 3 Jun 2005 23:56:00 +0100 (BST) Subject: genhdlist? (was: Re: updating a distro tree) In-Reply-To: <20050603223632.GF16989@thomson.net> References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> Message-ID: On Fri, 3 Jun 2005, Klaus Steden wrote: >I'm being a brave (foolish?) soul and am doing just that ... there's no yum in >EL4 (at least, I can't find one on the ISOs), so I'm bundling in APT instead. You could fetch the Fedora Core RPMs for yum, which would almost certainly just drop in. >Could someone out there possibly explain to me a little more about what >genhdlist does, and what its various options mean? It's got a usage >message, no man page, no info page, and the only docs I've found online >relate to RH7.2 and earlier. genhdlist builds the file which maps between package names and files. For example, comps.xml specifies that the "httpd" package is to be installed if the user selects the "Web Server" group. Anaconda looks up "httpd" to discover that relevant file is "httpd-2.0.52-3.1.i386.rpm" Therefore, say there were an update to "httpd-2.1.23-4.5.i386.rpm" you could replace the original file in RedHat/RPMS then rebuild base/hdlist{,2} by running genhdlist. The options are something like: --withnumbers used for building CD sets --fileorder used for building hdlist in the right order for dependencies --hdlist where to put the output --productpath used for Fedora (vs RedHat) (No research here, just guesswork and experience. Void where prohibited, YMMV.) Cheers, Phil From tibbs at math.uh.edu Fri Jun 3 23:15:15 2005 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Fri, 03 Jun 2005 18:15:15 -0500 Subject: genhdlist? In-Reply-To: (Philip Rowlands's message of "Fri, 3 Jun 2005 23:56:00 +0100 (BST)") References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> Message-ID: >>>>> "PR" == Philip Rowlands writes: PR> You could fetch the Fedora Core RPMs for yum, which would almost PR> certainly just drop in. The latest ones have sqlite dependencies which might be problematic. I recommend grabbing the 2.2.x release from http://linux.duke.edu/projects/yum/download.ptml and building a custom package with your custom repos already set up in /etc/yum.repos.d. The spec makes it pretty easy to add those. PR> genhdlist builds the file which maps between package names and PR> files. It is important to note that the hdlist is used only by the installer. It is completely immaterial to yum which has its own repository metadata format (currently generated by createrepo). I have a feeling that the hdlist will go away if the installer gains the ability to install from yum repositories in FC5. On x86 systems you can generally get away with just running genhdlist when you update or add a package. On x86_64 you run a significant chance of breaking your installs if you don't also run pkgorder. (If the i686 glibc installs after the x86_64 one, stuff can break.) I run it like this: /usr/lib/anaconda-runtime/genhdlist --productpath Fedora XXX PYTHONPATH=/usr/lib/anaconda /usr/lib/anaconda-runtime/pkgorder XXX i386 Fedora > pkgorder /usr/lib/anaconda-runtime/genhdlist --productpath Fedora --fileorder pkgorder XXX replacing XXX by the full patch to the top level of your install tree (the one that contains the Fedora directory). You can probably delete "--productpath" and "Fedora" on RHEL. - J< From gobbledegeek at gmail.com Sat Jun 4 06:53:12 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Sat, 4 Jun 2005 12:23:12 +0530 Subject: Automatic kickstart install In-Reply-To: <20050601195848.DAA4873A28@hormel.redhat.com> References: <20050601195848.DAA4873A28@hormel.redhat.com> Message-ID: <463aea570506032353498a7684@mail.gmail.com> Thanks both of you. Er.. where do I put isolinux.cfg on the cd? What is the format? Will I find the docs in /usr/share/kickstart* Do I need to install kickstart-tools to get them? THA :) -------------------------------------------- From: Brad Doctor Here is an example from isolinux.cfg: default linux prompt 1 display general.msg label install kernel vmlinuz append ks=cdrom:/myKS.cfg initrd=initrd.img lang= devfs=nomount ramdisk_size=8192 > Tell the bootloader (ISOLINUX) to include > ks=cdrom:/path/to/kickstart_file in the kernel command line. > > (Example would go here if today weren't so busy.) > > > Cheers, > Phil Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From mshuler at rackspace.com Sat Jun 4 13:23:15 2005 From: mshuler at rackspace.com (Michael Shuler) Date: Sat, 04 Jun 2005 08:23:15 -0500 Subject: genhdlist? In-Reply-To: References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> Message-ID: <42A1AB43.7040602@rackspace.com> Jason L Tibbitts III wrote: I just wanted to add one little tidbit here that tripped me up for a few minutes. Building an x86_64 install tree on an i686 kickstart server fails, unless: # to build x86_64 tree, we need to trick genhdlist echo x86_64-redhat-linux > /etc/rpm/platform > /usr/lib/anaconda-runtime/genhdlist --productpath Fedora XXX > PYTHONPATH=/usr/lib/anaconda /usr/lib/anaconda-runtime/pkgorder XXX i386 Fedora > pkgorder > /usr/lib/anaconda-runtime/genhdlist --productpath Fedora --fileorder pkgorder XXX # set back the correct /etc/rpm/platform echo i686-redhat-linux > /etc/rpm/platform Kind Regards, Michael Shuler From David.Knight at clubcorp.com Mon Jun 6 06:01:55 2005 From: David.Knight at clubcorp.com (David.Knight at clubcorp.com) Date: Mon, 6 Jun 2005 01:01:55 -0500 Subject: David Knight/CLUBCORP/US is out of the office. Message-ID: I will be out of the office starting 06/03/2005 and will not return until 06/13/2005. I am out of the office in training starting 5/13/05 and will be returning 05/13/05. I will be checking e-mail periodically and will try to respond as soon as time permits. From fvang at zantaz.com Mon Jun 6 21:10:25 2005 From: fvang at zantaz.com (Fong Vang) Date: Mon, 6 Jun 2005 14:10:25 -0700 Subject: kickstart exited abnormally -- sig-11 Message-ID: <2A0946C0CB543B4EA5E41C21D97D256E02E40A58@zorg.corp.zantaz.com> I tried kicking three different machines but they all exited with signal 11. In the past, I'm able work around this by disabling USB or ACPI. This problem exists for both RedHat 9 as well as RedHat Enterprise WS 3.0. System is able to PXE Boot.. loads kernel and RamDisk without any problem. Then right after system obtains an IP address and Host information, it errors out with a "install exited abnormally - received signal 11" message and terminates. There's nothing on the screen (all virtual terminals) to indicate a problem This is the last line displayed in vts-3: * ks location: nfs://192.168.1.1:/install/Kickstart/192.168.1.238-kickstart Any help is appreciated. This is the KS file for RH9: install nfs --server 192.168.1.1 --dir /install/redhat-9.0 lang en_US langsupport --default en_US.UTF-8 en_US.UTF-8 keyboard us mouse none --device psaux skipx xconfig --noprobe network --device eth0 --bootproto dhcp firewall --disabled authconfig --enableshadow --enablemd5 --enablecache timezone --utc GMT bootloader --location=mbr #clearpart --linux --drives=hda,sda --initlabel clearpart --drives=hda --initlabel part /boot --fstype ext3 --size=100 --ondisk=hda part swap --size=1024 --ondisk=hda part / --fstype ext3 --size=1024 --ondisk=hda part /var --size=1024 --fstype=ext3 --ondisk=hda part /opt --size=1024 --fstype=reiserfs --ondisk=hda part /var/log --size=1024 --grow --fstype=reiserfs --ondisk=hda reboot %packages @ Administration Tools @ Development Tools @ Dialup Networking Support @ Editors @ Network Servers @ System Tools @ Text-based Internet #@ X Window System ... ... -------------- next part -------------- An HTML attachment was scrubbed... URL: From jkeating at j2solutions.net Tue Jun 7 03:23:01 2005 From: jkeating at j2solutions.net (Jesse Keating) Date: Mon, 06 Jun 2005 20:23:01 -0700 Subject: kickstart exited abnormally -- sig-11 In-Reply-To: <2A0946C0CB543B4EA5E41C21D97D256E02E40A58@zorg.corp.zantaz.com> References: <2A0946C0CB543B4EA5E41C21D97D256E02E40A58@zorg.corp.zantaz.com> Message-ID: <1118114581.2797.19.camel@yoda.loki.me> On Mon, 2005-06-06 at 14:10 -0700, Fong Vang wrote: > I tried kicking three different machines but they all exited with > signal 11. In the past, I?m able work around this by disabling USB or > ACPI. This problem exists for both RedHat 9 as well as RedHat > Enterprise WS 3.0. > > > > System is able to PXE Boot.. loads kernel and RamDisk without any > problem. Then right after system obtains an IP address and Host > information, it errors out with a ?install exited abnormally ? > received signal 11? message and terminates. There?s nothing on the > screen (all virtual terminals) to indicate a problem > > This is the last line displayed in vts-3: > > > > * ks location: > nfs://192.168.1.1:/install/Kickstart/192.168.1.238-kickstart Hi Fong. I've ran into this before (I think you know where). Just as a sanity test, can you try accessing the ks config file over http rather than NFS? ks=http://192.168.1.1/install/Kickstart/192.168.1.238-kickstart I know this doesn't fit well into your environment, however I had to change my environment over when I consistently ran into this. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) Was I helpful? Let others know: http://svcs.affero.net/rm.php?r=jkeating From klaus.steden at thomson.net Tue Jun 7 03:54:45 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Mon, 6 Jun 2005 23:54:45 -0400 Subject: comps.xml, ordering in RPMs, and nVidia drivers In-Reply-To: <42A1AB43.7040602@rackspace.com> References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> <42A1AB43.7040602@rackspace.com> Message-ID: <20050607035445.GW16989@thomson.net> Hello again, I am putting together an automated installer that will include installation of nVidia drivers including the nVidia kernel modules. I think I've managed to get the necessary syntax for comps.xml, and the nvidia-graphics7174-kmdl RPM merges into the rest of my install tree ... but I am still puzzled about a few things: 1. What does it mean when genhdlist complains like this: WARNING: ordering not found for nvidia-graphics7174-kmdl-2.6.9-5.0.5.ELsmp-1.0_7174-68.el4.at.x86_64.rpm 2. The kernel and the nVidia kernel module I am trying to use are the same version (2.6.9-5.0.5.ELsmp); however, if I specify this in my comps.xml: -- cut here -- workstation-only Workstation-specific Packages false Drivers and Packages Strictly for Workstations. false mpg123 nvidia-graphics-devices nvidia-graphics7174-kmdl-2.6.9-5.0.5.ELsmp nvidia-graphics7174-libs tmake mpg123 xmms xmms-mp3 xxdiff -- cut here -- anaconda complains and dies. If I remove all the version info from the XML tag, anaconda doesn't install the module (makes sense). So how do I specify the right version of the kernel module in the comps.xml file so that it gets installed automagically? Is this the normal way to do things, or should I just install it manually in the %post section of kickstart? thanks, Klaus From phr at doc.ic.ac.uk Tue Jun 7 10:02:54 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 7 Jun 2005 11:02:54 +0100 (BST) Subject: comps.xml, ordering in RPMs, and nVidia drivers In-Reply-To: <20050607035445.GW16989@thomson.net> References: <429DFEC5.2060007@oracle.com> <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> <42A1AB43.7040602@rackspace.com> <20050607035445.GW16989@thomson.net> Message-ID: On Mon, 6 Jun 2005, Klaus Steden wrote: >So how do I specify the right version of the kernel module in the >comps.xml file so that it gets installed automagically? Is this the >normal way to do things, or should I just install it manually in the >%post section of kickstart? Sounds like you might be missing out the genhdlist and/or pkgorder steps required when making certain changes to comps.xml (see list archives for details). %post would probably be the quickest/easiest solution in this case. Cheers, Phil From fvang at zantaz.com Tue Jun 7 18:43:30 2005 From: fvang at zantaz.com (Fong Vang) Date: Tue, 7 Jun 2005 11:43:30 -0700 Subject: kickstart exited abnormally -- sig-11 Message-ID: <2A0946C0CB543B4EA5E41C21D97D256E02E40A62@zorg.corp.zantaz.com> I've already figured it out. Thanks for responding, though. It turned out to be an DNS+NFS issue on the Kickstart server. Too bad it's so silent. This should be easy to do on the client -- tell you it has problem with NFS. Anyway, I'll switch us to HTTP in the next release. -----Original Message----- From: Jesse Keating [mailto:jkeating at j2solutions.net] Sent: Monday, June 06, 2005 8:23 PM To: kickstart-list at redhat.com Subject: Re: kickstart exited abnormally -- sig-11 On Mon, 2005-06-06 at 14:10 -0700, Fong Vang wrote: > I tried kicking three different machines but they all exited with > signal 11. In the past, I'm able work around this by disabling USB or > ACPI. This problem exists for both RedHat 9 as well as RedHat > Enterprise WS 3.0. > > > > System is able to PXE Boot.. loads kernel and RamDisk without any > problem. Then right after system obtains an IP address and Host > information, it errors out with a "install exited abnormally - > received signal 11" message and terminates. There's nothing on the > screen (all virtual terminals) to indicate a problem > > This is the last line displayed in vts-3: > > > > * ks location: > nfs://192.168.1.1:/install/Kickstart/192.168.1.238-kickstart Hi Fong. I've ran into this before (I think you know where). Just as a sanity test, can you try accessing the ks config file over http rather than NFS? ks=http://192.168.1.1/install/Kickstart/192.168.1.238-kickstart I know this doesn't fit well into your environment, however I had to change my environment over when I consistently ran into this. -- Jesse Keating RHCE (geek.j2solutions.net) Fedora Legacy Team (www.fedoralegacy.org) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) Was I helpful? Let others know: http://svcs.affero.net/rm.php?r=jkeating From klaus.steden at thomson.net Tue Jun 7 21:30:02 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 7 Jun 2005 17:30:02 -0400 Subject: comps.xml, ordering in RPMs, and nVidia drivers In-Reply-To: References: <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> <42A1AB43.7040602@rackspace.com> <20050607035445.GW16989@thomson.net> Message-ID: <20050607213002.GY16989@thomson.net> > >So how do I specify the right version of the kernel module in the > >comps.xml file so that it gets installed automagically? Is this the > >normal way to do things, or should I just install it manually in the > >%post section of kickstart? > > Sounds like you might be missing out the genhdlist and/or pkgorder steps > required when making certain changes to comps.xml (see list archives for > details). > Hmmm, it seems I've got more than one package for a culprit, though. What causes this? It kills anaconda dead - I'd like to understand how it happens so I can prevent it wherever possible. > > %post would probably be the quickest/easiest solution in this case. > Yeah, that works if I've only got one or two packages that work that way, but I've got more than that. I would prefer to use the comps.xml framework as it is more portable and more flexible in the long term. cheers, Klaus From klaus.steden at thomson.net Tue Jun 7 22:29:23 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 7 Jun 2005 18:29:23 -0400 Subject: comps.xml, ordering in RPMs, and nVidia drivers In-Reply-To: <20050607213002.GY16989@thomson.net> References: <20050601210023.GN16989@thomson.net> <20050603223632.GF16989@thomson.net> <42A1AB43.7040602@rackspace.com> <20050607035445.GW16989@thomson.net> <20050607213002.GY16989@thomson.net> Message-ID: <20050607222923.GA16989@thomson.net> > Hmmm, it seems I've got more than one package for a culprit, though. What > causes this? It kills anaconda dead - I'd like to understand how it happens so > I can prevent it wherever possible. > It's a bad sign when I reply to myself but from the looking around I did, this bug/error message crops up a lot but there isn't a lot available to explain why. For what it's worth, I found the instructions/info here to be helpful in solving my problem as well: https://www.redhat.com/archives/kickstart-list/2005-May/msg00052.html (I was working with an RPM called 'XnView-static.i386.rpm'). cheers, Klaus From info at hostinthebox.net Wed Jun 8 01:38:54 2005 From: info at hostinthebox.net (dan) Date: Tue, 07 Jun 2005 18:38:54 -0700 Subject: Guide to comps.xml and friends Message-ID: <42A64C2E.2040308@hostinthebox.net> 'Evening, all - I've been working on creating my own comps.xml file for use with a kickstart install, to make a fully custom mini distribution based on CentOS and/or RHEL. Things are going well, and I'd be happy to publish the file for someone else to use, if they would like. However, I was wondering if anyone has found, or has in their posession, a comprehensive guide to comps.xml, detailing which packages have which dependencies, which groups are inherited, which groups are required, the true definition of each directive (, for instance)... stuff like that. > I'm sure that I can whip this out in maybe a few days, but I'd rather > not re-invent the wheel, ya know? > > I think that's about it for now. You guys have been very helpful over > the past couple months, and I greatly appreciate that and thank you all. > Dan, I found this page helpful (although not entirely comprehensive for my needs). http://penguin.triumf.ca/comps/ hth, Klaus From info at hostinthebox.net Wed Jun 8 02:21:52 2005 From: info at hostinthebox.net (dan) Date: Tue, 07 Jun 2005 19:21:52 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <20050608013926.GB16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <20050608013926.GB16989@thomson.net> Message-ID: <42A65640.2060706@hostinthebox.net> Klaus Steden wrote: >>'Evening, all - >> >>I've been working on creating my own comps.xml file for use with a >>kickstart install, to make a fully custom mini distribution based on >>CentOS and/or RHEL. >> >>Things are going well, and I'd be happy to publish the file for someone >>else to use, if they would like. >> >>However, I was wondering if anyone has found, or has in their posession, >>a comprehensive guide to comps.xml, detailing which packages have which >>dependencies, which groups are inherited, which groups are required, the >>true definition of each directive (>type="[default][mandatory][optional]>, for instance)... stuff like that. >> I'm sure that I can whip this out in maybe a few days, but I'd rather >>not re-invent the wheel, ya know? >> >>I think that's about it for now. You guys have been very helpful over >>the past couple months, and I greatly appreciate that and thank you all. >> > > Dan, > > I found this page helpful (although not entirely comprehensive for my needs). > > http://penguin.triumf.ca/comps/ > > hth, > Klaus > Klaus - THat's almost exactly what I'm looking for, but like you said, not as comprehensive as we had hoped ;) However, it is a good start. Thanks! -dant From katzj at redhat.com Wed Jun 8 03:47:40 2005 From: katzj at redhat.com (Jeremy Katz) Date: Tue, 07 Jun 2005 23:47:40 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <42A64C2E.2040308@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> Message-ID: <1118202461.3776.7.camel@bree.local.net> On Tue, 2005-06-07 at 18:38 -0700, dan wrote: > However, I was wondering if anyone has found, or has in their posession, > a comprehensive guide to comps.xml, detailing which packages have which > dependencies, which groups are inherited, which groups are required, the > true definition of each directive ( type="[default][mandatory][optional]>, for instance)... stuff like that. The format is defined at http://rhlinux.redhat.com/anaconda/comps.html Jeremy From brilong at cisco.com Wed Jun 8 14:14:53 2005 From: brilong at cisco.com (Brian Long) Date: Wed, 08 Jun 2005 10:14:53 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <1118202461.3776.7.camel@bree.local.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> Message-ID: <1118240093.16154.5.camel@brilong-lnx2.cisco.com> On Tue, 2005-06-07 at 23:47, Jeremy Katz wrote: > On Tue, 2005-06-07 at 18:38 -0700, dan wrote: > > However, I was wondering if anyone has found, or has in their posession, > > a comprehensive guide to comps.xml, detailing which packages have which > > dependencies, which groups are inherited, which groups are required, the > > true definition of each directive ( > type="[default][mandatory][optional]>, for instance)... stuff like that. > > The format is defined at http://rhlinux.redhat.com/anaconda/comps.html > > Jeremy Jeremy, this page was never updated for RHEL 3 where you removed the sections since Anaconda auto-computes dependencies. Also, the whole section on getfullcomps.py doesn't apply to supported releases. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From klaus.steden at thomson.net Wed Jun 8 17:48:41 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 13:48:41 -0400 Subject: manipulating init states in anaconda In-Reply-To: <1118240093.16154.5.camel@brilong-lnx2.cisco.com> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> Message-ID: <20050608174841.GH16989@thomson.net> Is it possible in anaconda to specify the desired init state of the machine during installation? Or is that automagically determined by the state of anaconda itself, and thus something I have to tinker with manually? thanks, Klaus From info at hostinthebox.net Wed Jun 8 18:26:21 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 11:26:21 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <1118240093.16154.5.camel@brilong-lnx2.cisco.com> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> Message-ID: <42A7384D.2060107@hostinthebox.net> Brian Long wrote: > On Tue, 2005-06-07 at 23:47, Jeremy Katz wrote: > >>On Tue, 2005-06-07 at 18:38 -0700, dan wrote: >> >>>However, I was wondering if anyone has found, or has in their posession, >>>a comprehensive guide to comps.xml, detailing which packages have which >>>dependencies, which groups are inherited, which groups are required, the >>>true definition of each directive (>>type="[default][mandatory][optional]>, for instance)... stuff like that. >> >>The format is defined at http://rhlinux.redhat.com/anaconda/comps.html >> >>Jeremy > > > Jeremy, this page was never updated for RHEL 3 where you removed the > sections since Anaconda auto-computes dependencies. > > Also, the whole section on getfullcomps.py doesn't apply to supported > releases. > > /Brian/ > Thanks for the heads-up Brian. Does this mean that the list that Jeremy had posted does not apply for RHEL3+ systems? Thanks! -dant From steve.mah at oracle.com Wed Jun 8 19:59:42 2005 From: steve.mah at oracle.com (Stephen Mah) Date: Wed, 08 Jun 2005 12:59:42 -0700 Subject: manipulating init states in anaconda In-Reply-To: <20050608174841.GH16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <20050608174841.GH16989@thomson.net> Message-ID: <42A74E2E.30303@oracle.com> would passing the append text parameter work for you? Anaconda would install in a text mode. Klaus Steden wrote: >Is it possible in anaconda to specify the desired init state of the machine >during installation? Or is that automagically determined by the state of >anaconda itself, and thus something I have to tinker with manually? > >thanks, >Klaus > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > From brilong at cisco.com Wed Jun 8 20:07:54 2005 From: brilong at cisco.com (Brian Long) Date: Wed, 08 Jun 2005 16:07:54 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <42A7384D.2060107@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> Message-ID: <1118261273.16154.19.camel@brilong-lnx2.cisco.com> On Wed, 2005-06-08 at 14:26, dan wrote: > Brian Long wrote: > > On Tue, 2005-06-07 at 23:47, Jeremy Katz wrote: > > > >>On Tue, 2005-06-07 at 18:38 -0700, dan wrote: > >> > >>>However, I was wondering if anyone has found, or has in their posession, > >>>a comprehensive guide to comps.xml, detailing which packages have which > >>>dependencies, which groups are inherited, which groups are required, the > >>>true definition of each directive ( >>>type="[default][mandatory][optional]>, for instance)... stuff like that. > >> > >>The format is defined at http://rhlinux.redhat.com/anaconda/comps.html > >> > >>Jeremy > > > > > > Jeremy, this page was never updated for RHEL 3 where you removed the > > sections since Anaconda auto-computes dependencies. > > > > Also, the whole section on getfullcomps.py doesn't apply to supported > > releases. > > > > /Brian/ > > > > > Thanks for the heads-up Brian. Does this mean that the list that Jeremy > had posted does not apply for RHEL3+ systems? It means you no longer need to have the packages section or run getfullcomps.py. You just have groups and grouphierarchy. Look at the comps.xml that comes on the ISO and you'll see what I mean. /Brian/ From klaus.steden at thomson.net Wed Jun 8 20:10:28 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 16:10:28 -0400 Subject: manipulating init states in anaconda In-Reply-To: <42A74E2E.30303@oracle.com> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <20050608174841.GH16989@thomson.net> <42A74E2E.30303@oracle.com> Message-ID: <20050608201028.GQ16989@thomson.net> > would passing the append text parameter work for you? > Anaconda would install in a text mode. > Well, it seems (or so I seem to have discovered but I'm willing to be corrected) that if you specify 'skipx', the default init state is 3, while specifying X11 parameters leaves you with a default init state of 5. This is reasonable enough for my needs. Klaus From klaus.steden at thomson.net Wed Jun 8 20:27:06 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 16:27:06 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <1118261273.16154.19.camel@brilong-lnx2.cisco.com> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> Message-ID: <20050608202706.GS16989@thomson.net> > > > > > Thanks for the heads-up Brian. Does this mean that the list that Jeremy > > had posted does not apply for RHEL3+ systems? > > It means you no longer need to have the packages section or run > getfullcomps.py. You just have groups and grouphierarchy. Look at the > comps.xml that comes on the ISO and you'll see what I mean. > What's the official procedure for rolling your own comps.xml in the meantime? I took the one from the ISO, inserted my own stuff, and copied it into my install tree ... but I'd like to do it the right way, rather than my way. Klaus From phr at doc.ic.ac.uk Wed Jun 8 20:34:24 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 8 Jun 2005 21:34:24 +0100 (BST) Subject: manipulating init states in anaconda In-Reply-To: <20050608201028.GQ16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <20050608174841.GH16989@thomson.net> <42A74E2E.30303@oracle.com> <20050608201028.GQ16989@thomson.net> Message-ID: On Wed, 8 Jun 2005, Klaus Steden wrote: >Well, it seems (or so I seem to have discovered but I'm willing to be >corrected) that if you specify 'skipx', the default init state is 3, >while specifying X11 parameters leaves you with a default init state of >5. This is reasonable enough for my needs. "runlevel" is the operative word here. As you've found, skipx will choose between 3/5, but for more fine-grained control (perhaps you want X to be configured but not start by default) this will work: %post perl -i -p -e 's/^id:.:initdefault:$/id:2:initdefault:/' /etc/inittab to set runlevel 2, for example. Cheers, Phil From info at hostinthebox.net Wed Jun 8 20:35:49 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 13:35:49 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <20050608202706.GS16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> Message-ID: <42A756A5.5050303@hostinthebox.net> Klaus Steden wrote: > > > >>>Thanks for the heads-up Brian. Does this mean that the list that Jeremy >>>had posted does not apply for RHEL3+ systems? >> >>It means you no longer need to have the packages section or run >>getfullcomps.py. You just have groups and grouphierarchy. Look at the >>comps.xml that comes on the ISO and you'll see what I mean. >> > > What's the official procedure for rolling your own comps.xml in the meantime? > > I took the one from the ISO, inserted my own stuff, and copied it into my > install tree ... but I'd like to do it the right way, rather than my way. > > Klaus > Yea, mine blew up, too. I'm still just in the "fooling around" stage. However, I would like to master this. Right now, I'm trying to prune Base and Core a bit, to see if I can get an even smaller install. I'll let you all know how it goes. It would be amazing to finally get this done, it would save me a lto of headache and time. Thanks! -dant From klaus.steden at thomson.net Wed Jun 8 20:35:15 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 16:35:15 -0400 Subject: manipulating init states in anaconda In-Reply-To: References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <20050608174841.GH16989@thomson.net> <42A74E2E.30303@oracle.com> <20050608201028.GQ16989@thomson.net> Message-ID: <20050608203515.GT16989@thomson.net> > >Well, it seems (or so I seem to have discovered but I'm willing to be > >corrected) that if you specify 'skipx', the default init state is 3, > >while specifying X11 parameters leaves you with a default init state of > >5. This is reasonable enough for my needs. > > "runlevel" is the operative word here. As you've found, skipx will > choose between 3/5, but for more fine-grained control (perhaps you want > X to be configured but not start by default) this will work: > > %post > perl -i -p -e 's/^id:.:initdefault:$/id:2:initdefault:/' /etc/inittab > > to set runlevel 2, for example. > Heh, yeah, that works ... but I'm hoping to avoid having to resort to that kind of approach. ;> cheers, Klaus From klaus.steden at thomson.net Wed Jun 8 20:53:58 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 16:53:58 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <42A756A5.5050303@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> Message-ID: <20050608205358.GU16989@thomson.net> > > Yea, mine blew up, too. I'm still just in the "fooling around" stage. > However, I would like to master this. > > Right now, I'm trying to prune Base and Core a bit, to see if I can get > an even smaller install. I'll let you all know how it goes. > > It would be amazing to finally get this done, it would save me a lto of > headache and time. > I just wanted to insert my own groups, which for the most part I think I've got working. Tweaking Base and Core, as much as it appeals to my administrative instincts, is more trouble to me than it's worth ... although for the life of me some of what is considered 'basic' stretches the definition somewhat. Klaus From info at hostinthebox.net Wed Jun 8 21:19:41 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 14:19:41 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <20050608205358.GU16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> <20050608205358.GU16989@thomson.net> Message-ID: <42A760ED.60506@hostinthebox.net> Klaus Steden wrote: >>Yea, mine blew up, too. I'm still just in the "fooling around" stage. >>However, I would like to master this. >> >>Right now, I'm trying to prune Base and Core a bit, to see if I can get >>an even smaller install. I'll let you all know how it goes. >> >>It would be amazing to finally get this done, it would save me a lto of >>headache and time. >> > > I just wanted to insert my own groups, which for the most part I think I've > got working. Tweaking Base and Core, as much as it appeals to my > administrative instincts, is more trouble to me than it's worth ... although > for the life of me some of what is considered 'basic' stretches the > definition somewhat. > > Klaus > Klaus - Yeap, that's what I'm looking for. I'm trying to beat RHEL and CentOS into a purposeful beast that suits our associaton's needs, as well as implementing some of our homegrown software. It's my goal to remove every single item that is not completely essential to doing a short list of very specific tasks. THanks -dant From nomis80 at lqt.ca Wed Jun 8 21:19:13 2005 From: nomis80 at lqt.ca (Simon Perreault) Date: Wed, 8 Jun 2005 17:19:13 -0400 Subject: manipulating init states in anaconda In-Reply-To: <20050608203515.GT16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <20050608203515.GT16989@thomson.net> Message-ID: <200506081719.14046.nomis80@lqt.ca> On Wednesday 08 June 2005 16:35, Klaus Steden wrote: > Heh, yeah, that works ... but I'm hoping to avoid having to resort to that > kind of approach. ;> Why? That's what I'd do. From phr at doc.ic.ac.uk Wed Jun 8 21:33:23 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 8 Jun 2005 22:33:23 +0100 (BST) Subject: Guide to comps.xml and friends In-Reply-To: <42A760ED.60506@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> <20050608205358.GU16989@thomson.net> <42A760ED.60506@hostinthebox.net> Message-ID: On Wed, 8 Jun 2005, dan wrote: >Yeap, that's what I'm looking for. I'm trying to beat RHEL and CentOS >into a purposeful beast that suits our associaton's needs, as well as >implementing some of our homegrown software. It's my goal to remove >every single item that is not completely essential to doing a short >list of very specific tasks. It's sometimes easier, unless you really need to make use of all the different packaging groups that comps.xml provides, to add/remove from within the kickstart file, e.g. %packages --resolvedeps @ Base @ Core @ Printing -glibc-devel -pam_smb nc ethereal (Nothing precludes building said kickstart file with a script) Cheers, Phil From info at hostinthebox.net Wed Jun 8 21:38:52 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 14:38:52 -0700 Subject: Guide to comps.xml and friends In-Reply-To: References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> <20050608205358.GU16989@thomson.net> <42A760ED.60506@hostinthebox.net> Message-ID: <42A7656C.6080203@hostinthebox.net> Philip Rowlands wrote: > On Wed, 8 Jun 2005, dan wrote: > > >>Yeap, that's what I'm looking for. I'm trying to beat RHEL and CentOS >>into a purposeful beast that suits our associaton's needs, as well as >>implementing some of our homegrown software. It's my goal to remove >>every single item that is not completely essential to doing a short >>list of very specific tasks. > > > It's sometimes easier, unless you really need to make use of all the > different packaging groups that comps.xml provides, to add/remove from > within the kickstart file, e.g. > > %packages --resolvedeps > @ Base > @ Core > @ Printing > -glibc-devel > -pam_smb > nc > ethereal > > (Nothing precludes building said kickstart file with a script) > > > Cheers, > Phil > Phil - Yup, and that's where I've started. But since @Base and @Core are forced includes by default, I want to just modify their included package lists, so I don't have to specify anything under %packages. Thanks -dant From ebrown at lanl.gov Wed Jun 8 21:49:44 2005 From: ebrown at lanl.gov (Ed Brown) Date: Wed, 08 Jun 2005 15:49:44 -0600 Subject: Guide to comps.xml and friends In-Reply-To: <42A760ED.60506@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> <20050608205358.GU16989@thomson.net> <42A760ED.60506@hostinthebox.net> Message-ID: <1118267384.15142.280.camel@edbrown.lanl.gov> On Wed, 2005-06-08 at 15:19, dan wrote: > Klaus Steden wrote: > >>Right now, I'm trying to prune Base and Core a bit, to see if I can get > >>an even smaller install. I'll let you all know how it goes. ... > > for the life of me some of what is considered 'basic' stretches the > > definition somewhat. ... > It's my goal to remove > every single item that is not completely essential to doing a short list > of very specific tasks. It would be great to see RedHat pare the Base and Core groups, to be able to start a server installation from a 'more' minimal set, rather than have to try to work backwards from the default install (base and core), which surely no one could defend as minimal. But while there is definitely some low-hanging fruit, after about 10-20 rpms, the difficulty of trying to define the 'completely essential' set is directly proportional to the number of people with a vote, and even if you're the only one, it's still hard deciding 'completely essential' vs. 'practical, convenient'. I'd be interested in seeing the lists people come up with. -Ed From info at hostinthebox.net Wed Jun 8 21:56:17 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 14:56:17 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <1118267384.15142.280.camel@edbrown.lanl.gov> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> <20050608205358.GU16989@thomson.net> <42A760ED.60506@hostinthebox.net> <1118267384.15142.280.camel@edbrown.lanl.gov> Message-ID: <42A76981.5030804@hostinthebox.net> Ed Brown wrote: > On Wed, 2005-06-08 at 15:19, dan wrote: > >>Klaus Steden wrote: >> >>>>Right now, I'm trying to prune Base and Core a bit, to see if I can get >>>>an even smaller install. I'll let you all know how it goes. > > ... > >>>for the life of me some of what is considered 'basic' stretches the >>>definition somewhat. > > ... > >>It's my goal to remove >>every single item that is not completely essential to doing a short list >>of very specific tasks. > > > It would be great to see RedHat pare the Base and Core groups, to be > able to start a server installation from a 'more' minimal set, rather > than have to try to work backwards from the default install (base and > core), which surely no one could defend as minimal. But while there is > definitely some low-hanging fruit, after about 10-20 rpms, the > difficulty of trying to define the 'completely essential' set is > directly proportional to the number of people with a vote, and even if > you're the only one, it's still hard deciding 'completely essential' vs. > 'practical, convenient'. > I'd be interested in seeing the lists people come up with. > > -Ed > > Ed - Yes, of course other people have different ideas of what 'completely essential' was. I'm just doing this in an effort to learn a bit about it in the process, and to see what I can get away with. My findings are surely nothing that anyone should abide by, but they are welcome to use them when I am complete. Thanks -dant From tru at pasteur.fr Wed Jun 8 22:11:16 2005 From: tru at pasteur.fr (Tru Huynh) Date: Thu, 9 Jun 2005 00:11:16 +0200 Subject: Guide to comps.xml and friends In-Reply-To: <1118267384.15142.280.camel@edbrown.lanl.gov> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> <42A756A5.5050303@hostinthebox.net> <20050608205358.GU16989@thomson.net> <42A760ED.60506@hostinthebox.net> <1118267384.15142.280.camel@edbrown.lanl.gov> Message-ID: <20050608221116.GA9766@sillage.bis.pasteur.fr> On Wed, Jun 08, 2005 at 03:49:44PM -0600, Ed Brown wrote: ... > I'd be interested in seeing the lists people come up with. you can check http://www.owlriver.com/tips/tiny-centos/ cheers, Tru From katzj at redhat.com Thu Jun 9 01:09:54 2005 From: katzj at redhat.com (Jeremy Katz) Date: Wed, 08 Jun 2005 21:09:54 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <1118240093.16154.5.camel@brilong-lnx2.cisco.com> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> Message-ID: <1118279395.3810.27.camel@bree.local.net> On Wed, 2005-06-08 at 10:14 -0400, Brian Long wrote: > On Tue, 2005-06-07 at 23:47, Jeremy Katz wrote: > > On Tue, 2005-06-07 at 18:38 -0700, dan wrote: > > > However, I was wondering if anyone has found, or has in their posession, > > > a comprehensive guide to comps.xml, detailing which packages have which > > > dependencies, which groups are inherited, which groups are required, the > > > true definition of each directive ( > > type="[default][mandatory][optional]>, for instance)... stuff like that. > > > > The format is defined at http://rhlinux.redhat.com/anaconda/comps.html > > > > Jeremy > > Jeremy, this page was never updated for RHEL 3 where you removed the > sections since Anaconda auto-computes dependencies. I'll take patches :-) Actually, I should move the document over to the wiki[1] so that people can live edit it. todo++ Jeremy [1] http://fedoraproject.org/wiki/Anaconda Although I'm planning to do some reorganization shortly. From info at hostinthebox.net Thu Jun 9 01:15:28 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 18:15:28 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <1118279395.3810.27.camel@bree.local.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <1118279395.3810.27.camel@bree.local.net> Message-ID: <42A79830.3080908@hostinthebox.net> Jeremy Katz wrote: > On Wed, 2005-06-08 at 10:14 -0400, Brian Long wrote: > >>On Tue, 2005-06-07 at 23:47, Jeremy Katz wrote: >> >>>On Tue, 2005-06-07 at 18:38 -0700, dan wrote: >>> >>>>However, I was wondering if anyone has found, or has in their posession, >>>>a comprehensive guide to comps.xml, detailing which packages have which >>>>dependencies, which groups are inherited, which groups are required, the >>>>true definition of each directive (>>>type="[default][mandatory][optional]>, for instance)... stuff like that. >>> >>>The format is defined at http://rhlinux.redhat.com/anaconda/comps.html >>> >>>Jeremy >> >>Jeremy, this page was never updated for RHEL 3 where you removed the >> sections since Anaconda auto-computes dependencies. > > > I'll take patches :-) > > Actually, I should move the document over to the wiki[1] so that people > can live edit it. todo++ > > Jeremy > > [1] http://fedoraproject.org/wiki/Anaconda Although I'm planning to do > some reorganization shortly. > Please pardon my ignorance, but is it safe to assume that the Anaconda installation process and procedure as I understand it, will remain the same in the general throughout Fedora, RHEL, and CentOS? I guess what I'm getting at, is you're talking about editing the fedoraproject.org's Wiki, but is that information going to benefit other guys attempting the same thing in FC, RHEL, and CentOS and a hadful of others? Thanks again! -dant From katzj at redhat.com Thu Jun 9 01:30:07 2005 From: katzj at redhat.com (Jeremy Katz) Date: Wed, 08 Jun 2005 21:30:07 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <42A79830.3080908@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <1118279395.3810.27.camel@bree.local.net> <42A79830.3080908@hostinthebox.net> Message-ID: <1118280608.3810.32.camel@bree.local.net> On Wed, 2005-06-08 at 18:15 -0700, dan wrote: > Please pardon my ignorance, but is it safe to assume that the Anaconda > installation process and procedure as I understand it, will remain the > same in the general throughout Fedora, RHEL, and CentOS? > > I guess what I'm getting at, is you're talking about editing the > fedoraproject.org's Wiki, but is that information going to benefit other > guys attempting the same thing in FC, RHEL, and CentOS and a hadful of > others? The "upstream" for anaconda is Fedora. So all of the (interesting) development for anaconda occurs there and then percolates into a later RHEL release. And since CentOS bases off of RHEL, ... :-) Jeremy From klaus.steden at thomson.net Thu Jun 9 01:27:35 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 21:27:35 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <42A79830.3080908@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <1118279395.3810.27.camel@bree.local.net> <42A79830.3080908@hostinthebox.net> Message-ID: <20050609012735.GV16989@thomson.net> > > Please pardon my ignorance, but is it safe to assume that the Anaconda > installation process and procedure as I understand it, will remain the > same in the general throughout Fedora, RHEL, and CentOS? > On that topic ... pardon my ignorance, but what is the difference between Fedora, RHEL, and CentOS, and how many other RH variants exist (current, not vintage). thanks, Klaus From skvidal at phy.duke.edu Thu Jun 9 01:36:05 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Wed, 08 Jun 2005 21:36:05 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <20050609012735.GV16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <1118279395.3810.27.camel@bree.local.net> <42A79830.3080908@hostinthebox.net> <20050609012735.GV16989@thomson.net> Message-ID: <1118280965.14757.6.camel@cutter> On Wed, 2005-06-08 at 21:27 -0400, Klaus Steden wrote: > > > > Please pardon my ignorance, but is it safe to assume that the Anaconda > > installation process and procedure as I understand it, will remain the > > same in the general throughout Fedora, RHEL, and CentOS? > > > On that topic ... pardon my ignorance, but what is the difference between > Fedora, RHEL, and CentOS, and how many other RH variants exist (current, not > vintage). > probably best checking out the fedora.redhat.com website for fedora vs rhel difference then talk to the centos folks for the centos vs rhel difference. -sv From info at hostinthebox.net Thu Jun 9 01:42:02 2005 From: info at hostinthebox.net (dan) Date: Wed, 08 Jun 2005 18:42:02 -0700 Subject: Guide to comps.xml and friends In-Reply-To: <20050609012735.GV16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <1118279395.3810.27.camel@bree.local.net> <42A79830.3080908@hostinthebox.net> <20050609012735.GV16989@thomson.net> Message-ID: <42A79E6A.3090807@hostinthebox.net> Klaus Steden wrote: >>Please pardon my ignorance, but is it safe to assume that the Anaconda >>installation process and procedure as I understand it, will remain the >>same in the general throughout Fedora, RHEL, and CentOS? >> > > On that topic ... pardon my ignorance, but what is the difference between > Fedora, RHEL, and CentOS, and how many other RH variants exist (current, not > vintage). > > thanks, > Klaus > Klaus - I was just curious if I should prepare for 10 slightly varying installers, or one overall installer. Thanks -dant From klaus.steden at thomson.net Thu Jun 9 02:04:15 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 8 Jun 2005 22:04:15 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <42A79E6A.3090807@hostinthebox.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <1118279395.3810.27.camel@bree.local.net> <42A79830.3080908@hostinthebox.net> <20050609012735.GV16989@thomson.net> <42A79E6A.3090807@hostinthebox.net> Message-ID: <20050609020415.GW16989@thomson.net> > > Klaus - > > I was just curious if I should prepare for 10 slightly varying > installers, or one overall installer. > Uh-oh, you're asking me for advice? ;> No, seriously ... I have only 'one' installer but I have a number of different install types, so what I did was I made some perl-based template files that can account for the variation, and my kickstart URL is actually a CGI that parses these files and returns a generated ks.cfg. This took a little bit of work to implement - basically, I had to intercept anaconda's invocation between /sbin/loader and /usr/bin/anaconda, insert my own code (which fetched the KS file), and then continue running and exec the actual anaconda script. If you have the time/patience for a variety of installers (I don't - I need to have them all available in one CD), you could simply specify your parameters of variation in the boot command line (you'd have to cut your own RH ISO, though). Probably easier, if you don't have heavy customization or a lot of major differences between your install types (I do). If you want more info, contact me off list. cheers, Klaus From brilong at cisco.com Fri Jun 10 02:00:41 2005 From: brilong at cisco.com (Brian Long) Date: Thu, 09 Jun 2005 22:00:41 -0400 Subject: Guide to comps.xml and friends In-Reply-To: <20050608202706.GS16989@thomson.net> References: <42A64C2E.2040308@hostinthebox.net> <1118202461.3776.7.camel@bree.local.net> <1118240093.16154.5.camel@brilong-lnx2.cisco.com> <42A7384D.2060107@hostinthebox.net> <1118261273.16154.19.camel@brilong-lnx2.cisco.com> <20050608202706.GS16989@thomson.net> Message-ID: <42A8F449.5030002@cisco.com> Klaus Steden wrote: > > > > >>>Thanks for the heads-up Brian. Does this mean that the list that Jeremy >>>had posted does not apply for RHEL3+ systems? >>> >>> >>It means you no longer need to have the packages section or run >>getfullcomps.py. You just have groups and grouphierarchy. Look at the >>comps.xml that comes on the ISO and you'll see what I mean. >> >> >> >What's the official procedure for rolling your own comps.xml in the meantime? > > I don't think Red Hat officially supports re-doing the kickstart image. We have developed some open-source tools that help you in this effort: http://kickstart-tools.sf.net. The kickstart-tools RPM includes "build_ks" which takes a config file full of RPMs and makes comps.xml, runs genhdlist, pkgorder, etc. I've also recently added yumrepo and yumgroup support to it, but that version is not on sourceforge yet. You can define a list of RPMs to include in a yumrepo and create yumgroups inside that repo all automated. /Brian/ From info at hostinthebox.net Sat Jun 11 02:23:11 2005 From: info at hostinthebox.net (Dan Trainor) Date: Fri, 10 Jun 2005 19:23:11 -0700 Subject: Automated kickstart of remote machine Message-ID: <42AA4B0F.2000902@hostinthebox.net> Hello, all - I've set up a small lab in VMWare to experiment with upgrading an older machine. I'd like to come up with a process to completely re-install RHEL4 on the machine. I'm trying to pull all this off in a little lab before I go for the big money - a server that is about 1200 miles away from me. I'm able to do a kickstart install off of a CD, or a disk - that's cake. However, I don't know much about doing a kickstart on a remote machine, a machine I will have no physical access to. All I can do (and hopefully after the kickstart will still be able to) is SSH to the machine as it stands. I guess what I'm asking here is what are my options in an effort to kickstart this machine remotely? Is there a process which I can hack at grub a little bit to load some additional kickstart paramaters next time it boots, and at that time, it would begin the kickstart? Is that how it's done in this situation? Thanks again for the time -dant From klaus.steden at thomson.net Sat Jun 11 02:54:48 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 10 Jun 2005 22:54:48 -0400 Subject: Automated kickstart of remote machine In-Reply-To: <42AA4B0F.2000902@hostinthebox.net> References: <42AA4B0F.2000902@hostinthebox.net> Message-ID: <20050611025448.GE16989@thomson.net> > Hello, all - > > I've set up a small lab in VMWare to experiment with upgrading an older > machine. I'd like to come up with a process to completely re-install > RHEL4 on the machine. I'm trying to pull all this off in a little lab > before I go for the big money - a server that is about 1200 miles away > from me. > > I'm able to do a kickstart install off of a CD, or a disk - that's cake. > However, I don't know much about doing a kickstart on a remote machine, > a machine I will have no physical access to. All I can do (and > hopefully after the kickstart will still be able to) is SSH to the > machine as it stands. > > I guess what I'm asking here is what are my options in an effort to > kickstart this machine remotely? Is there a process which I can hack at > grub a little bit to load some additional kickstart paramaters next time > it boots, and at that time, it would begin the kickstart? Is that how > it's done in this situation? > You -may- be able to do something like that if you can get a DHCP server up on the same segment, and then configure it, or preconfigure your install media, to find a kickstart file that way. There's some discussion of it in the anaconda docs, although I've never investigated it. hth, Klaus From error27 at gmail.com Sat Jun 11 05:29:18 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 10 Jun 2005 22:29:18 -0700 Subject: Automated kickstart of remote machine In-Reply-To: <42AA4B0F.2000902@hostinthebox.net> References: <42AA4B0F.2000902@hostinthebox.net> Message-ID: I don't think you've given us enough detail to work with but for the fun of it I'm going to use my imagination to fill in the details. cp ks.cfg / mount /mnt/cdrom/ cp /mnt/cdrom/isolinux/vmlinuz /boot/ cp /mnt/cdrom/isolinux/initrd.img /boot/ Add this to GRUB. title load root (hd0,0) kernel /boot/vmlinuz initrd /boot/initrd.img ks=hd:hda2/ks.cfg lang= text devfs=nomount ramdisk_size=16384 reboot #I assumed root was /dev/hda2 and that it was formated as ext3. Set GRUB to boot from load by default. What is your install media btw? I haven't tested this at all and if you have a typo anywhere or in your kickstart then you're screwed. It is simpler to save the ks.cfg to a floppy and have someone on site to make sure the install goes Ok. regards, dan carpenter From gobbledegeek at gmail.com Sat Jun 11 07:25:05 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Sat, 11 Jun 2005 12:55:05 +0530 Subject: Guide to comps.xml and friends (Brian Long) In-Reply-To: <20050610160024.711A573912@hormel.redhat.com> References: <20050610160024.711A573912@hormel.redhat.com> Message-ID: <463aea5705061100257ec557ff@mail.gmail.com> > >>It means you no longer need to have the packages section or run > >>getfullcomps.py. You just have groups and grouphierarchy. > I just rolled my own comps.xml for a custom single cd fc3. It has my non-standard package groups. Will everything work as usual.. ? > >What's the official procedure for rolling your own comps.xml in the > >meantime? Manually.. the good ol' way. Genuine, pure and handrolled on the thighs of local beauties.. ;} > http://kickstart-tools.sf.net. The kickstart-tools RPM includes > "build_ks" which takes a config file full of RPMs and makes comps.xml, From herrold at owlriver.com Sat Jun 11 15:56:45 2005 From: herrold at owlriver.com (R P Herrold) Date: Sat, 11 Jun 2005 11:56:45 -0400 (EDT) Subject: ks] Automated kickstart of remote machine In-Reply-To: <42AA4B0F.2000902@hostinthebox.net> References: <42AA4B0F.2000902@hostinthebox.net> Message-ID: On Fri, 10 Jun 2005, Dan Trainor wrote: > I guess what I'm asking here is what are my options in an effort to > kickstart this machine remotely? Is there a process which I can hack at > grub a little bit to load some additional kickstart paramaters next time > it boots, and at that time, it would begin the kickstart? Is that how > it's done in this situation? Exact method for this with grub: http://www.owlriver.com/tips/hands-off/ (method C) General discussion of kickstart and anaconda methods: http://www.owlriver.com/tips/pxe-install/ Please ask if you need any clarifications in the outlines. - Russ Herrold From ggerard at ggerard.com Sun Jun 12 00:05:30 2005 From: ggerard at ggerard.com (Gregory Gerard) Date: Sat, 11 Jun 2005 17:05:30 -0700 Subject: Ejecting CD after installation Message-ID: <200506120038.j5C0c5QS024130@mx2.redhat.com> What's the right way to eject a CD after installation? I have my kickstart set to reboot afterwards but it boots into the CD-ROM. My CD-ROM drive tray in the server (Dell 1U) is the portable type where the machine couldn't suck the tray and CD-ROM back in even if the BIOS so desired. While I'm using the CD-ROM, it's just for booting the machine, the install type is http based, not local media. Thanks, greg -------------- next part -------------- An HTML attachment was scrubbed... URL: From error27 at gmail.com Sun Jun 12 00:45:29 2005 From: error27 at gmail.com (Dan Carpenter) Date: Sat, 11 Jun 2005 17:45:29 -0700 Subject: Ejecting CD after installation In-Reply-To: <200506120038.j5C0c5QS024130@mx2.redhat.com> References: <200506120038.j5C0c5QS024130@mx2.redhat.com> Message-ID: Since you're not installing from the cdrom, maybe you could call `eject` in the %post script. regards, dan carpenter From Christian.Rohrmeier at SCHERING.DE Mon Jun 13 09:06:51 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Mon, 13 Jun 2005 11:06:51 +0200 Subject: Ejecting CD after installation In-Reply-To: <200506120038.j5C0c5QS024130@mx2.redhat.com> Message-ID: I find that the behaviour of the CDROM from an "http" (or other network)-based install is depedant on the version of RHEL that is installed. The CD is used to boot only, "prompt: linux ks=http://blahblah". RHEL 2.1 will allow me to eject the CD as soon as it has loaded the images from the URL (i.e. during install, at any point, PRE% or POST% script.) RHEL 3 and 4 do NOT allow the CD to be removed during install until after the machine has been rebooted or halted (i.e. you can't eject the CD during the POST% script either.) -Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 "Gregory Gerard" Sent by: To: kickstart-list-bounces cc: @redhat.com Subject: Ejecting CD after installation 12.06.2005 02:05 Please respond to Discussion list about Kickstart What?s the right way to eject a CD after installation? I have my kickstart set to reboot afterwards but it boots into the CD-ROM. My CD-ROM drive tray in the server (Dell 1U) is the portable type where the machine couldn?t suck the tray and CD-ROM back in even if the BIOS so desired. While I?m using the CD-ROM, it?s just for booting the machine, the install type is http based, not local media. Thanks, greg_______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From James_Martin at ao.uscourts.gov Mon Jun 13 13:36:53 2005 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Mon, 13 Jun 2005 09:36:53 -0400 Subject: Automated kickstart of remote machine In-Reply-To: <42AA4B0F.2000902@hostinthebox.net> Message-ID: Call me crazy, but this sounds like an awfully big hassle to get working.. And if it breaks, it's going to be a pain. Since this is only one machine, why use kickstart at all? Why not use the VNC option that lets you drive the upgrade/installation remotely? James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 06/10/2005 10:23:11 PM: > Hello, all - > > I've set up a small lab in VMWare to experiment with upgrading an older > machine. I'd like to come up with a process to completely re-install > RHEL4 on the machine. I'm trying to pull all this off in a little lab > before I go for the big money - a server that is about 1200 miles away > from me. > > I'm able to do a kickstart install off of a CD, or a disk - that's cake. > However, I don't know much about doing a kickstart on a remote machine, > a machine I will have no physical access to. All I can do (and > hopefully after the kickstart will still be able to) is SSH to the > machine as it stands. > > I guess what I'm asking here is what are my options in an effort to > kickstart this machine remotely? Is there a process which I can hack at > grub a little bit to load some additional kickstart paramaters next time > it boots, and at that time, it would begin the kickstart? Is that how > it's done in this situation? > > Thanks again for the time > -dant > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From chandan-dutta.chowdhury at hp.com Mon Jun 13 13:48:11 2005 From: chandan-dutta.chowdhury at hp.com (Chowdhury, Chandan Dutta) Date: Mon, 13 Jun 2005 19:18:11 +0530 Subject: Automated kickstart of remote machine Message-ID: Hello, It is possible to start a kickstart install by putting something like the following entries in the grub.conf default=0 timeout=5 title Ignite4Linux rhel4-as-i386 Install root (hd0,9) kernel /vmlinuz initrd /initrd.img ks=http://192.168.100.1/kickstart/ks.cfg Here the vmlinuz and initrd.img are images from the 1st-cdrom/images/pxeboot/ You need to put these images in the /boot dir and make the entries in the grub.conf You have to adjust the root(hd0,9) to your environment. Basically it points to the /boot partition if there is one or the / partition. You can make the ks.cfg through any other means. Of cource you should be absolutely sure of the ks.cfg . It is always a good idea to start a vnc install, in case there is a problem in the install. Regards -CDC -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of James_Martin at ao.uscourts.gov Sent: Monday, June 13, 2005 7:07 PM To: Discussion list about Kickstart Subject: Re: Automated kickstart of remote machine Call me crazy, but this sounds like an awfully big hassle to get working.. And if it breaks, it's going to be a pain. Since this is only one machine, why use kickstart at all? Why not use the VNC option that lets you drive the upgrade/installation remotely? James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 kickstart-list-bounces at redhat.com wrote on 06/10/2005 10:23:11 PM: > Hello, all - > > I've set up a small lab in VMWare to experiment with upgrading an > older machine. I'd like to come up with a process to completely > re-install > RHEL4 on the machine. I'm trying to pull all this off in a little lab > before I go for the big money - a server that is about 1200 miles away > from me. > > I'm able to do a kickstart install off of a CD, or a disk - that's cake. > However, I don't know much about doing a kickstart on a remote > machine, a machine I will have no physical access to. All I can do > (and hopefully after the kickstart will still be able to) is SSH to > the machine as it stands. > > I guess what I'm asking here is what are my options in an effort to > kickstart this machine remotely? Is there a process which I can hack > at grub a little bit to load some additional kickstart paramaters next > time it boots, and at that time, it would begin the kickstart? Is > that how it's done in this situation? > > Thanks again for the time > -dant > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From info at hostinthebox.net Mon Jun 13 18:12:55 2005 From: info at hostinthebox.net (Dan Trainor) Date: Mon, 13 Jun 2005 11:12:55 -0700 Subject: Automated kickstart of remote machine In-Reply-To: References: Message-ID: <42ADCCA7.1080506@hostinthebox.net> James_Martin at ao.uscourts.gov wrote: > Call me crazy, but this sounds like an awfully big hassle to get working.. > And if it breaks, it's going to be a pain. Since this is only one > machine, why use kickstart at all? Why not use the VNC option that lets > you drive the upgrade/installation remotely? > > James S. Martin, RHCE > Contractor > Administrative Office of the United States Courts > Washington, DC > (202) 502-2394 > > kickstart-list-bounces at redhat.com wrote on 06/10/2005 10:23:11 PM: > > >>Hello, all - >> >>I've set up a small lab in VMWare to experiment with upgrading an older >>machine. I'd like to come up with a process to completely re-install >>RHEL4 on the machine. I'm trying to pull all this off in a little lab >>before I go for the big money - a server that is about 1200 miles away >>from me. >> >>I'm able to do a kickstart install off of a CD, or a disk - that's cake. >> However, I don't know much about doing a kickstart on a remote machine, >>a machine I will have no physical access to. All I can do (and >>hopefully after the kickstart will still be able to) is SSH to the >>machine as it stands. >> >>I guess what I'm asking here is what are my options in an effort to >>kickstart this machine remotely? Is there a process which I can hack at >>grub a little bit to load some additional kickstart paramaters next time >>it boots, and at that time, it would begin the kickstart? Is that how >>it's done in this situation? >> >>Thanks again for the time >>-dant >> Hello, James - I'm sorry, this process will be implemented with many machines, not just one. How is this a hassle? Tell me, what do you think is more hassle - an attended installation, or a completely automated unattended installation? I'll be practicing this on many machines, and i'd rather perfect the arrt of a completely unattended remote installation, than have to actually sit down and babysit an install until it is done. Then VNC option looks cute, but again, that would require me to actually sit down and use it. I might consider the VNC option for debugging and such, that might be neat. Anyway, thanks for the ideas. -dant From James_Martin at ao.uscourts.gov Tue Jun 14 13:54:05 2005 From: James_Martin at ao.uscourts.gov (James_Martin at ao.uscourts.gov) Date: Tue, 14 Jun 2005 09:54:05 -0400 Subject: Automated kickstart of remote machine In-Reply-To: Message-ID: I can't read people's minds over the internet. Your message directly states you are doing this for one machine, as I've quoted you below. If you said you were doing this for many machines, I would have said something different. That being said, it depends how many machines you are talking about if kickstarting them is really beneficial--especially when you are doing it from 1200 miles away and no linux guru on the other end. > > Hello, all - > > > > I've set up a small lab in VMWare to experiment with upgrading an > > older machine. I'd like to come up with a process to completely > > re-install > > RHEL4 on the machine. I'm trying to pull all this off in a little lab > > > before I go for the big money - a server that is about 1200 miles away > > > from me. James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 From gurbuzzehra at hotmail.com Wed Jun 15 12:49:32 2005 From: gurbuzzehra at hotmail.com (.::afisb::. .....) Date: Wed, 15 Jun 2005 12:49:32 +0000 Subject: ERROR: could not find rrdtool. Use PathAdd: in mrtg.cfg to help mrtg find rrdtoo Message-ID: hi , i install mrtg-2.12.1 on RedHat systemand run it.my mrtg is working properly and creating .png, .log and .html files but when i add PathAdd in c.cfg file it gives this error message: "ERROR: could not find rrdtool. Use PathAdd: in mrtg.cfg to help mrtg find rrdtool" My c.cfg file includes: EnableIPv6: no WorkDir: /var/www/html/cihaz1 Options[_]: bits,growright LogFormat: rrdtool PathAdd: /home/zehra/rrdbuild/rrdtool-1.2.9/bindings/perl-shared //RRDs.pm file is in "/home/zehra/rrdbuild/rrdtool-1.2.9/bindings/perl-shared" directory so I wrote it with PathAdd. I wait your answers please help quickly!!!!!!!!!!!!!!! From avery at port25.com Wed Jun 15 20:20:28 2005 From: avery at port25.com (Avery Fay) Date: Wed, 15 Jun 2005 16:20:28 -0400 Subject: Problems with custom install CD Message-ID: <1118866829.5211.4.camel@localhost.localdomain> Hello, I'm trying to make a custom install CD (RHEL4) that is very stripped down and uses ks. I have two questions: 1.) Is there a way to specify a driverdisk image that is located on the cdrom? I've tried driverdisk --source=cdrom://driverdisk.img and --source=cdrom:/driverdisk.img without luck. If not, how can I integrate a driverdisk into the install media? 2.) Kickstart is loaded but complains that I don't have Red Hat Enterprise Linux install CD in the cdrom drive. What exactly is it checking for? I followed the directions from http://www.harkness.co.uk/other/kickstart.html and my cd should have everything it needs. Do I have to label my CD with particular text? Or disable media check or something? Thanks, -- Avery Fay From klaus.steden at thomson.net Thu Jun 16 04:19:54 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 16 Jun 2005 00:19:54 -0400 Subject: libstdc++.so.5 for i386 on x86_64 In-Reply-To: References: Message-ID: <20050616041954.GV16989@thomson.net> Hello, I'm finding myself in a bit of a sticky situation ... I need to have -both- compat-libstdc++-33 RPMs (i386 and x86_64) installed on my 64 bit machines for some 32 bit apps to work properly. I'm having a hell of a time finding out what depends on /usr/lib/libstdc++.so.5 - and specifying 'compat-libstdc++-33' in my ks.cfg only installs one of them. My thinking is that if I can find the right 32 bit app that's got a dependency on this library, I can install it and anaconda will install the 32 bit libstdc++ RPM as well. The original problem is a vendor package that isn't properly set up, but they're not going to budge on this one, I'm thinking. Anyone got any ideas? thanks, Klaus From espenas at gmail.com Thu Jun 16 10:50:19 2005 From: espenas at gmail.com (Espen Stefansen) Date: Thu, 16 Jun 2005 12:50:19 +0200 Subject: Trouble getting include to work in my scripts Message-ID: <9699b8ce0506160350b1e883d@mail.gmail.com> In my company we use kickstart to install several different computers, for instance laptop, desktop or servers. Each kickstart-file is configured to the specific needs. And in lots of those files, some sections are similar, like yum-settings. So for instance if i need to edit some of the yum-settings, i must edit all of the kickstart-files that uses that yum-setting. Instead i want to put the yum-settings in it's own file and then include it in the kickstart-files. That will make it much easier to update. All the files are in the same directory. So i try to use: %include yum_conf.inc. But that didn't work. So i tried the full path on the server, but that didn't help either. Does anyone have a suggestion to how i can make this work.... Regards Espen Stefansen From chandan-dutta.chowdhury at hp.com Thu Jun 16 10:57:06 2005 From: chandan-dutta.chowdhury at hp.com (Chowdhury, Chandan Dutta) Date: Thu, 16 Jun 2005 16:27:06 +0530 Subject: Trouble getting include to work in my scripts Message-ID: -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Espen Stefansen Sent: Thursday, June 16, 2005 4:20 PM To: kickstart-list at redhat.com Subject: Trouble getting include to work in my scripts In my company we use kickstart to install several different computers, for instance laptop, desktop or servers. Each kickstart-file is configured to the specific needs. And in lots of those files, some sections are similar, like yum-settings. So for instance if i need to edit some of the yum-settings, i must edit all of the kickstart-files that uses that yum-setting. Instead i want to put the yum-settings in it's own file and then include it in the kickstart-files. That will make it much easier to update. All the files are in the same directory. So i try to use: %include yum_conf.inc. But that didn't work. So i tried the full path on the server, but that didn't help either. Does anyone have a suggestion to how i can make this work.... Regards Espen Stefansen ---------------------------------- Hello, If I am not wrong, the %include files should be present on machine where the install is going on. So you will have to get the file to some temporary location (eg /tmp ) on the machine on which the installation is going on and then give the path to the include. You can get the file through a NFS mount or ftp in the %pre section of the ks.cfg -CDC From phr at doc.ic.ac.uk Thu Jun 16 11:02:05 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 16 Jun 2005 12:02:05 +0100 (BST) Subject: libstdc++.so.5 for i386 on x86_64 In-Reply-To: <20050616041954.GV16989@thomson.net> References: <20050616041954.GV16989@thomson.net> Message-ID: On Thu, 16 Jun 2005, Klaus Steden wrote: >I'm having a hell of a time finding out what depends on >/usr/lib/libstdc++.so.5 - and specifying 'compat-libstdc++-33' in my >ks.cfg only installs one of them. My thinking is that if I can find the >right 32 bit app that's got a dependency on this library, I can install >it and anaconda will install the 32 bit libstdc++ RPM as well. If you have the rpmdb-redhat package installed, you can ask: $ rpm --redhatrequires /usr/lib/libstdc++.so.5 This is not mentioned in the rpm(1) manpage - I had to look at the output of "rpm --help" to find it. Cheers, Phil From tibbs at math.uh.edu Thu Jun 16 13:18:18 2005 From: tibbs at math.uh.edu (Jason L Tibbitts III) Date: Thu, 16 Jun 2005 08:18:18 -0500 Subject: libstdc++.so.5 for i386 on x86_64 In-Reply-To: <20050616041954.GV16989@thomson.net> (Klaus Steden's message of "Thu, 16 Jun 2005 00:19:54 -0400") References: <20050616041954.GV16989@thomson.net> Message-ID: >>>>> "KS" == Klaus Steden writes: KS> My thinking is that if I can find the right 32 bit app that's got KS> a dependency on this library, I can install it and anaconda will KS> install the 32 bit libstdc++ RPM as well. You can just create an empty package that has the dependency you want, and install that. Or you can just manually install the package in %post, or set things up so that the machine installs it when it boots for the first time. I generally opt for the latter. I suppose it would be nice to be able to specify the arch of packages in %packages or in comps.xml, but in the absence of that functionality you have to do a few things manually. (I happen to need ncurses.i386.) - J< From Christian.Rohrmeier at SCHERING.DE Thu Jun 16 15:24:19 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 16 Jun 2005 17:24:19 +0200 Subject: Trouble getting include to work in my scripts In-Reply-To: Message-ID: Don't forget RHEL 2.1 anaconda doesn't support %include. (In case you are an unlucky type who still has to admin 2.1... like me...) ;) -Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 "Chowdhury, Chandan Dutta" "Espen Stefansen" Sent by: kickstart-list-bo "Discussion list about Kickstart" unces at redhat.com cc 16.06.2005 12:57 Subject RE: Trouble getting include to work in my scripts Please respond to Discussion list about Kickstart -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Espen Stefansen Sent: Thursday, June 16, 2005 4:20 PM To: kickstart-list at redhat.com Subject: Trouble getting include to work in my scripts In my company we use kickstart to install several different computers, for instance laptop, desktop or servers. Each kickstart-file is configured to the specific needs. And in lots of those files, some sections are similar, like yum-settings. So for instance if i need to edit some of the yum-settings, i must edit all of the kickstart-files that uses that yum-setting. Instead i want to put the yum-settings in it's own file and then include it in the kickstart-files. That will make it much easier to update. All the files are in the same directory. So i try to use: %include yum_conf.inc. But that didn't work. So i tried the full path on the server, but that didn't help either. Does anyone have a suggestion to how i can make this work.... Regards Espen Stefansen ---------------------------------- Hello, If I am not wrong, the %include files should be present on machine where the install is going on. So you will have to get the file to some temporary location (eg /tmp ) on the machine on which the installation is going on and then give the path to the include. You can get the file through a NFS mount or ftp in the %pre section of the ks.cfg -CDC _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From MWeiner at ag.com Fri Jun 17 17:31:02 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Fri, 17 Jun 2005 13:31:02 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> List readers - I am trying to setup a PXE/Kickstart box so that I can do some unattended installs remotely. What I have setup so far is: 1) a yum repository serving up my RPMS 2) the fc2 install base shared via nfs 3) a dhcp server dishing out the private Ips and paths to pxelinux.0 4) a tftp server serving up the pxelinux.0 bootloader and related kernel I have it successfully pulling the pxelinux.0 and according to the message log it appears to read the pxelinux.cfg/default file which has the following contents: [root at yum1 pxelinux.cfg]# cat 0A DEFAULT linux SERIAL 0,38400n8 LABEL linux KERNEL vmlinuz-fc2 APPEND ksdevice=eth0 console=tty0 console=ttyS0,38400 load_ramdisk=1 initrd=initrd-fc2.img network ip=dhcp ks=nfs:10.10.232.54:/tftp/ks-fc2.cfg Which also seems to work, but what isnt working properly is the ks=nfs statement. All services are running on the same box, and everything is served up via tftp and nfs, as in the following: ls -la /tftpboot lrwxrwxrwx 1 root root 10 Jun 17 08:22 10.10.232.54-kickstart -> ks-fc2.cfg -rw-r--r-- 1 root root 2443013 Jun 16 09:56 initrd-fc2.img -rw-r--r-- 1 root root 0 Jun 16 10:18 Kickstart_end drwxr-xr-x 2 root root 4096 Jun 16 19:08 ks -rw-r--r-- 1 root root 1411 Jun 17 08:22 ks-fc2.cfg -rw-r--r-- 1 root root 12952 Jun 16 10:08 pxelinux.0 drwxr-xr-x 2 root root 4096 Jun 17 10:12 pxelinux.cfg -rw-r--r-- 1 root root 1209805 Jun 16 09:56 vmlinuz-fc2 And ls -la /tftpboot/pxelinux.cfg/ lrwxrwxrwx 1 root root 15 Jun 17 09:59 0A -> default.netboot lrwxrwxrwx 1 root root 15 Jun 17 09:59 0A0A -> default.netboot -rw-r--r-- 1 root root 40 Jun 17 08:29 default lrwxrwxrwx 1 root root 19 Jun 17 09:59 default.netboot -> default.netboot-fc2 -rw-r--r-- 1 root root 216 Jun 17 10:12 default.netboot-fc2 I can see the server actually loading the bootloader and I can see this in the messages log now: Jun 16 14:11:48 yum1 rpc.mountd: authenticated mount request from 10.10.232.80:750 for /repo/fedora/linux/core/2/i386/os (/repo) But I am unsure if it is able to find my kickstart config file. The /repo above is of course the install base (Fedora/images/etc under the OS tree). Here is my current ks-fc2.cfg file for completeness: # Kickstart file automatically generated. install nfs --server=10.10.232.54 --dir=/repo/fedora/linux/core/2/i386/os lang en_US.UTF-8 langsupport --default en_US.UTF-8 en_US.UTF-8 keyboard us skipx network --device eth0 --bootproto dhcp --hostname ag800 network --device eth1 --bootproto dhcp --hostname ag800 rootpw --iscrypted $1$.pio2F.qewfj0iy2342^ firewall --disabled selinux --disabled authconfig --enableshadow --enablemd5 timezone America/New_York bootloader --location=mbr --append rhgb quiet # The following is the partition information you requested # Note that any partitions you deleted are not expressed # here so unless you clear all partitions first, this is # not guaranteed to work clearpart --linux part /boot --fstype ext3 --size=101 part /disk2 --fstype ext3 --size=12289 part / --fstype ext3 --size=12289 part swap --size=2048 part /weblog --fstype ext3 --size=1 --grow %packages --resolvedeps @ web-server @ mail-server @ dns-server @ dialup @ network-server @ sql-server @ editors @ admin-tools @ system-tools @ news-server @ smb-server @ authoring-and-publishing @ sound-and-video @ server-cfg @ graphics @ ftp-server @ development-tools @ engineering-and-scientific @ text-internet kernel-smp grub e2fsprogs %post # The Last Thing happening in Kickstart: # get a file via TFTP which indicates we are done. echo "get Kickstart_end" | /usr/bin/tftp 10.10.232.54 Now, when I rebooted this test box this morning, it grabs an IP from the dhcp server, is able to receive and load the pxelinux bootloader, but in the virtual console 3 I see the following after it runs through and loads the kernel: * NFS install method detected will use Rhupdates/ * Running anaconda script /usr/bin/anaconda * Display mode=t * Method = nfs://mnt/source/ . * ddcprobe returned bogus values: ID Name HorizSync VertSync * anaconda floppy devices is fd0 * Running kickstart %pre scripts * All kickstart %pre scripts have been run Which confuses me a little bit as I don't quite understand where its getting the nfs methof pointing to Rhupdates and some of the other stuff. Does anyone have any ideas? Thanks for any help in advance. Michael Weiner From error27 at gmail.com Fri Jun 17 17:46:43 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 17 Jun 2005 10:46:43 -0700 Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> Message-ID: > ks=nfs:10.10.232.54:/tftp/ks-fc2.cfg > ls -la /tftpboot Is tftp a symlink to tftpboot? regards, dan carpenter From MWeiner at ag.com Fri Jun 17 17:49:53 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Fri, 17 Jun 2005 13:49:53 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D305A0@orca.agcom.amgreetings.com> > ks=nfs:10.10.232.54:/tftp/ks-fc2.cfg > ls -la /tftpboot Is tftp a symlink to tftpboot? -- Why yes it is, you think that's whats breaking it? From klaus.steden at thomson.net Fri Jun 17 18:43:07 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 17 Jun 2005 14:43:07 -0400 Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D305A0@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D305A0@orca.agcom.amgreetings.com> Message-ID: <20050617184307.GG97868@thomson.net> > > ks=nfs:10.10.232.54:/tftp/ks-fc2.cfg > > ls -la /tftpboot > > Is tftp a symlink to tftpboot? > -- > Why yes it is, you think that's whats breaking it? > Possible. Depends on the tftp server. Some are more paranoid than others, given how insecure the protocol is. Klaus From MWeiner at ag.com Fri Jun 17 19:59:53 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Fri, 17 Jun 2005 15:59:53 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D3061D@orca.agcom.amgreetings.com> > > ks=nfs:10.10.232.54:/tftp/ks-fc2.cfg > > ls -la /tftpboot > > Is tftp a symlink to tftpboot? > -- > Why yes it is, you think that's whats breaking it? > Possible. Depends on the tftp server. Some are more paranoid than others, given how insecure the protocol is. -- Klaus - Made the change but still no go, any other ideas? Thanks Michael Weiner From error27 at gmail.com Fri Jun 17 23:20:37 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 17 Jun 2005 16:20:37 -0700 Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> Message-ID: I thought it was maybe a typo actually... ;) > But I am unsure if it is able to find my kickstart config file. Yep. It would be cool if anaconda printed a message instead of falling back to manual install. Sometimes what I do is go to tty3 and check to see if my ks.cfg file is present in /tmp/. (It always becomes /tmp/ks.cfg even if it starts out as ks-fc2.cfg). In your case, it is using the kickstart file. That's how it knows to install over NFS. > nfs --server=10.10.232.54 --dir=/repo/fedora/linux/core/2/i386/os I'm not sure I understand what you would like to be different about the install. > Which confuses me a little bit as I don't quite understand where its > getting the nfs methof pointing to Rhupdates I can't answer your question about Rhupdates. My guess is that it's probably a directory name somewhere in the install tree... > and some of the other stuff. I'm going to let someone else on the list address this aspect of your query. ;P regards, dan carpenter From phr at doc.ic.ac.uk Sat Jun 18 19:08:50 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 18 Jun 2005 20:08:50 +0100 (BST) Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D3058F@orca.agcom.amgreetings.com> Message-ID: On Fri, 17 Jun 2005, MW Mike Weiner (5028) wrote: >I am trying to setup a PXE/Kickstart box so that I can do some >unattended installs remotely. [snip] >Which also seems to work, but what isnt working properly is the ks=nfs >statement. My favourite 3 debugging questions: - What did you do? - What happened? - What do you think should happen? You've gone into plenty of detail for #1, but I can't see any report of an actual problem. >I can see the server actually loading the bootloader and I can see this >in the messages log now: > >Jun 16 14:11:48 yum1 rpc.mountd: authenticated mount request from >10.10.232.80:750 for /repo/fedora/linux/core/2/i386/os (/repo) Then it looks like the kickstart file has been found and read successfully. Remember that "ks=nfs" is only instructing anaconda to fetch the kickstart file and second stage over NFS; the kickstart file itself can specify a different source for the RPMs. >Now, when I rebooted this test box this morning, it grabs an IP from >the dhcp server, is able to receive and load the pxelinux bootloader, >but in the virtual console 3 I see the following after it runs through >and loads the kernel: > >* NFS install method detected will use Rhupdates/ That's a way of extending/fixing anaconda with patches. Can be ignored. [snip] >* Running kickstart %pre scripts >* All kickstart %pre scripts have been run > >Which confuses me a little bit as I don't quite understand where its >getting the nfs methof pointing to Rhupdates and some of the other >stuff. OK. All looks fine so far - does it hang at this point? Cheers, Phil From MWeiner at ag.com Mon Jun 20 11:19:07 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Mon, 20 Jun 2005 07:19:07 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D306F9@orca.agcom.amgreetings.com> >I am trying to setup a PXE/Kickstart box so that I can do some >unattended installs remotely. [snip] >Which also seems to work, but what isnt working properly is the ks=nfs >statement. My favourite 3 debugging questions: - What did you do? - What happened? - What do you think should happen? You've gone into plenty of detail for #1, but I can't see any report of an actual problem. >I can see the server actually loading the bootloader and I can see this >in the messages log now: > >Jun 16 14:11:48 yum1 rpc.mountd: authenticated mount request from >10.10.232.80:750 for /repo/fedora/linux/core/2/i386/os (/repo) Then it looks like the kickstart file has been found and read successfully. Remember that "ks=nfs" is only instructing anaconda to fetch the kickstart file and second stage over NFS; the kickstart file itself can specify a different source for the RPMs. >Now, when I rebooted this test box this morning, it grabs an IP from >the dhcp server, is able to receive and load the pxelinux bootloader, >but in the virtual console 3 I see the following after it runs through >and loads the kernel: > >* NFS install method detected will use Rhupdates/ That's a way of extending/fixing anaconda with patches. Can be ignored. [snip] >* Running kickstart %pre scripts >* All kickstart %pre scripts have been run > >Which confuses me a little bit as I don't quite understand where its >getting the nfs methof pointing to Rhupdates and some of the other >stuff. OK. All looks fine so far - does it hang at this point? --- Phil - Thanks for the response, and by the way, I love your debugging methodology - nice and simple and straightforward. Well, you're right, it DOES seem to find the kickstart file (ks-fc2.cfg) though I did not check to ensure it makes its way to /tmp/ks.cfg and it is being instructed to mount the repo directory containing the "Fedora" tree of the installation CDs. What is NOT happening is the actual install isnt taking place, it seems to hang after the pre scripts are run (which there arent any), and mounting VFS - just doesn't seem to want to do anything after that point. I have verified that the installation tree I am pointing at is valid (i.e. /repo/fedora/linux/core/2/i386/os *IS* a copy of whats on the mirrors). I also threw the ISOs into /repo/fedora/linux/core/2/i386/iso as I was interested to see if this might work better - doing the install from ISOs instead of the raw tree - though this isnt critical at this point, but I DO need to get the installer working properly, as I have a few hundred boxes to go through this process :) What else can I look for, or should I look at, to verify the install process at this point? Any way to verify that kickstart is doing the *right thing* ? Appreciate any and all responses and assistance. Michael Weiner From steve.mah at oracle.com Mon Jun 20 17:13:01 2005 From: steve.mah at oracle.com (Stephen Mah) Date: Mon, 20 Jun 2005 10:13:01 -0700 Subject: pxe deployment Message-ID: <42B6F91D.2050902@oracle.com> All, Is it possible for a unix based dhcp server to pass information to a client about tftp/pxe server information during bootup. We don't administer the dhcp servers, and I wanna do some inital research before approaching that group. -regards steve From Steven.Hajducko at DigitalInsight.com Mon Jun 20 17:17:39 2005 From: Steven.Hajducko at DigitalInsight.com (Steven Hajducko) Date: Mon, 20 Jun 2005 10:17:39 -0700 Subject: pxe deployment Message-ID: Certainly. Here's our dhcp.conf - [root at plague:/tmp]# cat /etc/dhcpd.conf # Deploy Kickstart Server DHCPd.conf ddns-update-style ad-hoc; default-lease-time 600; max-lease-time 7200; option subnet-mask 255.255.255.0; option broadcast-address 10.20.11.255; option routers 10.20.11.1; option domain-name-servers 10.20.11.2; #host deploy1 { # hardware ethernet 0:e0:81:21:d8:a4; # fixed-address 10.20.11.100; # filename "/var/www/html/kickstart/deploy.ks"; #} allow bootp; allow booting; # Definition of PXE-specific options # Code 1: Multicast IP address of boot file server # Code 2: UDP port that client should monitor for MTFTP responses # Code 3: UDP port that MTFTP servers are using to listen for MTFTP requests # Code 4: Number of seconds a client must listen for activity before trying # to start a new MTFTP transfer # Code 5: Number of seconds a client must listen before trying to restart # a MTFTP transfer option space PXE; option PXE.mtftp-ip code 1 = ip-address; option PXE.mtftp-cport code 2 = unsigned integer 16; option PXE.mtftp-sport code 3 = unsigned integer 16; option PXE.mtftp-tmout code 4 = unsigned integer 8; option PXE.mtftp-delay code 5 = unsigned integer 8; option PXE.discovery-control code 6 = unsigned integer 8; option PXE.discovery-mcast-addr code 7 = ip-address; class "pxeclients" { match if substring (option vendor-class-identifier, 0, 9) = "PXEClient"; option vendor-class-identifier "PXEClient"; vendor-option-space PXE; # At least one of the vendor-specific PXE options must be set in # order for the client boot ROMs to realize that we are a PXE-compliant # server. We set the MCAST IP address to 0.0.0.0 to tell the boot ROM # that we can't provide multicast TFTP (address 0.0.0.0 means no # address). option PXE.mtftp-ip 0.0.0.0; # This is the name of the file the boot ROMs should download. filename "pxelinux.0"; # This is the name of the server they should get it from. next-server 10.20.11.2; } option mask-supplier false; # Don't respond to ICMP Mask req subnet 10.20.11.0 netmask 255.255.255.0 { range 10.20.11.100 10.20.11.150; } -----Original Message----- From: Stephen Mah [mailto:steve.mah at oracle.com] Sent: Monday, June 20, 2005 10:13 AM To: Discussion list about Kickstart Subject: pxe deployment All, Is it possible for a unix based dhcp server to pass information to a client about tftp/pxe server information during bootup. We don't administer the dhcp servers, and I wanna do some inital research before approaching that group. -regards steve _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From andrew.w.robinson at mms.gov Mon Jun 20 19:37:47 2005 From: andrew.w.robinson at mms.gov (Robinson, Andrew W.) Date: Mon, 20 Jun 2005 13:37:47 -0600 Subject: pxe deployment Message-ID: <82ABF19AC0F67D40A0E2B2F5949BE8E2041393@IMSNEXPRI02.service.agency.mms.pri> I presume you mean you want the unix server to pass the pxe information while some other server is passing out the network information? The process is called dhcp-proxy and the capabilities are provided by the pxe rpm for Red Hat Enterprise Linux. I must confess I haven't mastered the configuration to get the process working robustly. HTH, Andrew Robinson > -----Original Message----- > From: Stephen Mah [mailto:steve.mah at oracle.com] > Sent: Monday, June 20, 2005 12:13 PM > To: Discussion list about Kickstart > Subject: pxe deployment > > All, > > Is it possible for a unix based dhcp server to pass information to a > client about tftp/pxe server information during bootup. We don't > administer the dhcp servers, and I wanna do some inital > research before > approaching that group. > > -regards > steve > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From phr at doc.ic.ac.uk Mon Jun 20 19:55:14 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Mon, 20 Jun 2005 20:55:14 +0100 (BST) Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D306F9@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D306F9@orca.agcom.amgreetings.com> Message-ID: On Mon, 20 Jun 2005, MW Mike Weiner (5028) wrote: >Thanks for the response, and by the way, I love your debugging >methodology - nice and simple and straightforward. Well, you're right, >it DOES seem to find the kickstart file (ks-fc2.cfg) though I did not >check to ensure it makes its way to /tmp/ks.cfg and it is being >instructed to mount the repo directory containing the "Fedora" tree of >the installation CDs. What is NOT happening is the actual install isnt >taking place, it seems to hang after the pre scripts are run (which >there arent any), and mounting VFS - just doesn't seem to want to do >anything after that point. If you don't change VC what's the last thing to print out on VC1 (the blue screen (hmm...))? Does anything happen if left for a full 5 minutes? Sometimes NFS timeouts can be rather slow. If you have access on the NFS server, does # tcpdump -n host show any activity? Cheers, Phil From Teemu.M.Jalonen at nokia.com Tue Jun 21 08:22:21 2005 From: Teemu.M.Jalonen at nokia.com (Teemu.M.Jalonen at nokia.com) Date: Tue, 21 Jun 2005 11:22:21 +0300 Subject: Problems with custom install CD Message-ID: > 2.) Kickstart is loaded but complains that I don't have Red Hat > Enterprise Linux install CD in the cdrom drive. What exactly is it > checking for? .discinfo file should be in the CD root From dns.prasad at ge.com Tue Jun 21 08:33:36 2005 From: dns.prasad at ge.com (Prasad, DNS (Corporate, consultant)) Date: Tue, 21 Jun 2005 14:03:36 +0530 Subject: Problems with custom install CD Message-ID: Hi Jalonen, Can U send the exact error and version UR using. Regards, Prasad.D.N.S -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Teemu.M.Jalonen at nokia.com Sent: Tuesday, June 21, 2005 1:52 PM To: kickstart-list at redhat.com Subject: RE: Problems with custom install CD > 2.) Kickstart is loaded but complains that I don't have Red Hat > Enterprise Linux install CD in the cdrom drive. What exactly is it > checking for? .discinfo file should be in the CD root _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From MWeiner at ag.com Tue Jun 21 13:30:27 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Tue, 21 Jun 2005 09:30:27 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D30B61@orca.agcom.amgreetings.com> On Mon, 20 Jun 2005, MW Mike Weiner (5028) wrote: >Thanks for the response, and by the way, I love your debugging >methodology - nice and simple and straightforward. Well, you're right, >it DOES seem to find the kickstart file (ks-fc2.cfg) though I did not >check to ensure it makes its way to /tmp/ks.cfg and it is being >instructed to mount the repo directory containing the "Fedora" tree of >the installation CDs. What is NOT happening is the actual install isnt >taking place, it seems to hang after the pre scripts are run (which >there arent any), and mounting VFS - just doesn't seem to want to do >anything after that point. If you don't change VC what's the last thing to print out on VC1 (the blue screen (hmm...))? Does anything happen if left for a full 5 minutes? Sometimes NFS timeouts can be rather slow. If you have access on the NFS server, does # tcpdump -n host show any activity? -- Phil - On the main console window, the last thing I see is the following: Md: ... Auto run Done. RAMDISK: Compressed image found at block 0 VFS: Mounted root (ext2 filesystem) And no, leaving it go, even for an hour or more doesn't do anything either. I am going to run tcpdump and see if I can gleam any additional information from it and will report back. Any ideas? Thanks in advance Michael Weiner From avery at port25.com Tue Jun 21 15:42:42 2005 From: avery at port25.com (Avery Fay) Date: Tue, 21 Jun 2005 11:42:42 -0400 Subject: Problems with custom install CD In-Reply-To: References: Message-ID: <1119368562.11655.9.camel@localhost.localdomain> On Tue, 2005-06-21 at 11:22 +0300, Teemu.M.Jalonen at nokia.com wrote: > > 2.) Kickstart is loaded but complains that I don't have Red Hat > > Enterprise Linux install CD in the cdrom drive. What exactly is it > > checking for? > > .discinfo file should be in the CD root Thanks, that fixed it. I didn't realise that cp -a would not copy dot files. -- Avery Fay From espenas at gmail.com Wed Jun 22 11:44:09 2005 From: espenas at gmail.com (Espen Stefansen) Date: Wed, 22 Jun 2005 13:44:09 +0200 Subject: Trouble getting include to work in my scripts In-Reply-To: References: Message-ID: <9699b8ce05062204441ced2da9@mail.gmail.com> On 6/16/05, Christian.Rohrmeier at schering.de wrote: > > Don't forget RHEL 2.1 anaconda doesn't support %include. (In case you are > an unlucky type who still has to admin 2.1... like me...) ;) I'm using Centos 4, RHEL 4 and Fedora 4. That's alot of 4's...:P > -----Original Message----- > From: kickstart-list-bounces at redhat.com > Sent: Thursday, June 16, 2005 4:20 PM > Subject: Trouble getting include to work in my scripts > > In my company we use kickstart to install several different computers, > for instance laptop, desktop or servers. Each kickstart-file is > configured to the specific needs. And in lots of those files, some > sections are similar, like yum-settings. > > So for instance if i need to edit some of the yum-settings, i must edit > all of the kickstart-files that uses that yum-setting. Instead i want to > put the yum-settings in it's own file and then include it in the > kickstart-files. That will make it much easier to update. > > All the files are in the same directory. So i try to use: %include > yum_conf.inc. But that didn't work. So i tried the full path on the > server, but that didn't help either. > > Does anyone have a suggestion to how i can make this work.... > > Regards > Espen Stefansen > > > ---------------------------------- > > Hello, > > If I am not wrong, the %include files should be present on machine > where the install is going on. So you will have to get the file to some > temporary location (eg /tmp ) on the machine on which the installation > is going on and then give the path to the include. You can get the file > through a NFS mount or ftp in the %pre section of the ks.cfg > I copied all of my inc-files to /tmp in the %pre-section. My kickstart now finds them, but it wont run all of them. Some of them get a "Error code 32256, can't run". For instance, in the first inc there's a simple perl one-liner who replaces a line in a file. It fails. The second inc-file does exactly the same, but more lines in other files, but this one doesnt work. The same scripts fail every time. I even tried using using ksappend, but it only includes one file. Does anybody include files in there kickstart-files, and have a working example to show me? And does anybody know what the 32256 error is? From MWeiner at ag.com Wed Jun 22 13:49:09 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Wed, 22 Jun 2005 09:49:09 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D30EEC@orca.agcom.amgreetings.com> OK, major steps forward, found there was an offending line in my kickstart script trying to get the console going on boot, removed it and all is well, the install starts, and things look GREAT! However I have a new issue this morning, I fire up the test box, it goes out, gets an IP, gets the kickstart script, begins the install albeit in an X window (strange as I told kickstart to skipX - though I believe that's for setting up the client X), all goes well until it hits an RPM it seems not to like and throws up the standard error: There was an error installing 'X'.rpm This can indicate a media failure, lack of disk space, and/or hardware problem. This is a fatal error and your install will be aborted. Please verify your media and try your install again. Well, when I reboot the system to try again, it fails on yet another RPM. I rsync'd the RPMS down from loop-mounted ISOs so I am unsure why it is complaining of bad media. Has anyone else had this problem, and how do we work around this issue? I see nothing in any logs on the server providing the install media. Thanks, any help would be greatly appreciated Michael Weiner From klaus.steden at thomson.net Wed Jun 22 14:15:58 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 22 Jun 2005 10:15:58 -0400 Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D30EEC@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D30EEC@orca.agcom.amgreetings.com> Message-ID: <20050622141558.GR2476@thomson.net> > However I have a new issue this morning, I fire up the test box, it goes > out, gets an IP, gets the kickstart script, begins the install albeit in > an X window (strange as I told kickstart to skipX - though I believe > that's for setting up the client X), all goes well until it hits an RPM > it seems not to like and throws up the standard error: > This is possibly a stupid question but ... 'skipx' ... or 'skipX'? Klaus From MWeiner at ag.com Wed Jun 22 14:18:36 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Wed, 22 Jun 2005 10:18:36 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D30F06@orca.agcom.amgreetings.com> > However I have a new issue this morning, I fire up the test box, it > goes out, gets an IP, gets the kickstart script, begins the install > albeit in an X window (strange as I told kickstart to skipX - though I > believe that's for setting up the client X), all goes well until it > hits an RPM it seems not to like and throws up the standard error: > This is possibly a stupid question but ... 'skipx' ... or 'skipX'? -- No such thing as a stupid question, here for clarity is my ks.cfg file I am currently using: [root at yum1 temp]# more /tftp/ks-fc2.cfg # Kickstart file automatically generated by anaconda. install nfs --server=10.10.232.54 --dir=/repo/fedora/linux/core/2/i386/os lang en_US.UTF-8 langsupport --default en_US.UTF-8 en_US.UTF-8 keyboard us skipx network --device eth0 --bootproto dhcp --hostname ag800 # network --device eth1 --bootproto dhcp --hostname ag800 rootpw --iscrypted $1$.pio2FpF$qdZBN7oU0oOG6R3oV6y5X. firewall --disabled selinux --disabled authconfig --enableshadow --enablemd5 timezone America/New_York bootloader --location=mbr --append rhgb quiet # The following is the partition information you requested # Note that any partitions you deleted are not expressed # here so unless you clear all partitions first, this is # not guaranteed to work clearpart --linux part /boot --fstype ext3 --size=101 part /disk2 --fstype ext3 --size=12289 part / --fstype ext3 --size=12289 part swap --size=2048 part /weblog --fstype ext3 --size=1 --grow %packages --resolvedeps @ web-server @ mail-server @ dns-server @ dialup @ network-server @ sql-server @ editors @ admin-tools @ system-tools @ news-server @ smb-server @ authoring-and-publishing @ sound-and-video @ server-cfg @ graphics @ ftp-server @ development-tools @ engineering-and-scientific @ text-internet kernel-smp grub e2fsprogs %post # run rpm error report rpm -Va --nofiles --nomd5 >> /tmp/rpm-problems.txt # add another nameserver echo "10.10.1.13 server.local server" >> /etc/resolv.conf echo "nameserver 10.10.1.24" >> /etc/resolv.conf echo "nameserver 10.10.1.55" >> /etc/resolv.conf # Update thyself yum -y update # The Last Thing happening in Kickstart: # get a file via TFTP which indicates we are done. echo "get Kickstart_end" | /usr/bin/tftp 10.10.232.54 Michael From pmatilai at laiskiainen.org Wed Jun 22 14:22:46 2005 From: pmatilai at laiskiainen.org (Panu Matilainen) Date: Wed, 22 Jun 2005 17:22:46 +0300 Subject: Kickstart Installs of FC2 In-Reply-To: <20050622141558.GR2476@thomson.net> References: <4FD2C985D5E2A642AE25823DFD61C2B006D30EEC@orca.agcom.amgreetings.com> <20050622141558.GR2476@thomson.net> Message-ID: <1119450166.31372.3.camel@localhost.localdomain> On Wed, 2005-06-22 at 10:15 -0400, Klaus Steden wrote: > > However I have a new issue this morning, I fire up the test box, it goes > > out, gets an IP, gets the kickstart script, begins the install albeit in > > an X window (strange as I told kickstart to skipX - though I believe > > that's for setting up the client X), all goes well until it hits an RPM > > it seems not to like and throws up the standard error: > > > This is possibly a stupid question but ... > > 'skipx' ... or 'skipX'? 'skipx' means don't configure X on the installed system. To use text mode during installation use 'text' directive instead (the resulting system can have X configured regardless of install-time gui/text mode). - Panu - From SSeremeth at anacomp.com Wed Jun 22 14:24:43 2005 From: SSeremeth at anacomp.com (Seremeth, Stephen) Date: Wed, 22 Jun 2005 10:24:43 -0400 Subject: Kickstart Installs of FC2 Message-ID: <66911C8E1F2DA24EB29DE289297A135E48573C@usrd104> > This is possibly a stupid question but ... > > 'skipx' ... or 'skipX'? > -- Let me add for anyone using HP's iLO in conjunction with PXE/tftp or the like, add this to your kernel command line for the PXE boot: "nofb". The iLO doesn't get confused. So our ended up looking something like this: [root at server scripts]# more /tftpboot/linux-install/pxelinux.cfg/default default 1 timeout 100 prompt 1 display msgs/boot.msg F1 msgs/boot.msg F2 msgs/general.msg F3 msgs/expert.msg F4 msgs/param.msg F5 msgs/rescue.msg F7 msgs/snake.msg label local localboot 1 label 0 localboot 1 label 1 kernel rhel-4-es/vmlinuz append ks=nfs:mynfsserver.org:/dist/kickstart/rh_es_4/ initrd=rhel-4-es/initrd.img ksdevice=eth0 linksleep=120 eth0_ethtool="speed=100 autoneg=off duplex=full delay_link=1" nofb From phr at doc.ic.ac.uk Wed Jun 22 15:26:40 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 22 Jun 2005 16:26:40 +0100 (BST) Subject: Kickstart Installs of FC2 In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B006D30EEC@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B006D30EEC@orca.agcom.amgreetings.com> Message-ID: On Wed, 22 Jun 2005, MW Mike Weiner (5028) wrote: >However I have a new issue this morning, I fire up the test box, it >goes out, gets an IP, gets the kickstart script, begins the install >albeit in an X window Include "text" on the kernel command line. (see /usr/share/doc/anaconda-*/command-line.txt) >There was an error installing 'X'.rpm This can indicate a media >failure, lack of disk space, and/or hardware problem. This is a fatal >error and your install will be aborted. Please verify your media and >try your install again. > >Well, when I reboot the system to try again, it fails on yet another >RPM. I rsync'd the RPMS down from loop-mounted ISOs so I am unsure why >it is complaining of bad media. Has anyone else had this problem, and >how do we work around this issue? I see nothing in any logs on the >server providing the install media. Bad media was one of the three possible problems anaconda mentioned. What about hardware - can you replicate this random RPM hiccup on a different client? If not, I'd be testing the box for RAM/disk probs. Cheers, Phil From MWeiner at ag.com Wed Jun 22 16:03:24 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Wed, 22 Jun 2005 12:03:24 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D30F9B@orca.agcom.amgreetings.com> On Wed, 22 Jun 2005, MW Mike Weiner (5028) wrote: >However I have a new issue this morning, I fire up the test box, it >goes out, gets an IP, gets the kickstart script, begins the install >albeit in an X window Include "text" on the kernel command line. (see /usr/share/doc/anaconda-*/command-line.txt) >There was an error installing 'X'.rpm This can indicate a media >failure, lack of disk space, and/or hardware problem. This is a fatal >error and your install will be aborted. Please verify your media and >try your install again. > >Well, when I reboot the system to try again, it fails on yet another >RPM. I rsync'd the RPMS down from loop-mounted ISOs so I am unsure why >it is complaining of bad media. Has anyone else had this problem, and >how do we work around this issue? I see nothing in any logs on the >server providing the install media. Bad media was one of the three possible problems anaconda mentioned. What about hardware - can you replicate this random RPM hiccup on a different client? If not, I'd be testing the box for RAM/disk probs. -- Added text to kernel command line, and I will try to run this on several test boxes to see what the issue is, just for sanity's sake I re-copied the offending RPMS off the disk again. Thanks as always Michael Weiner From MWeiner at ag.com Wed Jun 22 16:39:40 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Wed, 22 Jun 2005 12:39:40 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D30FCD@orca.agcom.amgreetings.com> On Wed, 22 Jun 2005, MW Mike Weiner (5028) wrote: >However I have a new issue this morning, I fire up the test box, it >goes out, gets an IP, gets the kickstart script, begins the install >albeit in an X window Include "text" on the kernel command line. (see /usr/share/doc/anaconda-*/command-line.txt) >There was an error installing 'X'.rpm This can indicate a media >failure, lack of disk space, and/or hardware problem. This is a fatal >error and your install will be aborted. Please verify your media and >try your install again. > >Well, when I reboot the system to try again, it fails on yet another >RPM. I rsync'd the RPMS down from loop-mounted ISOs so I am unsure why >it is complaining of bad media. Has anyone else had this problem, and >how do we work around this issue? I see nothing in any logs on the >server providing the install media. Bad media was one of the three possible problems anaconda mentioned. What about hardware - can you replicate this random RPM hiccup on a different client? If not, I'd be testing the box for RAM/disk probs. -- OK, I grabbed another test box, exactly like the first one, and now it cant seem to see the dhcpd server - OY, one step forward, and three steps back. Any suggestions? Thanks as always Michael Weiner _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From MWeiner at ag.com Wed Jun 22 17:26:04 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Wed, 22 Jun 2005 13:26:04 -0400 Subject: Kickstart Installs of FC2 Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B006D31011@orca.agcom.amgreetings.com> Bad media was one of the three possible problems anaconda mentioned. What about hardware - can you replicate this random RPM hiccup on a different client? If not, I'd be testing the box for RAM/disk probs. -- Yes, seems to carry over to another new box, I had to flash the bios to upgrade the version of PXE on the test box and now that's fine. Just seems to randomly pick an RPM and complain about it, then wants a reboot. Thanks as always Michael Weiner _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From dky at utcc.utoronto.ca Wed Jun 22 17:31:57 2005 From: dky at utcc.utoronto.ca (Derek Yeung) Date: Wed, 22 Jun 2005 13:31:57 -0400 (EDT) Subject: boot: prompt parameters Message-ID: Hello! Is it possible to put parameters at the boot: prompt so that post install scripts can process it and do appropriate things, instead of multiple configuration files? i.e.: boot: ks=http://blah.host.dom/place.cfg parameter1 parameter2 parameter3 Thanks! /dky From error27 at gmail.com Wed Jun 22 17:42:36 2005 From: error27 at gmail.com (Dan Carpenter) Date: Wed, 22 Jun 2005 10:42:36 -0700 Subject: boot: prompt parameters In-Reply-To: References: Message-ID: On 6/22/05, Derek Yeung wrote: > > Is it possible to put parameters at the boot: prompt so that post install > scripts can process it and do appropriate things, instead of multiple > configuration files? > > > i.e.: > boot: ks=http://blah.host.dom/place.cfg parameter1 parameter2 parameter3 > > Yes. The %post script can access the parameters from /proc/cmdline. regards, dan carpenter From pmatilai at laiskiainen.org Wed Jun 22 17:45:13 2005 From: pmatilai at laiskiainen.org (Panu Matilainen) Date: Wed, 22 Jun 2005 20:45:13 +0300 Subject: boot: prompt parameters In-Reply-To: References: Message-ID: <1119462313.31372.7.camel@localhost.localdomain> On Wed, 2005-06-22 at 13:31 -0400, Derek Yeung wrote: > Hello! > > Is it possible to put parameters at the boot: prompt so that post install > scripts can process it and do appropriate things, instead of multiple > configuration files? > > > i.e.: > boot: ks=http://blah.host.dom/place.cfg parameter1 parameter2 parameter3 Sure, as long as you avoid parameters that are interpreted by the kernel itself or anaconda. Just parse your parameters from /proc/cmdline contents in %post or whatever. - Panu - From phr at doc.ic.ac.uk Wed Jun 22 17:46:10 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 22 Jun 2005 18:46:10 +0100 (BST) Subject: boot: prompt parameters In-Reply-To: References: Message-ID: On Wed, 22 Jun 2005, Derek Yeung wrote: >Is it possible to put parameters at the boot: prompt so that post >install scripts can process it and do appropriate things, instead of >multiple configuration files? > >i.e.: >boot: ks=http://blah.host.dom/place.cfg parameter1 parameter2 parameter3 Yes, pluck them out of /proc/cmdline (would have to be in %post --nochroot). Cheers, Phil From brilong at cisco.com Wed Jun 22 18:40:34 2005 From: brilong at cisco.com (Brian Long) Date: Wed, 22 Jun 2005 14:40:34 -0400 Subject: part --grow Message-ID: <1119465634.7352.17.camel@localhost.localdomain> If I have multiple partitions specified in ks.cfg and some of them have --grow specified as well as --maxsize, what is the behavior in distributing free space on the drive in RHEL 3 Update 4? Will free space get evenly divided across all --grow partitions and then maxsize gets taken into account? Or will the first partition be grown to maxsize and then the second partition, etc? Thanks. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From ebrown at lanl.gov Wed Jun 22 19:30:28 2005 From: ebrown at lanl.gov (Ed Brown) Date: Wed, 22 Jun 2005 13:30:28 -0600 Subject: linksleep (was: RE: Kickstart Installs of FC2) In-Reply-To: <66911C8E1F2DA24EB29DE289297A135E48573C@usrd104> References: <66911C8E1F2DA24EB29DE289297A135E48573C@usrd104> Message-ID: <1119468627.17084.472.camel@edbrown.lanl.gov> On Wed, 2005-06-22 at 08:24, Seremeth, Stephen wrote: > label 1 > kernel rhel-4-es/vmlinuz > append ks=nfs:mynfsserver.org:/dist/kickstart/rh_es_4/ > initrd=rhel-4-es/initrd.img ksdevice=eth0 linksleep=120 > eth0_ethtool="speed=100 autoneg=off duplex=full delay_link=1" nofb I don't see the 'linksleep' or 'eth0_ethtool' options in the anaconda(9.1.5.8) docs for RHEL3. I did find a mention of 'linksleep' in a letter from over a year ago from Philip Rowlands, describing the patch that introduced it. The only mention of 'eth0_ethtool' I can find is in RHEL2.1 documentation. Are either of these options and their possible values documented for RHEL3? I assume these are here to address the longstanding problems some users have experienced when kickstarting with the e1000 driver through gigabit switches. Are you (or others?) using these options successfully for other types of installs (e.g., ftp install, static addressing)? I'd prefer not to specify ksdevice in the syslinux config, and to not have multiple labels, for situations where another interface is used for the install. Is it possible to include more than one 'ethX_ethtool' entry, or to set the settings for all e1000 interfaces? thanks for any pointers, -Ed From GavinJ at trutest.co.nz Wed Jun 22 20:58:33 2005 From: GavinJ at trutest.co.nz (Gavin Jackson) Date: Thu, 23 Jun 2005 08:58:33 +1200 Subject: Anaconda and Asus P4T Message-ID: <2460c92dd1b44fb18f67faaf559e6ea442b9d0c1@trutest.co.nz> I tried unsuccessfully to install FC3 this last weekend using initially a kickstart file that I had created and then resorting to using the graphical interface to attempt a manual install and answering all the questions when prompted. I was using a USB flash drive as the location for the kickstart file and half way through the install, I'd get the following error: SystemError: (16, 'Device or resource busy') together with a whole lot of other call trace lines and my config options. I initially though that the USB system might be the problem and transferred the kickstart file to a floppy. I don't recall if the error was the same, but half way thought the install it too spat the dummy. I then attempted the manual graphical way selecting to install a 'minimum' system. This install too would get about halfway and stop, one of the errors given was: RuntimeError: /usr/sbin/kudzu can not be run I did some searching and it turned up quite a few hits of people who had a problem install FC3 on a box that used an Asus P4T mother board. One solution was to update the bios to the latest version which I did but it made no difference. Has anyone here had any problem with using an Asus P4T and if so, was it every solved and how? I have downloaded RH9 and will try that again when I get a chance. I've had Mandrake 9 running and have even built Linux from scratch without any problems, so the hardware should be good. System: Mother board: Asus P4T Ram: 512MB RDRAM CPU Pentium 1.3GHz HDD Two 160GB IDE drives I'd really appreciate any help I can get to solve this problem. Thanks Gavin Jackson ------------------------------------------- Research & Development Software Engineer Tru-Test Ltd PO Box 51078, Pakuranga Auckland, New Zealand DDI +64 9 978 8757 ------------------------------------------- From info at hostinthebox.net Wed Jun 22 22:35:52 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 22 Jun 2005 15:35:52 -0700 Subject: Edited comps.xml - now what? Message-ID: <42B9E7C8.9090603@hostinthebox.net> Hello, all - I've used many suggestions that you guys have provided over the last few weeks to create a distribution that I am comfortable with. However, I'd like to take it one step further here. I have created a comps.xml file that I am happy with. It's my understanding that this file is referenced to 'RedHat/Base/hdlist{2}' to map a package alias name against a real filename. Am I understanding things correctly that if I want to add my own packages to comps.xml, I would somehow need to rebuild a 'RedHat/Base/hdlist{2}' file(s), including the package names of which I would like to add? Documentation regarding anaconda-runtime is very sparse. The best info I've found came from mailing lists. Right now, I'm plowing through the anaconda-runtime files trying to see which steps I need to preform. But any help from you all would be great. Thanks! -dant From klaus.steden at thomson.net Wed Jun 22 22:41:35 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 22 Jun 2005 18:41:35 -0400 Subject: Edited comps.xml - now what? In-Reply-To: <42B9E7C8.9090603@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> Message-ID: <20050622224135.GW2476@thomson.net> > Hello, all - > > I've used many suggestions that you guys have provided over the last few > weeks to create a distribution that I am comfortable with. However, I'd > like to take it one step further here. > > I have created a comps.xml file that I am happy with. It's my > understanding that this file is referenced to 'RedHat/Base/hdlist{2}' to > map a package alias name against a real filename. > > Am I understanding things correctly that if I want to add my own > packages to comps.xml, I would somehow need to rebuild a > 'RedHat/Base/hdlist{2}' file(s), including the package names of which I > would like to add? > > Documentation regarding anaconda-runtime is very sparse. The best info > I've found came from mailing lists. Right now, I'm plowing through the > anaconda-runtime files trying to see which steps I need to preform. But > any help from you all would be great. > Someone posted a link to a python script that fairly neatly encapsulates all the necessary steps (dependency checks, package ordering, and genhdlist) ... I don't have the link handy - but I've been using it for the last three weeks (with a few tweaks) and it's working for both 32 and 64-bit versions of RHEL4 for me. If you want a copy I can send you one. hth, Klaus From info at hostinthebox.net Wed Jun 22 22:48:16 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 22 Jun 2005 15:48:16 -0700 Subject: Edited comps.xml - now what? In-Reply-To: <20050622224135.GW2476@thomson.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> Message-ID: <42B9EAB0.2030807@hostinthebox.net> Klaus Steden wrote: >>Hello, all - >> >>I've used many suggestions that you guys have provided over the last few >>weeks to create a distribution that I am comfortable with. However, I'd >>like to take it one step further here. >> >>I have created a comps.xml file that I am happy with. It's my >>understanding that this file is referenced to 'RedHat/Base/hdlist{2}' to >>map a package alias name against a real filename. >> >>Am I understanding things correctly that if I want to add my own >>packages to comps.xml, I would somehow need to rebuild a >>'RedHat/Base/hdlist{2}' file(s), including the package names of which I >>would like to add? >> >>Documentation regarding anaconda-runtime is very sparse. The best info >>I've found came from mailing lists. Right now, I'm plowing through the >>anaconda-runtime files trying to see which steps I need to preform. But >>any help from you all would be great. >> > > Someone posted a link to a python script that fairly neatly encapsulates all > the necessary steps (dependency checks, package ordering, and genhdlist) ... I > don't have the link handy - but I've been using it for the last three weeks > (with a few tweaks) and it's working for both 32 and 64-bit versions of RHEL4 > for me. > > If you want a copy I can send you one. > > hth, > Klaus > Thanks for responding, Klaus - Mind just telling me the name of the script? Perhaps I can google for it and get some more information on it, such as some documentation. I'd honestly like to fully understand the entire process, rather than just use a script that does it all for me - however, the idea of an all-in-one tool does sound very appealing. Thanks! -dant From klaus.steden at thomson.net Wed Jun 22 22:54:12 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 22 Jun 2005 18:54:12 -0400 Subject: Edited comps.xml - now what? In-Reply-To: <42B9EAB0.2030807@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> Message-ID: <20050622225412.GY2476@thomson.net> > > > > Someone posted a link to a python script that fairly neatly encapsulates > > all the necessary steps (dependency checks, package ordering, and > > genhdlist) ... I don't have the link handy - but I've been using it for > > the last three weeks (with a few tweaks) and it's working for both 32 and > > 64-bit versions of RHEL4 for me. > > > > If you want a copy I can send you one. > > > > hth, > > Klaus > > > > Thanks for responding, Klaus - > > Mind just telling me the name of the script? Perhaps I can google for > it and get some more information on it, such as some documentation. > > I'd honestly like to fully understand the entire process, rather than > just use a script that does it all for me - however, the idea of an > all-in-one tool does sound very appealing. > It was called 'do_update'. http://www.math.uh.edu/~tibbs/do_update (thanks to my Firefox history!). Seems pretty straightforward - I haven't learned python yet, but it wasn't too daunting. Klaus From info at hostinthebox.net Thu Jun 23 03:38:08 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 22 Jun 2005 20:38:08 -0700 Subject: Edited comps.xml - now what? In-Reply-To: <20050622225412.GY2476@thomson.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> Message-ID: <42BA2EA0.6050108@hostinthebox.net> Klaus Steden wrote: >>>Someone posted a link to a python script that fairly neatly encapsulates >>>all the necessary steps (dependency checks, package ordering, and >>>genhdlist) ... I don't have the link handy - but I've been using it for >>>the last three weeks (with a few tweaks) and it's working for both 32 and >>>64-bit versions of RHEL4 for me. >>> >>>If you want a copy I can send you one. >>> >>>hth, >>>Klaus >>> >> >>Thanks for responding, Klaus - >> >>Mind just telling me the name of the script? Perhaps I can google for >>it and get some more information on it, such as some documentation. >> >>I'd honestly like to fully understand the entire process, rather than >>just use a script that does it all for me - however, the idea of an >>all-in-one tool does sound very appealing. >> > > It was called 'do_update'. > > http://www.math.uh.edu/~tibbs/do_update (thanks to my Firefox history!). > > Seems pretty straightforward - I haven't learned python yet, but it wasn't too > daunting. > > Klaus > Klaus - Looks to be pretty neat. Seems a bit out of date, but I think the concept remains the same. However, I'm still looking for some documentation that clearly explains how the procedure is done - what to do after a comps.xml file is edited, how to create the hdlist files, etc etc. I guess I'll keep searching here. ANy feeback by others would also be greatly appreciated. Again, thanks for the time! -dant From phr at doc.ic.ac.uk Thu Jun 23 06:24:10 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 23 Jun 2005 07:24:10 +0100 (BST) Subject: Edited comps.xml - now what? In-Reply-To: <42BA2EA0.6050108@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> <42BA2EA0.6050108@hostinthebox.net> Message-ID: On Wed, 22 Jun 2005, Dan Trainor wrote: >However, I'm still looking for some documentation that clearly explains >how the procedure is done - what to do after a comps.xml file is >edited, how to create the hdlist files, etc etc. genhdlist, also with (depending whether the anaconda bug still exists) pkgorder. The exact incantations appear in the list archives. Cheers, Phil From brilong at cisco.com Thu Jun 23 11:53:11 2005 From: brilong at cisco.com (Brian Long) Date: Thu, 23 Jun 2005 07:53:11 -0400 Subject: Edited comps.xml - now what? In-Reply-To: <42BA2EA0.6050108@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> <42BA2EA0.6050108@hostinthebox.net> Message-ID: <1119527591.8807.1.camel@localhost.localdomain> > However, I'm still looking for some documentation that clearly explains > how the procedure is done - what to do after a comps.xml file is edited, > how to create the hdlist files, etc etc. I guess I'll keep searching > here. ANy feeback by others would also be greatly appreciated. There is a community Wiki set up by Alexander Rau here: http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProject I believe someone on the Fedora development team is working to move the content here: http://www.fedoraproject.org/wiki/Anaconda /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From klaus.steden at thomson.net Thu Jun 23 18:05:32 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 23 Jun 2005 14:05:32 -0400 Subject: rpm error ... what does it mean? In-Reply-To: <42B9E7C8.9090603@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> Message-ID: <20050623180532.GD2476@thomson.net> Hello, I've noticed some of my systems now emit messages like this when I'm querying the installed RPMs ... [root at gilliam ~]# rpm -qa pwlib error: rpmdbNextIterator: skipping h# 608 blob size(3740): BAD, 8 + 16 * il(68) + dl(7444) error: rpmdbNextIterator: skipping h# 608 blob size(3740): BAD, 8 + 16 * il(68) + dl(7444) pwlib-1.6.5-11 ... what does that mean, and what could have caused it? thanks, Klaus From brandono at berkeley.edu Thu Jun 23 18:11:55 2005 From: brandono at berkeley.edu (Brandon Ooi) Date: Thu, 23 Jun 2005 11:11:55 -0700 Subject: FC4 kickstart broken? Message-ID: <42BAFB6B.4080809@berkeley.edu> Hi guys, I was hoping to shed some light on FC4's kickstart (anaconda). Graphical installs work great but the Fedora team seems to be leaving the server guys behind. My kickstart file is fairly basic. Because of our high-load server environment, we try to trim down our builds and remove things a server should never use (bluetooth, cups, wireless tools... etc...etc...). I usually do this in the %packages section with --resolvedeps. Simplified example: %package --resolvedeps @ editors kernel-smp gcc libpng-devel -bluez-utils <== breakage occurs here -bluez-hcidump This kickstart file worked perfectly in FC3. In FC4, when you try to actually remove a package the installer crashes with an "unhandled exception" telling me to report this as a bug. Something about a key error in kickstart.py? If you comment out all the package removals, it "works." By works, i mean it seems to install all the packages, but the resulting anaconda-ks.cfg file in /root does not have anything meaningful in the %package section. It almost seems like package selection in the kickstart files is not supported anymore? I have not found a similar bug on bugzilla which makes me wonder if something changed and I'm not aware? if not I will file a bugzilla bug. Doing a graphical install of FC4 on all our machines would take forever! Any help would be appreciated :) Brandon Ooi From espenas at gmail.com Thu Jun 23 18:34:50 2005 From: espenas at gmail.com (Espen Stefansen) Date: Thu, 23 Jun 2005 20:34:50 +0200 Subject: FC4 kickstart broken? In-Reply-To: <42BAFB6B.4080809@berkeley.edu> References: <42BAFB6B.4080809@berkeley.edu> Message-ID: <9699b8ce050623113470ca2425@mail.gmail.com> Hi I noticed this myself. It took me several hours to figure this out. A working day well spent. ;) So this is definately a bug. Looks almost like a showstopper to me. All of my Fedora kickstart-files fail. Definately not good. I'm thinking of trying to use yum to remove such things. But i'm afraid that dependencies will be the problem here. Regards Espen On 6/23/05, Brandon Ooi wrote: > Hi guys, > > I was hoping to shed some light on FC4's kickstart (anaconda). Graphical > installs work great but the Fedora team seems to be leaving the server > guys behind. > > My kickstart file is fairly basic. Because of our high-load server > environment, we try to trim down our builds and remove things a server > should never use (bluetooth, cups, wireless tools... etc...etc...). I > usually do this in the %packages section with --resolvedeps. > > Simplified example: > > %package --resolvedeps > @ editors > kernel-smp > gcc > libpng-devel > -bluez-utils <== breakage occurs here > -bluez-hcidump > > This kickstart file worked perfectly in FC3. In FC4, when you try to > actually remove a package the installer crashes with an "unhandled > exception" telling me to report this as a bug. Something about a key > error in kickstart.py? > > If you comment out all the package removals, it "works." By works, i > mean it seems to install all the packages, but the resulting > anaconda-ks.cfg file in /root does not have anything meaningful in the > %package section. It almost seems like package selection in the > kickstart files is not supported anymore? I have not found a similar bug > on bugzilla which makes me wonder if something changed and I'm not > aware? if not I will file a bugzilla bug. Doing a graphical install of > FC4 on all our machines would take forever! > > Any help would be appreciated :) > > Brandon Ooi > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From brandono at berkeley.edu Thu Jun 23 20:07:12 2005 From: brandono at berkeley.edu (Brandon Ooi) Date: Thu, 23 Jun 2005 13:07:12 -0700 Subject: FC4 kickstart broken? In-Reply-To: <9699b8ce050623113470ca2425@mail.gmail.com> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> Message-ID: <42BB1670.1060309@berkeley.edu> Hi, I agree with Espen, I tried removing some of the packages with rpm -e and yum but there are problems. I'm not sure how such a big bug could get into a Fedora major release. Brandon Espen Stefansen wrote: >Hi > >I noticed this myself. It took me several hours to figure this out. A >working day well spent. ;) > >So this is definately a bug. Looks almost like a showstopper to me. >All of my Fedora kickstart-files fail. Definately not good. > >I'm thinking of trying to use yum to remove such things. But i'm >afraid that dependencies will be the problem here. > >Regards >Espen > > >On 6/23/05, Brandon Ooi wrote: > > >>Hi guys, >> >>I was hoping to shed some light on FC4's kickstart (anaconda). Graphical >>installs work great but the Fedora team seems to be leaving the server >>guys behind. >> >>My kickstart file is fairly basic. Because of our high-load server >>environment, we try to trim down our builds and remove things a server >>should never use (bluetooth, cups, wireless tools... etc...etc...). I >>usually do this in the %packages section with --resolvedeps. >> >>Simplified example: >> >>%package --resolvedeps >>@ editors >>kernel-smp >>gcc >>libpng-devel >>-bluez-utils <== breakage occurs here >>-bluez-hcidump >> >>This kickstart file worked perfectly in FC3. In FC4, when you try to >>actually remove a package the installer crashes with an "unhandled >>exception" telling me to report this as a bug. Something about a key >>error in kickstart.py? >> >>If you comment out all the package removals, it "works." By works, i >>mean it seems to install all the packages, but the resulting >>anaconda-ks.cfg file in /root does not have anything meaningful in the >>%package section. It almost seems like package selection in the >>kickstart files is not supported anymore? I have not found a similar bug >>on bugzilla which makes me wonder if something changed and I'm not >>aware? if not I will file a bugzilla bug. Doing a graphical install of >>FC4 on all our machines would take forever! >> >>Any help would be appreciated :) >> >>Brandon Ooi >> >>_______________________________________________ >>Kickstart-list mailing list >>Kickstart-list at redhat.com >>https://www.redhat.com/mailman/listinfo/kickstart-list >> >> >> > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > From skvidal at phy.duke.edu Thu Jun 23 20:08:31 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Thu, 23 Jun 2005 16:08:31 -0400 Subject: FC4 kickstart broken? In-Reply-To: <42BB1670.1060309@berkeley.edu> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> Message-ID: <1119557311.32522.55.camel@opus.phy.duke.edu> On Thu, 2005-06-23 at 13:07 -0700, Brandon Ooi wrote: > Hi, > > I agree with Espen, I tried removing some of the packages with rpm -e > and yum but there are problems. I'm not sure how such a big bug could > get into a Fedora major release. there are problems removing packages with yum in %post of fc4? What problems? -sv From espenas at gmail.com Thu Jun 23 20:41:10 2005 From: espenas at gmail.com (Espen Stefansen) Date: Thu, 23 Jun 2005 22:41:10 +0200 Subject: FC4 kickstart broken? In-Reply-To: <1119557311.32522.55.camel@opus.phy.duke.edu> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> <1119557311.32522.55.camel@opus.phy.duke.edu> Message-ID: <9699b8ce05062313413fb3f8fa@mail.gmail.com> On 6/23/05, seth vidal wrote: > On Thu, 2005-06-23 at 13:07 -0700, Brandon Ooi wrote: > > Hi, > > > > I agree with Espen, I tried removing some of the packages with rpm -e > > and yum but there are problems. I'm not sure how such a big bug could > > get into a Fedora major release. > > there are problems removing packages with yum in %post of fc4? > > What problems? > Yum is not the problem. The problem is that in %packages u cant remove packages. Anaconda will fail. And what i meant with dependencies, is that it's easier not to install, than to install and then uninstall later with yum. Espen From skvidal at phy.duke.edu Thu Jun 23 20:44:26 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Thu, 23 Jun 2005 16:44:26 -0400 Subject: FC4 kickstart broken? In-Reply-To: <9699b8ce05062313413fb3f8fa@mail.gmail.com> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> <1119557311.32522.55.camel@opus.phy.duke.edu> <9699b8ce05062313413fb3f8fa@mail.gmail.com> Message-ID: <1119559466.32522.64.camel@opus.phy.duke.edu> On Thu, 2005-06-23 at 22:41 +0200, Espen Stefansen wrote: > On 6/23/05, seth vidal wrote: > > On Thu, 2005-06-23 at 13:07 -0700, Brandon Ooi wrote: > > > Hi, > > > > > > I agree with Espen, I tried removing some of the packages with rpm -e > > > and yum but there are problems. I'm not sure how such a big bug could > > > get into a Fedora major release. > > > > there are problems removing packages with yum in %post of fc4? > > > > What problems? > > > > Yum is not the problem. The problem is that in %packages u cant remove > packages. Anaconda will fail. > > And what i meant with dependencies, is that it's easier not to > install, than to install and then uninstall later with yum. > oh, okay. I must have misread that. thanks! -sv From brandono at berkeley.edu Thu Jun 23 20:49:31 2005 From: brandono at berkeley.edu (Brandon Ooi) Date: Thu, 23 Jun 2005 13:49:31 -0700 Subject: FC4 kickstart broken? In-Reply-To: <1119559466.32522.64.camel@opus.phy.duke.edu> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> <1119557311.32522.55.camel@opus.phy.duke.edu> <9699b8ce05062313413fb3f8fa@mail.gmail.com> <1119559466.32522.64.camel@opus.phy.duke.edu> Message-ID: <42BB205B.9000801@berkeley.edu> hi, the problem is in %post yum isn't configured correctly yet (not to mention the machine doesn't have an external connection yet). yum doesn't run without an accessible baseurl? brandon seth vidal wrote: >On Thu, 2005-06-23 at 22:41 +0200, Espen Stefansen wrote: > > >>On 6/23/05, seth vidal wrote: >> >> >>>On Thu, 2005-06-23 at 13:07 -0700, Brandon Ooi wrote: >>> >>> >>>>Hi, >>>> >>>>I agree with Espen, I tried removing some of the packages with rpm -e >>>>and yum but there are problems. I'm not sure how such a big bug could >>>>get into a Fedora major release. >>>> >>>> >>>there are problems removing packages with yum in %post of fc4? >>> >>>What problems? >>> >>> >>> >>Yum is not the problem. The problem is that in %packages u cant remove >>packages. Anaconda will fail. >> >>And what i meant with dependencies, is that it's easier not to >>install, than to install and then uninstall later with yum. >> >> >> > >oh, okay. I must have misread that. > >thanks! >-sv > > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > From skvidal at phy.duke.edu Thu Jun 23 20:51:10 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Thu, 23 Jun 2005 16:51:10 -0400 Subject: FC4 kickstart broken? In-Reply-To: <42BB205B.9000801@berkeley.edu> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> <1119557311.32522.55.camel@opus.phy.duke.edu> <9699b8ce05062313413fb3f8fa@mail.gmail.com> <1119559466.32522.64.camel@opus.phy.duke.edu> <42BB205B.9000801@berkeley.edu> Message-ID: <1119559870.32522.73.camel@opus.phy.duke.edu> On Thu, 2005-06-23 at 13:49 -0700, Brandon Ooi wrote: > hi, > > the problem is in %post yum isn't configured correctly yet (not to > mention the machine doesn't have an external connection yet). yum > doesn't run without an accessible baseurl? > it's not configured yet? Why not? Moreover why can't you just echo in your new yum.conf and repo files at the start of the %post and go from there. -sv From brandono at berkeley.edu Thu Jun 23 21:00:18 2005 From: brandono at berkeley.edu (Brandon Ooi) Date: Thu, 23 Jun 2005 14:00:18 -0700 Subject: FC4 kickstart broken? In-Reply-To: <1119559870.32522.73.camel@opus.phy.duke.edu> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> <1119557311.32522.55.camel@opus.phy.duke.edu> <9699b8ce05062313413fb3f8fa@mail.gmail.com> <1119559466.32522.64.camel@opus.phy.duke.edu> <42BB205B.9000801@berkeley.edu> <1119559870.32522.73.camel@opus.phy.duke.edu> Message-ID: <42BB22E2.8070405@berkeley.edu> that's true I could do that but the machines don't have external connections. Is it possible to use yum to remove packages without a connection to an actual repository? I'd rather do that than do the same with rpm -e. brandon seth vidal wrote: >On Thu, 2005-06-23 at 13:49 -0700, Brandon Ooi wrote: > > >>hi, >> >>the problem is in %post yum isn't configured correctly yet (not to >>mention the machine doesn't have an external connection yet). yum >>doesn't run without an accessible baseurl? >> >> >> > >it's not configured yet? Why not? Moreover why can't you just echo in >your new yum.conf and repo files at the start of the %post and go from >there. > >-sv > > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > From skvidal at phy.duke.edu Thu Jun 23 21:02:57 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Thu, 23 Jun 2005 17:02:57 -0400 Subject: FC4 kickstart broken? In-Reply-To: <42BB22E2.8070405@berkeley.edu> References: <42BAFB6B.4080809@berkeley.edu> <9699b8ce050623113470ca2425@mail.gmail.com> <42BB1670.1060309@berkeley.edu> <1119557311.32522.55.camel@opus.phy.duke.edu> <9699b8ce05062313413fb3f8fa@mail.gmail.com> <1119559466.32522.64.camel@opus.phy.duke.edu> <42BB205B.9000801@berkeley.edu> <1119559870.32522.73.camel@opus.phy.duke.edu> <42BB22E2.8070405@berkeley.edu> Message-ID: <1119560577.32522.87.camel@opus.phy.duke.edu> On Thu, 2005-06-23 at 14:00 -0700, Brandon Ooi wrote: > that's true I could do that but the machines don't have external > connections. Is it possible to use yum to remove packages without a > connection to an actual repository? I'd rather do that than do the same > with rpm -e. > sorta. You could enable caching (yum -C) and try that out. I think that should work w/o fault. -sv From info at hostinthebox.net Thu Jun 23 22:53:34 2005 From: info at hostinthebox.net (Dan Trainor) Date: Thu, 23 Jun 2005 15:53:34 -0700 Subject: Redirecting Anaconda output Message-ID: <42BB3D6E.1040200@hostinthebox.net> Hello, all - I was wondering if it was possible somehow to redirect Anaconda's output, including debug output detailing crashes, to an external source - such as a remote syslog. The reason why I ask is that I'd like to keep records of these failed installs so that I can have more information to better troubleshoot this information - and writing down 100 lines of code by reading a monitor to the side of me, should not be an option. Thanks! -dant From info at hostinthebox.net Thu Jun 23 23:10:52 2005 From: info at hostinthebox.net (Dan Trainor) Date: Thu, 23 Jun 2005 16:10:52 -0700 Subject: Edited comps.xml - now what? In-Reply-To: <42BA2EA0.6050108@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> <42BA2EA0.6050108@hostinthebox.net> Message-ID: <42BB417C.60200@hostinthebox.net> Dan Trainor wrote: > Klaus Steden wrote: > >>>>Someone posted a link to a python script that fairly neatly encapsulates >>>>all the necessary steps (dependency checks, package ordering, and >>>>genhdlist) ... I don't have the link handy - but I've been using it for >>>>the last three weeks (with a few tweaks) and it's working for both 32 and >>>>64-bit versions of RHEL4 for me. >>>> >>>>If you want a copy I can send you one. >>>> >>>>hth, >>>>Klaus >>>> >>> >>>Thanks for responding, Klaus - >>> >>>Mind just telling me the name of the script? Perhaps I can google for >>>it and get some more information on it, such as some documentation. >>> >>>I'd honestly like to fully understand the entire process, rather than >>>just use a script that does it all for me - however, the idea of an >>>all-in-one tool does sound very appealing. >>> >> >>It was called 'do_update'. >> >>http://www.math.uh.edu/~tibbs/do_update (thanks to my Firefox history!). >> >>Seems pretty straightforward - I haven't learned python yet, but it wasn't too >>daunting. >> >>Klaus >> > > > Klaus - > > Looks to be pretty neat. Seems a bit out of date, but I think the > concept remains the same. > > However, I'm still looking for some documentation that clearly explains > how the procedure is done - what to do after a comps.xml file is edited, > how to create the hdlist files, etc etc. I guess I'll keep searching > here. ANy feeback by others would also be greatly appreciated. > > Again, thanks for the time! > -dant > Forgive me for letting these emails get out-of-sync. I downloaded them to my laptop at home, and have no way to retrieve them - but this was a response, retreived rfom the kickstart-list archives, by Phil: ...snip... On Wed, 22 Jun 2005, Dan Trainor wrote: >However, I'm still looking for some documentation that clearly explains >how the procedure is done - what to do after a comps.xml file is >edited, how to create the hdlist files, etc etc. genhdlist, also with (depending whether the anaconda bug still exists) pkgorder. The exact incantations appear in the list archives. ...snip... Phil - I think you're correct, that genhdlist is my answer. I still don't understand exactly how it works. And the exact incantations (nice word, btw - very fitting) do not appear in the list archives. I think I found a total of ~20 emails that had any reference to the string 'genhdlist'. I'm a bit lost for information. The best I've found is a few examples of people who are trying to modify 'base/hdlist{2}'. So I run it, and I get: genhdlist: genhdlist [--withnumbers] [--fileorder ] [--hdlist ] [--productpath ] + What does --with-numbers do? What does --fileorder do? What is --hdlist referencing? What is a --productpath? Which additional paths is it asking for? ...and we're just getting started. I'd like to become familiar with all the tools that come with anaconda-runtime-10.0-5, because they *look* to be pretty important and useful. I just can't find any documentation anywhere. I've been telling myself the entire time: "Hold back, don't ask these guys, just RTFM!" - then I think to myself, "Where is TFM?". Is there any formal documentation on the anaconda-runtime suite of utilities? Again, I greatly appreciate the time. Thanks -dant From dns.prasad at ge.com Fri Jun 24 01:11:55 2005 From: dns.prasad at ge.com (Prasad, DNS (Corporate, consultant)) Date: Fri, 24 Jun 2005 06:41:55 +0530 Subject: Redirecting Anaconda output Message-ID: Hi Dant, You can get the installed RPMS information from install.log you find install.log under /root. install.log stores the information about which RPMS get installed and which get uninstalled during the installation time Regards, Prasad. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Dan Trainor Sent: Friday, June 24, 2005 4:24 AM To: Discussion list about Kickstart Subject: Redirecting Anaconda output Hello, all - I was wondering if it was possible somehow to redirect Anaconda's output, including debug output detailing crashes, to an external source - such as a remote syslog. The reason why I ask is that I'd like to keep records of these failed installs so that I can have more information to better troubleshoot this information - and writing down 100 lines of code by reading a monitor to the side of me, should not be an option. Thanks! -dant _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From info at hostinthebox.net Fri Jun 24 01:58:57 2005 From: info at hostinthebox.net (Dan Trainor) Date: Thu, 23 Jun 2005 18:58:57 -0700 Subject: Redirecting Anaconda output In-Reply-To: References: Message-ID: <42BB68E1.3010804@hostinthebox.net> Prasad, DNS (Corporate, consultant) wrote: > Hi Dant, > > You can get the installed RPMS information from install.log > you find install.log under /root. > > install.log stores the information about which RPMS get installed and which get uninstalled during the installation time > > > Regards, > Prasad. > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com]On Behalf Of Dan Trainor > Sent: Friday, June 24, 2005 4:24 AM > To: Discussion list about Kickstart > Subject: Redirecting Anaconda output > > > Hello, all - > > I was wondering if it was possible somehow to redirect Anaconda's > output, including debug output detailing crashes, to an external source > - such as a remote syslog. > > The reason why I ask is that I'd like to keep records of these failed > installs so that I can have more information to better troubleshoot this > information - and writing down 100 lines of code by reading a monitor to > the side of me, should not be an option. > > Thanks! > -dant > Hello, Prasad, and thanks for the response - I am aware of the install.log, but I am loking for is a way to redirect Anaconda's messages during the actuall install, or attempt to install, so that I can debug failed installs. Thanks! -dant From klaus.steden at thomson.net Fri Jun 24 02:06:40 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 23 Jun 2005 22:06:40 -0400 Subject: Redirecting Anaconda output In-Reply-To: <42BB68E1.3010804@hostinthebox.net> References: <42BB68E1.3010804@hostinthebox.net> Message-ID: <20050624020640.GJ2476@thomson.net> > Hello, Prasad, and thanks for the response - > > I am aware of the install.log, but I am loking for is a way to redirect > Anaconda's messages during the actuall install, or attempt to install, > so that I can debug failed installs. > I have to say that that wouldn't suck, as an option. >From what I've seen looking through the anaconda source code, it wouldn't be too difficult to implement, at least the C portion. Klaus From info at hostinthebox.net Fri Jun 24 02:12:55 2005 From: info at hostinthebox.net (Dan Trainor) Date: Thu, 23 Jun 2005 19:12:55 -0700 Subject: Redirecting Anaconda output In-Reply-To: <20050624020640.GJ2476@thomson.net> References: <42BB68E1.3010804@hostinthebox.net> <20050624020640.GJ2476@thomson.net> Message-ID: <42BB6C27.1060802@hostinthebox.net> Klaus Steden wrote: >>Hello, Prasad, and thanks for the response - >> >>I am aware of the install.log, but I am loking for is a way to redirect >>Anaconda's messages during the actuall install, or attempt to install, >>so that I can debug failed installs. >> > > I have to say that that wouldn't suck, as an option. > >>From what I've seen looking through the anaconda source code, it wouldn't be > too difficult to implement, at least the C portion. > > Klaus > Wonder if I can write something to broadcast over syslog to another machine. Do you think you'd be able to write this in C? I'm sure others would be interested in this. Thanks -dant From klaus.steden at thomson.net Fri Jun 24 02:55:37 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 23 Jun 2005 22:55:37 -0400 Subject: Redirecting Anaconda output In-Reply-To: <42BB6C27.1060802@hostinthebox.net> References: <42BB68E1.3010804@hostinthebox.net> <20050624020640.GJ2476@thomson.net> <42BB6C27.1060802@hostinthebox.net> Message-ID: <20050624025537.GK2476@thomson.net> I could have sworn Dan Trainor said this Thu, Jun 23, 2005 at 07:12:55PM -0700 ... > Klaus Steden wrote: > >>Hello, Prasad, and thanks for the response - > >> > >>I am aware of the install.log, but I am loking for is a way to redirect > >>Anaconda's messages during the actuall install, or attempt to install, so > >>that I can debug failed installs. > >> > > > > I have to say that that wouldn't suck, as an option. > > > >>From what I've seen looking through the anaconda source code, it wouldn't > >>be too difficult to implement, at least the C portion. > > > > Klaus > > > > Wonder if I can write something to broadcast over syslog to another machine. > In which language? I only know the C one well - although I'd be willing to surmise that the Python approach (like everything else Python) is easy. > > Do you think you'd be able to write this in C? I'm sure others would be > interested in this. > I think I could, although right now I'm way short on time due to work demands. The C API for syslog is fairly straightforward, and is part of libc ... however, the client side of it is pretty dumb and it can only forward logging messages to a remote syslogd if there is a local syslogd - since the daemon is responsible for the forwarding, as instructed by /etc/syslog.conf. An alternative approach may be to use the 'logger' shell command with the '-u' switch, after initializing something to hand off data received from a local Unix domain socket to a remote syslogd. A better approach, imo, would be to incorporate functionality that the *BSD people put in their 'logger' command - a '-h' switch that allows you to specify a destination host. In the end, syslog is simply a UDP message-passing framework with some filtering logic on top. Shouldn't be too hard - but I know it'll be some weeks before I can even entertain the idea. What's everyone else out there think of the idea? Klaus From gotosathya at gmail.com Fri Jun 24 05:07:02 2005 From: gotosathya at gmail.com (Sathya M) Date: Fri, 24 Jun 2005 10:37:02 +0530 Subject: About boot information Message-ID: Hello all, I have one doubt regarding Installation. At the time of installation we will give the boot information in boot promt. This is taken by Python program during installation. Doubt is where that information is stored & retrive by installation program? Thanks, Sathy From klaus.steden at thomson.net Fri Jun 24 05:18:57 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 24 Jun 2005 01:18:57 -0400 Subject: About boot information In-Reply-To: References: Message-ID: <20050624051857.GL2476@thomson.net> > Hello all, > I have one doubt regarding Installation. At the time of > installation we will give the boot information in boot promt. This is > taken by Python program during installation. > Doubt is where that information is stored & retrive by installation program? > That info is stored in /proc/cmdline by the Linux bootstrap process, so this is where you should start looking. hth, Klaus From SSeremeth at anacomp.com Fri Jun 24 12:16:56 2005 From: SSeremeth at anacomp.com (Seremeth, Stephen) Date: Fri, 24 Jun 2005 08:16:56 -0400 Subject: Redirecting Anaconda output Message-ID: <66911C8E1F2DA24EB29DE289297A135E485746@usrd104> > What's everyone else out there think of the idea? I think it's something that's long overdue and would be well appreciated by _everybody_ using ks. Recently I had some kickstart attempts that bombed at the _end_ of kickstarting -- and when that happens, kickstart actually _does_ save some good logs for you. I think they were in /var/log but I'm not remembering completely now, but it included some/all of the stuff from the other virtual consoles, etc.. So, the idea has been thought of before, I guess, even if the logs are even more valuable when ks bombs half way through rather than after you have a mostly working install. If I can dig up more info, I'll post it. Also, re: anaconda docs and genhdlist -- the best info I had found was on the previously referenced site which I couldn't reach yesterday: http://rau.homedns.org/twiki/bin/view/Anaconda/AnacondaDocumentationProj ect And as the previous post said it looks to be migrating (not nearly completely yet) to: http://www.fedoraproject.org/wiki/Anaconda There's also a slightly outdated linux journal article which goes into some detail -- be sure to read the comments at the end of the article for additional hints. Steve From klaus.steden at thomson.net Fri Jun 24 22:43:20 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 24 Jun 2005 18:43:20 -0400 Subject: Redirecting Anaconda output In-Reply-To: <66911C8E1F2DA24EB29DE289297A135E485746@usrd104> References: <66911C8E1F2DA24EB29DE289297A135E485746@usrd104> Message-ID: <20050624224320.GK2476@thomson.net> > > What's everyone else out there think of the idea? > > I think it's something that's long overdue and would be well appreciated by > _everybody_ using ks. Recently I had some kickstart attempts that bombed at > the _end_ of kickstarting -- and when that happens, kickstart actually > _does_ save some good logs for you. I think they were in /var/log but I'm > not remembering completely now, but it included some/all of the stuff from > the other virtual consoles, etc.. So, the idea has been thought of before, > I guess, even if the logs are even more valuable when ks bombs half way > through rather than after you have a mostly working install. If I can dig > up more info, I'll post it. > > Also, re: anaconda docs and genhdlist -- the best info I had found was on > the previously referenced site which I couldn't reach yesterday: > I just took a glance through the source code to 'logger' in the FreeBSD source tree. It's pretty compact, as these things go - no direct syslog interface - just straight socket I/O and some parsing of stdio. The file itself is less than 300 lines - some 45 of which are copyright and versioning info. The actual business end of the code consists of five functions, including main(). It should be fairly trivial to incorporate this kind of facility into the source tree of the 'loader' binary. Is this the right place for it? Or would it be more a matter of adding a logger command to the bootstrap/busybox tools and invoking it with system() calls from anaconda? Klaus From info at hostinthebox.net Fri Jun 24 22:46:19 2005 From: info at hostinthebox.net (Dan Trainor) Date: Fri, 24 Jun 2005 15:46:19 -0700 Subject: Redirecting Anaconda output In-Reply-To: <20050624224320.GK2476@thomson.net> References: <66911C8E1F2DA24EB29DE289297A135E485746@usrd104> <20050624224320.GK2476@thomson.net> Message-ID: <42BC8D3B.3060708@hostinthebox.net> Klaus Steden wrote: >>>What's everyone else out there think of the idea? >> >>I think it's something that's long overdue and would be well appreciated by >>_everybody_ using ks. Recently I had some kickstart attempts that bombed at >>the _end_ of kickstarting -- and when that happens, kickstart actually >>_does_ save some good logs for you. I think they were in /var/log but I'm >>not remembering completely now, but it included some/all of the stuff from >>the other virtual consoles, etc.. So, the idea has been thought of before, >>I guess, even if the logs are even more valuable when ks bombs half way >>through rather than after you have a mostly working install. If I can dig >>up more info, I'll post it. >> >>Also, re: anaconda docs and genhdlist -- the best info I had found was on >>the previously referenced site which I couldn't reach yesterday: >> > > I just took a glance through the source code to 'logger' in the FreeBSD source > tree. > > It's pretty compact, as these things go - no direct syslog interface - just > straight socket I/O and some parsing of stdio. > > The file itself is less than 300 lines - some 45 of which are copyright and > versioning info. > > The actual business end of the code consists of five functions, including > main(). It should be fairly trivial to incorporate this kind of facility into > the source tree of the 'loader' binary. > > Is this the right place for it? Or would it be more a matter of adding a > logger command to the bootstrap/busybox tools and invoking it with system() > calls from anaconda? > > Klaus > Klaus - I cannot answer that, I'm sorry. I lack intimate knowledge of Anaconda. However, if you do complete this, I will buy you a case of beer. And I mean it. Thanks! -dant From info at hostinthebox.net Fri Jun 24 23:39:41 2005 From: info at hostinthebox.net (Dan Trainor) Date: Fri, 24 Jun 2005 16:39:41 -0700 Subject: Dropping to a shell in %post Message-ID: <42BC99BD.9030108@hostinthebox.net> Quick one for you guys again - Is it possible to drop to a sh shell in or after %post, so that I can poke around and get familiar with the %post enviornment to see about donig some things later on? Thanks! -dant From error27 at gmail.com Sat Jun 25 00:17:31 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 24 Jun 2005 17:17:31 -0700 Subject: Dropping to a shell in %post In-Reply-To: <42BC99BD.9030108@hostinthebox.net> References: <42BC99BD.9030108@hostinthebox.net> Message-ID: On 6/24/05, Dan Trainor wrote: > Is it possible to drop to a sh shell in or after %post, so that I can > poke around and get familiar with the %post enviornment to see about > donig some things later on? add these lines to your %post. echo "Dropping to a shell" open -c 8 bash read tmp The echo prints to tty3 so you probably won't see it... Not sure how to put stuff on the first tty. This will pause the install until you press Enter on the front screen. You can switch to tty8 to do whatever you want. regards, dan carpenter From info at hostinthebox.net Sat Jun 25 00:27:37 2005 From: info at hostinthebox.net (Dan Trainor) Date: Fri, 24 Jun 2005 17:27:37 -0700 Subject: Dropping to a shell in %post In-Reply-To: References: <42BC99BD.9030108@hostinthebox.net> Message-ID: <42BCA4F9.6030805@hostinthebox.net> Dan Carpenter wrote: > On 6/24/05, Dan Trainor wrote: > > >>Is it possible to drop to a sh shell in or after %post, so that I can >>poke around and get familiar with the %post enviornment to see about >>donig some things later on? > > > add these lines to your %post. > > echo "Dropping to a shell" > open -c 8 bash > read tmp > > The echo prints to tty3 so you probably won't see it... Not sure how > to put stuff on the first tty. This will pause the install until you > press Enter on the front screen. You can switch to tty8 to do > whatever you want. > > regards, > dan carpenter > Dan - Awesome, that's what I was looking for. I'm having troule finding out why I can't install two additional RPMs in %post, saying that the files are not present. I am copying them from /mnt/source/thedirthatiputfilesin to /mnt/sysimage/tmp/mynewdir, so I'd like to troubleshoot that. Thanks -dant From error27 at gmail.com Sat Jun 25 01:40:26 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 24 Jun 2005 18:40:26 -0700 Subject: Dropping to a shell in %post In-Reply-To: <42BCA4F9.6030805@hostinthebox.net> References: <42BC99BD.9030108@hostinthebox.net> <42BCA4F9.6030805@hostinthebox.net> Message-ID: > /mnt/source/thedirthatiputfilesin to /mnt/sysimage/tmp/mynewdir %post takes place in `chroot /mnt/sysimage/` so there is no source/ under /mnt any more but you would have figured that out on your own by dropping to a shell. regards, dan carpenter From dan at half-asleep.com Sat Jun 25 01:43:18 2005 From: dan at half-asleep.com (Daniel Segall) Date: Fri, 24 Jun 2005 21:43:18 -0400 (EDT) Subject: Dropping to a shell in %post In-Reply-To: <42BCA4F9.6030805@hostinthebox.net> References: <42BC99BD.9030108@hostinthebox.net> <42BCA4F9.6030805@hostinthebox.net> Message-ID: <50706.69.33.32.187.1119663798.squirrel@webmail.half-asleep.com> Are you actually specifying that full path? If so, that would be your problem. You are already chroot'd in /mnt/sysimage, so if you told it /mnt/sysimage/somepath, it would be interpreted literally. -Dan > Dan Carpenter wrote: >> On 6/24/05, Dan Trainor wrote: >> >> >>>Is it possible to drop to a sh shell in or after %post, so that I can >>>poke around and get familiar with the %post enviornment to see about >>>donig some things later on? >> >> >> add these lines to your %post. >> >> echo "Dropping to a shell" >> open -c 8 bash >> read tmp >> >> The echo prints to tty3 so you probably won't see it... Not sure how >> to put stuff on the first tty. This will pause the install until you >> press Enter on the front screen. You can switch to tty8 to do >> whatever you want. >> >> regards, >> dan carpenter >> > > Dan - > > Awesome, that's what I was looking for. I'm having troule finding out > why I can't install two additional RPMs in %post, saying that the files > are not present. I am copying them from > /mnt/source/thedirthatiputfilesin to /mnt/sysimage/tmp/mynewdir, so I'd > like to troubleshoot that. > > Thanks > -dant > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > From hahaha_30k at yahoo.com Sat Jun 25 01:45:17 2005 From: hahaha_30k at yahoo.com (ha haha) Date: Fri, 24 Jun 2005 18:45:17 -0700 (PDT) Subject: kickstart waits when machines have two network cards Message-ID: <20050625014517.27458.qmail@web30213.mail.mud.yahoo.com> Hi, I try to begin unattended kickstart installation for FC4 onto my machine but it fails, because unfortunately the machine has e two network cards and so the kickstart hangs for manual input, to select which network card to continue. Is there a way to specify a network card explicitly to grasp ks.cfg for NFS installation? Thanks. __________________________________ Discover Yahoo! Get on-the-go sports scores, stock quotes, news and more. Check it out! http://discover.yahoo.com/mobile.html From joe at swelltech.com Sat Jun 25 01:49:43 2005 From: joe at swelltech.com (Joe Cooper) Date: Fri, 24 Jun 2005 20:49:43 -0500 Subject: kickstart waits when machines have two network cards In-Reply-To: <20050625014517.27458.qmail@web30213.mail.mud.yahoo.com> References: <20050625014517.27458.qmail@web30213.mail.mud.yahoo.com> Message-ID: <42BCB837.80406@swelltech.com> ha haha wrote: > Hi, > > I try to begin unattended kickstart installation for > FC4 onto my machine but it fails, because > unfortunately the machine has e two network cards and > so the kickstart hangs for manual input, to select > which network card to continue. > > Is there a way to specify a network card explicitly > to grasp ks.cfg for NFS installation? Yes: network --device eth0 --bootproto dhcp From st.moloney at gte.net Sat Jun 25 01:50:14 2005 From: st.moloney at gte.net (Sharon and Tim Moloney) Date: Fri, 24 Jun 2005 21:50:14 -0400 Subject: kickstart waits when machines have two network cards References: <20050625014517.27458.qmail@web30213.mail.mud.yahoo.com> Message-ID: <021001c57928$3bb33460$3301a8c0@saporo> "ha haha" wrote: > I try to begin unattended kickstart installation for > FC4 onto my machine but it fails, because > unfortunately the machine has e two network cards and > so the kickstart hangs for manual input, to select > which network card to continue. > > Is there a way to specify a network card explicitly > to grasp ks.cfg for NFS installation? On the boot command line, you can add "ksdevice=eth0" to specify the first NIC or "ksdevice=link" to specify the one that has an active link status (assuming only one has an active link status). From hahaha_30k at yahoo.com Sat Jun 25 01:54:49 2005 From: hahaha_30k at yahoo.com (ha haha) Date: Fri, 24 Jun 2005 18:54:49 -0700 (PDT) Subject: SATA disk modules load sequence for FC4 kickstart Message-ID: <20050625015449.30341.qmail@web30211.mail.mud.yahoo.com> Hi, Ive got another problem for FC4 kickstart installation. The problem is related to SATA module loading sequence. I have a HP workstation XW4200 machine. two SATA disks are connected to motherboard's SATA ports directly, and be detected to /dev/sda, /dev/sdb by former FC2. two other hard disks are connected through Promise SATA cards and detected as /dev/sdc and /dev/sdd. Now when I tried to upgrade it to FC4, it detects the hard drives in different order. FC4 loads sata_promise module first and then ata_piix module. so that former /dev/sd{c,d} are detected as /dev/sd{a,b}, and vice versa. So now I can not do upgrade to FC4 while preserve orignal data on hard disks. If there is a way for me to specify SATA module loading order during unattended kickstart and post-install? Or I have to recompile a new kernel to static-build ata_piix into kernel, and compile sata_promise as a loadable module? MANY thanks. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From hahaha_30k at yahoo.com Sat Jun 25 02:04:31 2005 From: hahaha_30k at yahoo.com (ha haha) Date: Fri, 24 Jun 2005 19:04:31 -0700 (PDT) Subject: kickstart waits when machines have two network cards In-Reply-To: <021001c57928$3bb33460$3301a8c0@saporo> Message-ID: <20050625020431.77910.qmail@web30202.mail.mud.yahoo.com> Joe, Sharon and Tim Moloney, Thanks a lot!!! Now the problem goes a step further. It's related to modules load sequence for network cards. I have a machine with two networks cards, originaly the built-in Broadcom network card is detected as eth0, and external Intel gigabit as eth1, by former FC2. And the installs only happens on eth0, the built-in Broadcom NIC. Now the FC4 detects network cards in reverse order, so former eth0(built-in Broadcom) is detected as eth1, while the former eth1 is assigned as eth0. Is there configuration file I can change (/etc/modprobe.conf doesn't help) to alter the network card modules' load sequence, both during kickstart and post-install? Thanks....... The --- Sharon and Tim Moloney wrote: > "ha haha" wrote: > > > I try to begin unattended kickstart installation > for > > FC4 onto my machine but it fails, because > > unfortunately the machine has e two network cards > and > > so the kickstart hangs for manual input, to select > > which network card to continue. > > > > Is there a way to specify a network card > explicitly > > to grasp ks.cfg for NFS installation? > > On the boot command line, you can add > "ksdevice=eth0" to specify the first > NIC or "ksdevice=link" to specify the one that has > an active link status > (assuming only one has an active link status). > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > ____________________________________________________ Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football http://football.fantasysports.yahoo.com From st.moloney at gte.net Sat Jun 25 02:12:41 2005 From: st.moloney at gte.net (Sharon and Tim Moloney) Date: Fri, 24 Jun 2005 22:12:41 -0400 Subject: kickstart waits when machines have two network cards References: <20050625020431.77910.qmail@web30202.mail.mud.yahoo.com> Message-ID: <023501c5792b$5edf6320$3301a8c0@saporo> "ha haha" wrote: > Now the problem goes a step further. It's related to > modules load sequence for network cards. > > I have a machine with two networks cards, originaly > the built-in Broadcom network card is detected as > eth0, and external Intel gigabit as eth1, by former > FC2. > > And the installs only happens on eth0, the built-in > Broadcom NIC. > > Now the FC4 detects network cards in reverse order, > so former eth0(built-in Broadcom) is detected as eth1, > while the former eth1 is assigned as eth0. > > Is there configuration file I can change > (/etc/modprobe.conf doesn't help) to alter the network > card modules' load sequence, both during kickstart and > post-install? I think (but I'm not positive) that you can specify ethernet device order (or any device?) by passing arguments on the command line, like "eth0=3fe,10" where 3fe is the I/O port and 10 is the interrupt. Check the kernel documentation to verify and for details. It's been a while since I've used it and I don't have a Linux box available now to check it for you. From phr at doc.ic.ac.uk Sat Jun 25 06:52:34 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Sat, 25 Jun 2005 07:52:34 +0100 (BST) Subject: Dropping to a shell in %post In-Reply-To: <42BC99BD.9030108@hostinthebox.net> References: <42BC99BD.9030108@hostinthebox.net> Message-ID: On Fri, 24 Jun 2005, Dan Trainor wrote: >Is it possible to drop to a sh shell in or after %post, so that I can >poke around and get familiar with the %post enviornment to see about >donig some things later on? One method is to "sleep 3600" in the script and use the anaconda-provided shell on VT2 (press Alt+F2). When done debugging, kill the sleep process - repeat as necessary. To see things exactly as %post (implied chroot) sees them, "chroot /mnt/sysimage". Cheers, Phil From regatta at gmail.com Mon Jun 27 05:20:24 2005 From: regatta at gmail.com (regatta) Date: Mon, 27 Jun 2005 08:20:24 +0300 Subject: Reinstallaing a live PC Message-ID: <5a3ed56505062622206bec7aa4@mail.gmail.com> Hi Not sure if this the right maillist but anyway , I want to reinstall my machines remotely using kickstart or askmethod option. can I add askmethod or kickstart path in the GRUB so when I reboot and select the kernel with askmethod the re installation will start ? Thanks -- Best Regards, -------------------- -*- If Linux doesn't have the solution, you have the wrong problem -*- From clumens at redhat.com Mon Jun 27 17:05:17 2005 From: clumens at redhat.com (Chris Lumens) Date: Mon, 27 Jun 2005 13:05:17 -0400 (EDT) Subject: FC4 kickstart broken? In-Reply-To: <42BAFB6B.4080809@berkeley.edu> References: <42BAFB6B.4080809@berkeley.edu> Message-ID: > If you comment out all the package removals, it "works." By works, i mean it > seems to install all the packages, but the resulting anaconda-ks.cfg file in > /root does not have anything meaningful in the %package section. It almost > seems like package selection in the kickstart files is not supported anymore? > I have not found a similar bug on bugzilla which makes me wonder if something > changed and I'm not aware? if not I will file a bugzilla bug. Doing a > graphical install of FC4 on all our machines would take forever! See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160209. The quick explanation is that we added support to exclude by architecture at the last minute and broke the normal case. A workaround would be to append ".i386" or whatever to the names of the packages you want to exclude. - Chris From brandono at berkeley.edu Mon Jun 27 17:10:30 2005 From: brandono at berkeley.edu (Brandon Ooi) Date: Mon, 27 Jun 2005 10:10:30 -0700 Subject: FC4 kickstart broken? In-Reply-To: References: <42BAFB6B.4080809@berkeley.edu> Message-ID: <42C03306.5070506@berkeley.edu> Thanks! That will save me from having to do a bunch of rpm -e's. Does this mean that if we were to do an x86_64 install, we append .x86_64? Thanks again! Brandon Chris Lumens wrote: >> If you comment out all the package removals, it "works." By works, i >> mean it seems to install all the packages, but the resulting >> anaconda-ks.cfg file in /root does not have anything meaningful in >> the %package section. It almost seems like package selection in the >> kickstart files is not supported anymore? I have not found a similar >> bug on bugzilla which makes me wonder if something changed and I'm >> not aware? if not I will file a bugzilla bug. Doing a graphical >> install of FC4 on all our machines would take forever! > > > See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160209. The > quick explanation is that we added support to exclude by architecture at > the last minute and broke the normal case. A workaround would be to > append ".i386" or whatever to the names of the packages you want to > exclude. > > - Chris > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From extern.Tobias.Kronwitter at AUDI.DE Tue Jun 28 11:21:52 2005 From: extern.Tobias.Kronwitter at AUDI.DE (extern.Tobias.Kronwitter at AUDI.DE) Date: Tue, 28 Jun 2005 13:21:52 +0200 Subject: kickstart seems not to process: ks.cfg Message-ID: Hello list I'm trying to install RedHat Enterprise LINUX V 3.0 over the network. I have no keyboard, mouse or monitor connected to the server. Instead I'm using the serial port as a console. The server itself is a SUN VX60. The following steps are performed: 1. dhcp / pxeboot 2. load ks.cfg via nfs 3. install OS via nfs Step 1 works fine. However there are complaints about missing parameters: +--------------------------------------------------------------------------- CLIENT MAC ADDR: 00 0E 0C 30 B5 E0 GUID: DDFD185E 9FB8 11D8 B9DC 00108365A7E7 CLIENT IP: 10.199.131.149 MASK: 255.255.255.224 DHCP IP: 10.199.135.107 GATEWAY IP: 10.199.131.130 10.199.131.129 PXELINUX 3.09 2005-06-17 Copyright (C) 1994-2005 H. Peter Anvin UNDI data segment at: 00090D40 UNDI data segment size: B450 UNDI code segment at: 0009C190 UNDI code segment size: 30E0 PXE entry point found (we hope) at 9C19:0106 My IP address seems to be 0AC78395 10.199.131.149 ip=10.199.131.149:10.199.131.150:10.199.131.129:255.255.255.224 TFTP prefix: /el-3.0/sun/ Trying to load: pxelinux.cfg/01-00-0e-0c-30-b5-e0 Trying to load: pxelinux.cfg/0AC78395 Missing parameter in config file. Missing parameter in config file. Missing parameter in config file. 0AC78395: PXE installing Enterprise Linux 3.0 in 10 seconds boot: +--------------------------------------------------------------------------- This is the corresponding pxeboot - config file (0AC78395): serial 1 9600 default do_inst display pxelinux.cfg/bootinfo_0AC78395.txt prompt 1 timeout 100 label do_inst kernel ../images/pxeboot/vmlinuz append ksdevice=eth1 console=ttyS1,9600n8 load_ramdisk=1 initrd=install/initrd_rdh-es-v3.img ks=nfs:10.199.131.150:/tftpboot/ks_0AC78395.cfg network # After 10 seconds, the machine boots vmlinuz via network, loads the adaptec driver: +--------------------------------------------------------------------------- Welcome to Red Hat Enterprise Linux +--------+ Loading SCSI driver +---------+ | | | Loading aic79xx driver... | | | +----------------------------------------+ / between elements | selects | next screen +--------------------------------------------------------------------------- loads and configures ethernet: +--------------------------------------------------------------------------- Welcome to Red Hat Enterprise Linux +--------------------+ Dynamic IP +---------------------+ | | | Sending request for IP information for eth1... | | | +-------------------------------------------------------+ / between elements | selects | next screen +--------------------------------------------------------------------------- and finally ends in the interactive installation, instead of processing ks.cfg The file: ks_0AC78395.cfg looks like: # @(#)ks_0AC78395.cfg 1.0 # # T. K - 050610 # # # install lang en_US.UTF-8 langsupport --default en_US.UTF-8 en_US.UTF-8 keyboard de mouse none text skipx network --device eth1 --bootproto static --ip 10.199.131.149 --netmask 255.255.255.224 --gateway 10.199.131.129 --nameserver 10.199.254.15 --hostname iuts003 rootpw --iscrypted $1$o/hsDB3v$mQ8BZ7doWf4R9549EtRvH/ firewall --disabled authconfig --enableshadow --enablemd5 timezone --utc Europe/Berlin # bootloader --location=mbr --append console=ttyS1,9600n8 zerombr yes clearpart --all --initlabel # #part /boot --asprimary --fstype ext3 --size 50 --ondisk sda #part / --asprimary --fstype ext3 --size 8192 --ondisk sda #part /vol1 --asprimary --fstype ext3 --size 2500 --grow --ondisk sda #part swap --recommended --ondisk sda # part raid.01 --size=2048 --ondisk=sda part raid.02 --size=2048 --ondisk=sdb # part raid.11 --size=50 --ondisk=sda part raid.12 --size=50 --ondisk=sdb # part raid.21 --size=8192 --ondisk=sda part raid.22 --size=8192 --ondisk=sdb # part raid.31 --size=1 --grow --ondisk=sda part raid.32 --size=1 --grow --ondisk=sdb # raid swap --level=1 --device=md0 raid.01 raid.02 raid /boot --level=1 --device=md1 raid.11 raid.12 raid / --level=1 --device=md2 raid.21 raid.22 raid /vol1 --level=1 --device=md3 raid.31 raid.32 # nfs --server 10.199.135.107 --dir /shark/install/RedHat_ES_v3 # %packages @ Network Servers # Other package groups may be installed by uncommenting one or more of the # following. If no packages are listed, the installation will prompt for # which packages to install #@ Everything @ Core @ Base ## @ KDE Desktop Environment @ Graphical Internet @ Text-based Internet ## @ Sound and Video ## @ Graphics ## @ Office/Productivity ## @ Mail Server @ Network Servers ## @ Legacy Network Server ## @ News Server ## @ Windows File Server @ Server Configuration Tools @ FTP Server ## @ SQL Database ## @ MySQL Database @ Web Server ## @ DNS Name Server ## @ Authoring and Publishing ## @ Engineering and Scientific @ Editors ## @ Emacs ## @ XEmacs ## @ Ruby @ System Tools @ Administration Tools ## @ Games and Entertainment ## @ ISO8859-2 Support ## @ ISO8859-9 Support ## @ ISO8859-14 Support @ ISO8859-15 Support @ Development Tools @ Development Libraries ## @ Kernel Development ## @ Legacy Software Development ## @ Compatibility Arch Support ## @ Compatibility Arch Development Support ## @ Legacy Software Support ## @ X Software Development ## @ GNOME Software Development ## @ KDE Software Development # kernel-smp kernel grub # ### Commands to be run immediately after this file has been parsed %pre echo "Kickstart-installed Red Hat Enterprise Linux `/bin/date`" ### Commands to be run post-installation %post echo "Kickstart installation" > /tmp/message # Any ideas ? Help is greatly appreciated. Thank you very much Tobias From error27 at gmail.com Tue Jun 28 13:23:03 2005 From: error27 at gmail.com (Dan Carpenter) Date: Tue, 28 Jun 2005 15:23:03 +0200 Subject: kickstart seems not to process: ks.cfg In-Reply-To: References: Message-ID: On 6/28/05, extern.Tobias.Kronwitter at audi.de wrote: > However there are complaints about missing parameters: That could result in the wrong stuff being passed to the install kernel. Check by doing an install and typing `cat /proc/cmdline` on tty3 (or something like that since you don't have a keyboard). Also check that the nfs directory is accessible from tty3. mkdir /mnt/nfs mount 10.199.131.150:/tftpboot/ /mnt/nfs less /mnt/nfs/ks_0AC78395.cfg regards, dan carpenter From phr at doc.ic.ac.uk Tue Jun 28 14:06:15 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 28 Jun 2005 15:06:15 +0100 (BST) Subject: kickstart seems not to process: ks.cfg In-Reply-To: References: Message-ID: On Tue, 28 Jun 2005 extern.Tobias.Kronwitter at AUDI.DE wrote: >The following steps are performed: > >1. dhcp / pxeboot >2. load ks.cfg via nfs >3. install OS via nfs > >Step 1 works fine. >However there are complaints about missing parameters: [snip] >PXE entry point found (we hope) at 9C19:0106 > >My IP address seems to be 0AC78395 10.199.131.149 >ip=10.199.131.149:10.199.131.150:10.199.131.129:255.255.255.224 > >TFTP prefix: /el-3.0/sun/ >Trying to load: pxelinux.cfg/01-00-0e-0c-30-b5-e0 >Trying to load: pxelinux.cfg/0AC78395 >Missing parameter in config file. >Missing parameter in config file. >Missing parameter in config file. >0AC78395: PXE installing Enterprise Linux 3.0 in 10 seconds >boot: >+--------------------------------------------------------------------------- > >This is the corresponding pxeboot - config file (0AC78395): > >serial 1 9600 >default do_inst >display pxelinux.cfg/bootinfo_0AC78395.txt >prompt 1 >timeout 100 > >label do_inst > kernel ../images/pxeboot/vmlinuz > append ksdevice=eth1 console=ttyS1,9600n8 load_ramdisk=1 >initrd=install/initrd_rdh-es-v3.img >ks=nfs:10.199.131.150:/tftpboot/ks_0AC78395.cfg network ># Is that "append" line actually wrapped? If so, try un-wrapping. Those "Missing parameter" errors mean that your pxelinux.cfg files is probably not being parsed as intended. Do relative paths works for the "kernel" arg? (Remember these files are being retreived by PXElinux via TFTP.) Cheers, Phil From gotosathya at gmail.com Tue Jun 28 15:25:09 2005 From: gotosathya at gmail.com (Sathya M) Date: Tue, 28 Jun 2005 20:55:09 +0530 Subject: Regd CDDialog.py In-Reply-To: References: Message-ID: Hello All, I have one doubt regarding how the CDDialog.py will be created after the installation.Because if i have to upgrade my packages then it will display the messeges like "insert Disk 1 to install" from this CDDialog.py.I have to change the display.If you guys know this then help me. Regrads, sathy From info at hostinthebox.net Wed Jun 29 00:24:21 2005 From: info at hostinthebox.net (Dan Trainor) Date: Tue, 28 Jun 2005 17:24:21 -0700 Subject: Dropping to a shell in %post In-Reply-To: <50706.69.33.32.187.1119663798.squirrel@webmail.half-asleep.com> References: <42BC99BD.9030108@hostinthebox.net> <42BCA4F9.6030805@hostinthebox.net> <50706.69.33.32.187.1119663798.squirrel@webmail.half-asleep.com> Message-ID: <42C1EA35.1040203@hostinthebox.net> Daniel Segall wrote: > Are you actually specifying that full path? If so, that would be your > problem. You are already chroot'd in /mnt/sysimage, so if you told it > /mnt/sysimage/somepath, it would be interpreted literally. > > -Dan > > >>Dan Carpenter wrote: >> >>>On 6/24/05, Dan Trainor wrote: >>> >>> >>> >>>>Is it possible to drop to a sh shell in or after %post, so that I can >>>>poke around and get familiar with the %post enviornment to see about >>>>donig some things later on? >>> >>> >>>add these lines to your %post. >>> >>>echo "Dropping to a shell" >>>open -c 8 bash >>>read tmp >>> >>>The echo prints to tty3 so you probably won't see it... Not sure how >>>to put stuff on the first tty. This will pause the install until you >>>press Enter on the front screen. You can switch to tty8 to do >>>whatever you want. >>> >>>regards, >>>dan carpenter >>> >> >>Dan - >> >>Awesome, that's what I was looking for. I'm having troule finding out >>why I can't install two additional RPMs in %post, saying that the files >>are not present. I am copying them from >>/mnt/source/thedirthatiputfilesin to /mnt/sysimage/tmp/mynewdir, so I'd >>like to troubleshoot that. >> >>Thanks >>-dant >> Hey, dan - This past post has taught be a little bit more about the env. How might I go about finding out where my cdrom drive is mounted (is it even still mounted in %post?) inside of the chroot? Thanks! -dant From info at hostinthebox.net Wed Jun 29 00:33:46 2005 From: info at hostinthebox.net (Dan Trainor) Date: Tue, 28 Jun 2005 17:33:46 -0700 Subject: /proc/cmdline Message-ID: <42C1EC6A.603@hostinthebox.net> Hello, all - What's the weapon of choice when extracting values from /proc/cmdline? grep, cut, awk, sed? perl? I don't know. I'm just trying to get a better idea of how I can use cmdline. Thanks! -dant From klaus.steden at thomson.net Wed Jun 29 00:40:38 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 28 Jun 2005 20:40:38 -0400 Subject: /proc/cmdline In-Reply-To: <42C1EC6A.603@hostinthebox.net> References: <42C1EC6A.603@hostinthebox.net> Message-ID: <20050629004037.GZ2476@thomson.net> > Hello, all - > > What's the weapon of choice when extracting values from /proc/cmdline? > grep, cut, awk, sed? perl? I don't know. I'm just trying to get a > better idea of how I can use cmdline. > Dan, I typically use grep to test for the presence of what I'm looking to match, and then sed to extract it. Options in /proc/cmdline are usually of the form or =. So, for instance ... -- cut -- if grep -i -q "class=[a-zA-Z0-9]" /proc/cmdline then CLASS=`cat /proc/cmdline | sed 's/.*class=\([^ ]*\).*/\1/'` fi -- cut -- hth, Klaus From info at hostinthebox.net Wed Jun 29 00:44:41 2005 From: info at hostinthebox.net (Dan Trainor) Date: Tue, 28 Jun 2005 17:44:41 -0700 Subject: /proc/cmdline In-Reply-To: <20050629004037.GZ2476@thomson.net> References: <42C1EC6A.603@hostinthebox.net> <20050629004037.GZ2476@thomson.net> Message-ID: <42C1EEF9.2050903@hostinthebox.net> Klaus Steden wrote: >>Hello, all - >> >>What's the weapon of choice when extracting values from /proc/cmdline? >>grep, cut, awk, sed? perl? I don't know. I'm just trying to get a >>better idea of how I can use cmdline. >> > > Dan, > > I typically use grep to test for the presence of what I'm looking to match, > and then sed to extract it. Options in /proc/cmdline are usually of the form > or =. So, for instance ... > > -- cut -- > if grep -i -q "class=[a-zA-Z0-9]" /proc/cmdline > then > CLASS=`cat /proc/cmdline | sed 's/.*class=\([^ ]*\).*/\1/'` > fi > -- cut -- > > hth, > Klaus > > Klaus - It does help, thanks. I just want to make sure that I am using cmdline to it's full potential. Thanks -dant From info at hostinthebox.net Wed Jun 29 00:51:34 2005 From: info at hostinthebox.net (Dan Trainor) Date: Tue, 28 Jun 2005 17:51:34 -0700 Subject: Dropping to a shell in %post In-Reply-To: <42C1EA35.1040203@hostinthebox.net> References: <42BC99BD.9030108@hostinthebox.net> <42BCA4F9.6030805@hostinthebox.net> <50706.69.33.32.187.1119663798.squirrel@webmail.half-asleep.com> <42C1EA35.1040203@hostinthebox.net> Message-ID: <42C1F096.9010408@hostinthebox.net> Dan Trainor wrote: > Daniel Segall wrote: > >>Are you actually specifying that full path? If so, that would be your >>problem. You are already chroot'd in /mnt/sysimage, so if you told it >>/mnt/sysimage/somepath, it would be interpreted literally. >> >>-Dan >> >> >> >>>Dan Carpenter wrote: >>> >>> >>>>On 6/24/05, Dan Trainor wrote: >>>> >>>> >>>> >>>> >>>>>Is it possible to drop to a sh shell in or after %post, so that I can >>>>>poke around and get familiar with the %post enviornment to see about >>>>>donig some things later on? >>>> >>>> >>>>add these lines to your %post. >>>> >>>>echo "Dropping to a shell" >>>>open -c 8 bash >>>>read tmp >>>> >>>>The echo prints to tty3 so you probably won't see it... Not sure how >>>>to put stuff on the first tty. This will pause the install until you >>>>press Enter on the front screen. You can switch to tty8 to do >>>>whatever you want. >>>> >>>>regards, >>>>dan carpenter >>>> >>> >>>Dan - >>> >>>Awesome, that's what I was looking for. I'm having troule finding out >>>why I can't install two additional RPMs in %post, saying that the files >>>are not present. I am copying them from >>>/mnt/source/thedirthatiputfilesin to /mnt/sysimage/tmp/mynewdir, so I'd >>>like to troubleshoot that. >>> >>>Thanks >>>-dant >>> > > > Hey, dan - > > This past post has taught be a little bit more about the env. How might > I go about finding out where my cdrom drive is mounted (is it even still > mounted in %post?) inside of the chroot? > > Thanks! > -dant > > Hi - I think I can use /tmp/cdrom as the block device to mount, in an attempt to copy additional files off of the CD, into the env setup that I use from tty2. What I'd like to do, in %post, is install a bunch of additonal RPMs off of the install CD. However, I get a whole lot of dependency errors. I believe this is caused iether because nothing is in the library search path, an RPM database is not yet constructed, or because this is just one stupid idea in itself and it will never work and I'm just wasting my time. Can someone shed some light on this subject? I would be very greatful. Thanks! -dant From klaus.steden at thomson.net Wed Jun 29 02:31:13 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 28 Jun 2005 22:31:13 -0400 Subject: forcing RPM install order In-Reply-To: <42C1EC6A.603@hostinthebox.net> References: <42C1EC6A.603@hostinthebox.net> Message-ID: <20050629023113.GA2476@thomson.net> Are there ways to force RPMs to install in a particular order? I've got a really messy linkage situation with 32 bits apps on a 64 bit system, and of the four possible permutations (A, B, AB, or BA) - only one of them actually works with everything. Anyone got any advice on the matter? I've specified an order in my ks.cfg but anaconda seems to disagree with me. I suppose I could specify package B in the %post section but that's a little clumsy. thanks, Klaus From chandan-dutta.chowdhury at hp.com Wed Jun 29 08:59:12 2005 From: chandan-dutta.chowdhury at hp.com (Chowdhury, Chandan Dutta) Date: Wed, 29 Jun 2005 14:29:12 +0530 Subject: Console based kickstart file generator Message-ID: Hello all, Is there a console based tool for generating the kickstart file? Currently the tool provided by redhat works only with X. Is there a similar tool, which I can use on the console? Regards -CDC From phr at doc.ic.ac.uk Wed Jun 29 12:17:10 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 29 Jun 2005 13:17:10 +0100 (BST) Subject: forcing RPM install order In-Reply-To: <20050629023113.GA2476@thomson.net> References: <42C1EC6A.603@hostinthebox.net> <20050629023113.GA2476@thomson.net> Message-ID: On Tue, 28 Jun 2005, Klaus Steden wrote: >Are there ways to force RPMs to install in a particular order? > >I've got a really messy linkage situation with 32 bits apps on a 64 bit >system, and of the four possible permutations (A, B, AB, or BA) - only one of >them actually works with everything. Fix the dependencies, and rpm will do it for you. This assumes you built the packages in the first place; if they're 3rd-party binary RPMs with horrendous %postinstall scripts, then it'll be easier to use a kickstart %post. Cheers, Phil From clumens at redhat.com Wed Jun 29 14:08:00 2005 From: clumens at redhat.com (Chris Lumens) Date: Wed, 29 Jun 2005 10:08:00 -0400 (EDT) Subject: langsupport heads up Message-ID: As of anaconda-10.3.0.3-1, the langsupport keyword is no longer supported in kickstart. Instead of using this keyword, you should now be using the appropriate packages in the %packages section. So before where you had this: langsupport de fr it You should now have this: %packages @german-support @french-support @italian-support You'll get an exception with a helpful error message if you use kickstart from Rawhide and don't update things. Good luck. - Chris From klaus.steden at thomson.net Wed Jun 29 15:33:53 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 29 Jun 2005 11:33:53 -0400 Subject: Console based kickstart file generator In-Reply-To: References: Message-ID: <20050629153352.GB2476@thomson.net> > Hello all, > > > Is there a console based tool for generating the kickstart file? > Currently the tool provided by redhat works only with X. Is there a > similar tool, which I can use on the console? > vi. ;> I wrote one in Perl, based on templates, that I call from within kickstart to do my bidding. Klaus From info at hostinthebox.net Wed Jun 29 23:22:13 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 29 Jun 2005 16:22:13 -0700 Subject: Edited comps.xml - now what? In-Reply-To: <42BB417C.60200@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> <42BA2EA0.6050108@hostinthebox.net> <42BB417C.60200@hostinthebox.net> Message-ID: <42C32D25.2060006@hostinthebox.net> Dan Trainor wrote: > Dan Trainor wrote: > >>Klaus Steden wrote: >> >> >>>>>Someone posted a link to a python script that fairly neatly encapsulates >>>>>all the necessary steps (dependency checks, package ordering, and >>>>>genhdlist) ... I don't have the link handy - but I've been using it for >>>>>the last three weeks (with a few tweaks) and it's working for both 32 and >>>>>64-bit versions of RHEL4 for me. >>>>> >>>>>If you want a copy I can send you one. >>>>> >>>>>hth, >>>>>Klaus >>>>> >>>> >>>>Thanks for responding, Klaus - >>>> >>>>Mind just telling me the name of the script? Perhaps I can google for >>>>it and get some more information on it, such as some documentation. >>>> >>>>I'd honestly like to fully understand the entire process, rather than >>>>just use a script that does it all for me - however, the idea of an >>>>all-in-one tool does sound very appealing. >>>> >>> >>>It was called 'do_update'. >>> >>>http://www.math.uh.edu/~tibbs/do_update (thanks to my Firefox history!). >>> >>>Seems pretty straightforward - I haven't learned python yet, but it wasn't too >>>daunting. >>> >>>Klaus >>> >> >> >>Klaus - >> >>Looks to be pretty neat. Seems a bit out of date, but I think the >>concept remains the same. >> >>However, I'm still looking for some documentation that clearly explains >>how the procedure is done - what to do after a comps.xml file is edited, >>how to create the hdlist files, etc etc. I guess I'll keep searching >>here. ANy feeback by others would also be greatly appreciated. >> >>Again, thanks for the time! >>-dant >> > > > Forgive me for letting these emails get out-of-sync. I downloaded them > to my laptop at home, and have no way to retrieve them - but this was a > response, retreived rfom the kickstart-list archives, by Phil: > > ...snip... > > On Wed, 22 Jun 2005, Dan Trainor wrote: > > >>However, I'm still looking for some documentation that clearly explains >>how the procedure is done - what to do after a comps.xml file is >>edited, how to create the hdlist files, etc etc. > > > genhdlist, also with (depending whether the anaconda bug still exists) > pkgorder. The exact incantations appear in the list archives. > > ...snip... > > Phil - > > I think you're correct, that genhdlist is my answer. I still don't > understand exactly how it works. And the exact incantations (nice word, > btw - very fitting) do not appear in the list archives. I think I found > a total of ~20 emails that had any reference to the string 'genhdlist'. > I'm a bit lost for information. The best I've found is a few examples > of people who are trying to modify 'base/hdlist{2}'. > > So I run it, and I get: > > genhdlist: genhdlist [--withnumbers] [--fileorder ] > [--hdlist ] [--productpath ] + > > What does --with-numbers do? > What does --fileorder do? > What is --hdlist referencing? > What is a --productpath? > Which additional paths is it asking for? > > ...and we're just getting started. I'd like to become familiar with all > the tools that come with anaconda-runtime-10.0-5, because they *look* to > be pretty important and useful. I just can't find any documentation > anywhere. > > I've been telling myself the entire time: "Hold back, don't ask these > guys, just RTFM!" - then I think to myself, "Where is TFM?". Is there > any formal documentation on the anaconda-runtime suite of utilities? > > Again, I greatly appreciate the time. > > Thanks > -dant > Hey, all - Are these questions legit, or am I just being crazy here? ;) Anyone have any documentation on genhdlist? Can I expect my RPMs to "install" normally when installing from Anaconda, or do the RPMs have to be "special"? Frustrated, but appreciating all your time - thanks. -dant From phr at doc.ic.ac.uk Thu Jun 30 00:33:34 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 30 Jun 2005 01:33:34 +0100 (BST) Subject: Edited comps.xml - now what? In-Reply-To: <42C32D25.2060006@hostinthebox.net> References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> <42BA2EA0.6050108@hostinthebox.net> <42BB417C.60200@hostinthebox.net> <42C32D25.2060006@hostinthebox.net> Message-ID: On Wed, 29 Jun 2005, Dan Trainor wrote: >Are these questions legit, or am I just being crazy here? ;) > >Anyone have any documentation on genhdlist? Can I expect my RPMs to >"install" normally when installing from Anaconda, or do the RPMs have to >be "special"? No special requirements, but also precious little "official" documentation for anaconda-utils. If the mailing list archives don't cover it, it's not worth knowing... in a pinch the sourcecode is always available. Cheers, Phil From info at hostinthebox.net Thu Jun 30 00:37:33 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 29 Jun 2005 17:37:33 -0700 Subject: Edited comps.xml - now what? In-Reply-To: References: <42B9E7C8.9090603@hostinthebox.net> <20050622224135.GW2476@thomson.net> <42B9EAB0.2030807@hostinthebox.net> <20050622225412.GY2476@thomson.net> <42BA2EA0.6050108@hostinthebox.net> <42BB417C.60200@hostinthebox.net> <42C32D25.2060006@hostinthebox.net> Message-ID: <42C33ECD.6000707@hostinthebox.net> Philip Rowlands wrote: > On Wed, 29 Jun 2005, Dan Trainor wrote: > > >>Are these questions legit, or am I just being crazy here? ;) >> >>Anyone have any documentation on genhdlist? Can I expect my RPMs to >>"install" normally when installing from Anaconda, or do the RPMs have to >>be "special"? > > > No special requirements, but also precious little "official" > documentation for anaconda-utils. If the mailing list archives don't > cover it, it's not worth knowing... in a pinch the sourcecode is always > available. > > > Cheers, > Phil > Phil - Very good point, indeed. After all, what good is Open Source without the source? Unfortunately, I do not agree with you about if it's not covered, it's not worth knowing. I'm sure Anaconda has plenty of potential, and many others would immediate benefit from these little details. However, not all of us are gifted enough to have an intimate understanding of major and minor languages at the drop of the hat. By trade, I am by no means a programmer - however, I can generally figure something out. Thanks! -dant From chandan-dutta.chowdhury at hp.com Thu Jun 30 04:45:42 2005 From: chandan-dutta.chowdhury at hp.com (Chowdhury, Chandan Dutta) Date: Thu, 30 Jun 2005 10:15:42 +0530 Subject: Console based kickstart file generator Message-ID: Hello Klaus, Can you share the same with me ? Where can I find thepackage. Thanks in advance for all the help Regards -CDC -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Klaus Steden Sent: Wednesday, June 29, 2005 9:04 PM To: Discussion list about Kickstart Subject: Re: Console based kickstart file generator > Hello all, > > > Is there a console based tool for generating the kickstart file? > Currently the tool provided by redhat works only with X. Is there a > similar tool, which I can use on the console? > vi. ;> I wrote one in Perl, based on templates, that I call from within kickstart to do my bidding. Klaus _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From MWeiner at ag.com Thu Jun 30 13:49:17 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Thu, 30 Jun 2005 09:49:17 -0400 Subject: Kickstarting FC2 installs Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B007037354@orca.agcom.amgreetings.com> I have been down this road before, but am running into some issues. OK, what I am trying to do is create a "boot disk" such that someone in our data center can simply stick the disk in the floppy, reboot the server and it will boot off of that, and begin the kickstart install. So to that end, I have the following in my syslinux.cfg: default linux prompt 1 timeout 600 display snake.msg F1 boot.msg F2 general.msg F3 expert.msg F4 param.msg F5 rescue.msg F7 snake.msg DEFAULT linux LABEL linux KERNEL vmlinuz APPEND ksdevice=eth0 console=tty0 load_ramdisk=1 prompt_ramdisk=0 \ initrd=initrd.img network ip=dhcp ks=nfs:10.10.232.54:/tftpboot/ks.cfg \ selinux=0 However, when it boots off of the diskette, I see the following on VC3: Bootp: no boot file specified Ks server: 10.10.232.54:/tftpboot file: ks-fc2.cfg ... kickstarting through device eth0 ... Mounting nfs path 10.10.232.54:/repo/fedora/linux/core/2/i386/os No valid tree or isos found in 10.10.232.54:/repo/fedora/linux/core/2/i386/os And the install basically aborts. Now, to test this, I ran the same thing only via PXE boot, and that works perfectly mounting the SAME server and location, etc same ks file. Is it possible to do what I am trying (eg create a ks bootdisk and do an nfs install of fc2)? Everything I have read leads me to believe it can. So my question is, why am I seeing two different behaviors? The only REAL difference is via PXE I load the vmlinuz and initrd.img from the FC2/images/pxeboot tree rather than using say bootimg and dd to create a boot disk. Would this cause this type of behavior? I tried to get those two files onto my boot disk but alas they are too large for a 1.44M floppy. Any ideas or pointers?Any help would be greatly appreciated. Thanks in advance Michael Weiner -- Update, I changed the syslinux.cfg to say ks=floppy since I have the ks.cfg on the local boot floppy. And that seems to get around that issue. Now, it goes to NFS mount 10.10.232.54:/repo/fedora/linux/core/2/i386/os and it fails to mount and basically bombs. BTW, I have all this working via PXE, the issue is that I would have to have the person in the data center redo the BIOS' across all 450 odd machines to get the PXE boot to work, and a boot disk seemed to be less work to me, simply stick it in, rather than changing default BIOS settings. Any thoughts? Michael Weiner ... Says it Best. Michael Weiner, Linux+, Linux+ SME Senior Systems Administrator AmericanGreetings.com One American Road Cleveland, OH 44144 MWeiner at ag.com IM: hUnTeRoZe http://www.interactive.ag.com tel: fax: mobile: (216) 889-5028 (216) 889-5029 (216) 965-8619 Add me to your address book... Want a signature like this? From klaus.steden at thomson.net Thu Jun 30 14:09:46 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 30 Jun 2005 10:09:46 -0400 Subject: Console based kickstart file generator In-Reply-To: References: Message-ID: <20050630140945.GA842@thomson.net> > Hello Klaus, > > Can you share the same with me ? Where can I find thepackage. > > Thanks in advance for all the help > Sorry, no such luck, I put this together by hand. The approach I've taken was to inject code in between the /sbin/loader starting (the kickstart init program) and anaconda being launched - something I wouldn't advise unless you're brave and possibly foolish. Realistically, though, the kickstart.cfg is just plain text - I grabbed the docs, and a copy of a ks file I had used on another machine, and I used that as my reference point for generating things. Make one using the GUI, and start experimenting with it, would be my best advice. hth, Klaus From phr at doc.ic.ac.uk Thu Jun 30 14:27:47 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 30 Jun 2005 15:27:47 +0100 (BST) Subject: Console based kickstart file generator In-Reply-To: <20050630140945.GA842@thomson.net> References: <20050630140945.GA842@thomson.net> Message-ID: On Thu, 30 Jun 2005, Klaus Steden wrote: >Realistically, though, the kickstart.cfg is just plain text - I grabbed the >docs, and a copy of a ks file I had used on another machine, and I used that >as my reference point for generating things. Make one using the GUI, and start >experimenting with it, would be my best advice. Another approach to avoids having to write kickstart files from scratch would be to build one machine by hand, then check /root/anaconda-ks.cfg which is left behind. Cheers, Phil From Christian.Rohrmeier at SCHERING.DE Thu Jun 30 14:39:15 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 30 Jun 2005 16:39:15 +0200 Subject: Console based kickstart file generator In-Reply-To: <20050630140945.GA842@thomson.net> Message-ID: Hi, I did something similar. I put tags in a template ks.cfg that I then replace dynamically with the apropriate text via a perl script. The original starting point was a standard kickstart script. The documentation for KS is actually quite good, other than a few omissions. Most important for me was to see others' KS files to get ideas, and then ofcourse to spend quality time writing and testing my own to learn what can and can't be done with KS. Cheers, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Klaus Steden To Sent by: Discussion list about Kickstart kickstart-list-bo unces at redhat.com cc Subject 30.06.2005 16:09 Re: Console based kickstart file generator Please respond to Discussion list about Kickstart > Hello Klaus, > > Can you share the same with me ? Where can I find thepackage. > > Thanks in advance for all the help > Sorry, no such luck, I put this together by hand. The approach I've taken was to inject code in between the /sbin/loader starting (the kickstart init program) and anaconda being launched - something I wouldn't advise unless you're brave and possibly foolish. Realistically, though, the kickstart.cfg is just plain text - I grabbed the docs, and a copy of a ks file I had used on another machine, and I used that as my reference point for generating things. Make one using the GUI, and start experimenting with it, would be my best advice. hth, Klaus _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From tdiehl at rogueind.com Thu Jun 30 19:53:28 2005 From: tdiehl at rogueind.com (Tom Diehl) Date: Thu, 30 Jun 2005 15:53:28 -0400 (EDT) Subject: FC4 kickstart broken? In-Reply-To: References: <42BAFB6B.4080809@berkeley.edu> Message-ID: On Mon, 27 Jun 2005, Chris Lumens wrote: > > If you comment out all the package removals, it "works." By works, i mean it > > seems to install all the packages, but the resulting anaconda-ks.cfg file in > > /root does not have anything meaningful in the %package section. It almost > > seems like package selection in the kickstart files is not supported anymore? > > I have not found a similar bug on bugzilla which makes me wonder if something > > changed and I'm not aware? if not I will file a bugzilla bug. Doing a > > graphical install of FC4 on all our machines would take forever! > > See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160209. The > quick explanation is that we added support to exclude by architecture at > the last minute and broke the normal case. A workaround would be to > append ".i386" or whatever to the names of the packages you want to > exclude. Is an update going to be issued for this?? Regards, Tom Diehl tdiehl at rogueind.com Spamtrap address mtd123 at rogueind.com From phr at doc.ic.ac.uk Thu Jun 30 20:29:09 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 30 Jun 2005 21:29:09 +0100 (BST) Subject: Kickstarting FC2 installs In-Reply-To: <4FD2C985D5E2A642AE25823DFD61C2B007037354@orca.agcom.amgreetings.com> References: <4FD2C985D5E2A642AE25823DFD61C2B007037354@orca.agcom.amgreetings.com> Message-ID: On Thu, 30 Jun 2005, MW Mike Weiner (5028) wrote: >Is it possible to do what I am trying (eg create a ks bootdisk and do >an nfs install of fc2)? Everything I have read leads me to believe it >can. So my question is, why am I seeing two different behaviors? The >only REAL difference is via PXE I load the vmlinuz and initrd.img from >the FC2/images/pxeboot tree rather than using say bootimg and dd to >create a boot disk. Would this cause this type of behavior? I tried to >get those two files onto my boot disk but alas they are too large for a >1.44M floppy. Erm, which files *did* fit? You'd need both vmlinuz and initrd to get the network drivers, without which you'd see, well, the issues you're having. You'd have to jump through a painful number of hoops to get working floppy bootdisk with FC2. Is CD-ROM an option? Cheers, Phil From MWeiner at ag.com Thu Jun 30 21:08:38 2005 From: MWeiner at ag.com (MW Mike Weiner (5028)) Date: Thu, 30 Jun 2005 17:08:38 -0400 Subject: Kickstarting FC2 installs Message-ID: <4FD2C985D5E2A642AE25823DFD61C2B0070375EF@orca.agcom.amgreetings.com> -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Philip Rowlands Sent: Thursday, June 30, 2005 4:29 PM To: Discussion list about Kickstart Subject: Re: Kickstarting FC2 installs On Thu, 30 Jun 2005, MW Mike Weiner (5028) wrote: >Is it possible to do what I am trying (eg create a ks bootdisk and do >an nfs install of fc2)? Everything I have read leads me to believe it >can. So my question is, why am I seeing two different behaviors? The >only REAL difference is via PXE I load the vmlinuz and initrd.img from >the FC2/images/pxeboot tree rather than using say bootimg and dd to >create a boot disk. Would this cause this type of behavior? I tried to >get those two files onto my boot disk but alas they are too large for a >1.44M floppy. Erm, which files *did* fit? You'd need both vmlinuz and initrd to get the network drivers, without which you'd see, well, the issues you're having. You'd have to jump through a painful number of hoops to get working floppy bootdisk with FC2. Is CD-ROM an option? -- Ya, unfortunately a CD ks boot disk is required. The initrd and vmlinuz alone are well over the size of the floppy, plus, it seems that if you use a boot disk it wont find the ks.cfg file if provided over a network, as the network hasn't been loaded yet and this will fail. At least from what I am seeing, and a number of resources I have read. So to that end, unfortunately, I will burn to a miniCD and go that route, but what a waste. Thanks for the response Michael Weiner