From Gerrard.Geldenhuis at datacash.com Fri Feb 1 12:21:19 2008 From: Gerrard.Geldenhuis at datacash.com (Gerrard Geldenhuis) Date: Fri, 1 Feb 2008 12:21:19 -0000 Subject: Weird issue trying to boot Dell 2450 ... In-Reply-To: References: Message-ID: I'm purely speculating here but maybe the bios is looking in the wrong place on the CD. There might be an incompatibility between modern boot cds and old bios... Regards > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list- > bounces at redhat.com] On Behalf Of Klaus Steden > Sent: 31 January 2008 19:23 > To: Discussion list about Kickstart > Subject: Weird issue trying to boot Dell 2450 ... > > > Hi there, > > I've encountered a strange issue trying to bootstrap an old Dell 2450. It > will boot from floppy, it will boot from a Dell provisioning CD, but it > won't boot from a CentOS disc (neither the LiveCD nor a CentOS 4.5 ISO), > and > it won't boot its own hard drive. > > I'm working around it using a boot floppy (!! how vintage) to bootstrap > the > OS into loading, but I'm baffled as to why it won't load itself by any > other > means. When Windows 2003 was installed, it booted off the hard drive fine, > so I know it -can- do it ... but I have no clue why it won't load Linux on > its own. > > Has anyone ever had an experience like this with a machine, Dell or > otherwise? It's running Fedora Core 6, so it's a fairly current OS, and > has > no issues when the OS is actually loaded. > > I'm okay with it booting off the floppy but that seems like kind of a > ghetto > solution, so I'd rather not have to if I don't have to. > > thanks, > Klaus > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From accarlson at gmail.com Fri Feb 1 17:46:32 2008 From: accarlson at gmail.com (Augusto Castelan Carlson) Date: Fri, 1 Feb 2008 15:46:32 -0200 Subject: Two network interfaces. Message-ID: Hello! I will have tow install a server with two network interfaces. The host will have two guest, Dom0 e Dom1. All fedora 8. I would like to have one guest and the host and the Dom0 using the interface eth0 after install, and Dom1 using the interface eth1. The host and guest will be in the same network. I automated the install with kickstart. How can I set this info during the install with kickstart? Only saying to use eth0 and eth1 in the network parameter is enough? Because guest would get peth0 and peth1 right? Thanks in advance. -- Augusto Castelan Carlson From dwight at supercomputer.org Sat Feb 2 07:31:31 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Fri, 1 Feb 2008 23:31:31 -0800 Subject: Weird issue trying to boot Dell 2450 ... In-Reply-To: References: Message-ID: <200802012331.32243.dwight@supercomputer.org> Many older systems have CD-ROM drives that only support CD-R format. They won't recognize a CD-RW disk. That would be the first thing that I'd check. I used to have a system like that. The key question here is whether Fedora ir CentOS will even read the CD that you have. If it is indeed a media incompatibility, then no OS will be able to read it. If the CD can indeed be read, then I'd suspect a BIOS issue. -dwight- On Thursday 31 January 2008 11:23:27 am Klaus Steden wrote: > Hi there, > > I've encountered a strange issue trying to bootstrap an old Dell > 2450. It will boot from floppy, it will boot from a Dell > provisioning CD, but it won't boot from a CentOS disc (neither the > LiveCD nor a CentOS 4.5 ISO), and it won't boot its own hard > drive. > > I'm working around it using a boot floppy (!! how vintage) to > bootstrap the OS into loading, but I'm baffled as to why it won't > load itself by any other means. When Windows 2003 was installed, > it booted off the hard drive fine, so I know it -can- do it ... > but I have no clue why it won't load Linux on its own. > > Has anyone ever had an experience like this with a machine, Dell > or otherwise? It's running Fedora Core 6, so it's a fairly current > OS, and has no issues when the OS is actually loaded. > > I'm okay with it booting off the floppy but that seems like kind > of a ghetto solution, so I'd rather not have to if I don't have > to. > > thanks, > Klaus > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From klaus.steden at thomson.net Mon Feb 4 18:54:03 2008 From: klaus.steden at thomson.net (Klaus Steden) Date: Mon, 04 Feb 2008 10:54:03 -0800 Subject: Weird issue trying to boot Dell 2450 ... In-Reply-To: <200802012331.32243.dwight@supercomputer.org> Message-ID: Yeah, the CD is fine (I can read it elsewhere), so I think you?re right about it not being able to read CD-RWs. I wouldn?t be surprised, this machine is quite old (Pentium 3!!) cheers, Klaus On 2/1/08 11:31 PM, "dwight at supercomputer.org" did etch on stone tablets: > Many older systems have CD-ROM drives that only support CD-R format. > They won't recognize a CD-RW disk. That would be the first thing > that I'd check. I used to have a system like that. > > The key question here is whether Fedora ir CentOS will even read the > CD that you have. If it is indeed a media incompatibility, then no > OS will be able to read it. If the CD can indeed be read, then I'd > suspect a BIOS issue. > > -dwight- > > On Thursday 31 January 2008 11:23:27 am Klaus Steden wrote: >> > Hi there, >> > >> > I've encountered a strange issue trying to bootstrap an old Dell >> > 2450. It will boot from floppy, it will boot from a Dell >> > provisioning CD, but it won't boot from a CentOS disc (neither the >> > LiveCD nor a CentOS 4.5 ISO), and it won't boot its own hard >> > drive. >> > >> > I'm working around it using a boot floppy (!! how vintage) to >> > bootstrap the OS into loading, but I'm baffled as to why it won't >> > load itself by any other means. When Windows 2003 was installed, >> > it booted off the hard drive fine, so I know it -can- do it ... >> > but I have no clue why it won't load Linux on its own. >> > >> > Has anyone ever had an experience like this with a machine, Dell >> > or otherwise? It's running Fedora Core 6, so it's a fairly current >> > OS, and has no issues when the OS is actually loaded. >> > >> > I'm okay with it booting off the floppy but that seems like kind >> > of a ghetto solution, so I'd rather not have to if I don't have >> > to. >> > >> > thanks, >> > Klaus >> > >> > _______________________________________________ >> > 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 Matt.Fahrner at coat.com Mon Feb 4 19:00:12 2008 From: Matt.Fahrner at coat.com (Matt Fahrner) Date: Mon, 04 Feb 2008 14:00:12 -0500 Subject: Weird issue trying to boot Dell 2450 ... In-Reply-To: References: Message-ID: <47A760BC.4060002@coat.com> There might be a firmware image that you can use to upgrade the CD drive. I've been able to get CD readers to recognize additional media by getting the latest firmware in (or in some cases, hacked firmwares)... - Matt Klaus Steden wrote: > > Yeah, the CD is fine (I can read it elsewhere), so I think you?re right > about it not being able to read CD-RWs. I wouldn?t be surprised, this > machine is quite old (Pentium 3!!) > > cheers, > Klaus > > On 2/1/08 11:31 PM, "dwight at supercomputer.org" > did etch on stone tablets: > > Many older systems have CD-ROM drives that only support CD-R format. > They won't recognize a CD-RW disk. That would be the first thing > that I'd check. I used to have a system like that. > > The key question here is whether Fedora ir CentOS will even read the > CD that you have. If it is indeed a media incompatibility, then no > OS will be able to read it. If the CD can indeed be read, then I'd > suspect a BIOS issue. > > -dwight- > > On Thursday 31 January 2008 11:23:27 am Klaus Steden wrote: > > Hi there, > > > > I've encountered a strange issue trying to bootstrap an old Dell > > 2450. It will boot from floppy, it will boot from a Dell > > provisioning CD, but it won't boot from a CentOS disc (neither the > > LiveCD nor a CentOS 4.5 ISO), and it won't boot its own hard > > drive. > > > > I'm working around it using a boot floppy (!! how vintage) to > > bootstrap the OS into loading, but I'm baffled as to why it won't > > load itself by any other means. When Windows 2003 was installed, > > it booted off the hard drive fine, so I know it -can- do it ... > > but I have no clue why it won't load Linux on its own. > > > > Has anyone ever had an experience like this with a machine, Dell > > or otherwise? It's running Fedora Core 6, so it's a fairly current > > OS, and has no issues when the OS is actually loaded. > > > > I'm okay with it booting off the floppy but that seems like kind > > of a ghetto solution, so I'd rather not have to if I don't have > > to. > > > > thanks, > > Klaus > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > -- --------------------------------------------------------------------- Matt Fahrner 2 South Park St. Chief Systems Architect Willis House Burlington Coat Factory Warehouse Lebanon, N.H. 03766 Tel: (603) 448-4100 x5150 USA Fax: (603) 443-6190 Matt.Fahrner at COAT.COM --------------------------------------------------------------------- From klaus.steden at thomson.net Mon Feb 4 20:13:54 2008 From: klaus.steden at thomson.net (Klaus Steden) Date: Mon, 04 Feb 2008 12:13:54 -0800 Subject: Weird issue trying to boot Dell 2450 ... In-Reply-To: <47A760BC.4060002@coat.com> Message-ID: Hmmm, I wonder ... it might be easier just to swap out the drive for a different (newer) one, if I can find one lying around. However ... I'm still a little baffled as to why it won't boot the hard drive directly, either. Klaus On 2/4/08 11:00 AM, "Matt Fahrner" did etch on stone tablets: > There might be a firmware image that you can use to upgrade the CD > drive. I've been able to get CD readers to recognize additional media by > getting the latest firmware in (or in some cases, hacked firmwares)... > > - Matt > > Klaus Steden wrote: >> >> Yeah, the CD is fine (I can read it elsewhere), so I think you?re right >> about it not being able to read CD-RWs. I wouldn?t be surprised, this >> machine is quite old (Pentium 3!!) >> >> cheers, >> Klaus >> >> On 2/1/08 11:31 PM, "dwight at supercomputer.org" >> did etch on stone tablets: >> >> Many older systems have CD-ROM drives that only support CD-R format. >> They won't recognize a CD-RW disk. That would be the first thing >> that I'd check. I used to have a system like that. >> >> The key question here is whether Fedora ir CentOS will even read the >> CD that you have. If it is indeed a media incompatibility, then no >> OS will be able to read it. If the CD can indeed be read, then I'd >> suspect a BIOS issue. >> >> -dwight- >> >> On Thursday 31 January 2008 11:23:27 am Klaus Steden wrote: >>> Hi there, >>> >>> I've encountered a strange issue trying to bootstrap an old Dell >>> 2450. It will boot from floppy, it will boot from a Dell >>> provisioning CD, but it won't boot from a CentOS disc (neither the >>> LiveCD nor a CentOS 4.5 ISO), and it won't boot its own hard >>> drive. >>> >>> I'm working around it using a boot floppy (!! how vintage) to >>> bootstrap the OS into loading, but I'm baffled as to why it won't >>> load itself by any other means. When Windows 2003 was installed, >>> it booted off the hard drive fine, so I know it -can- do it ... >>> but I have no clue why it won't load Linux on its own. >>> >>> Has anyone ever had an experience like this with a machine, Dell >>> or otherwise? It's running Fedora Core 6, so it's a fairly current >>> OS, and has no issues when the OS is actually loaded. >>> >>> I'm okay with it booting off the floppy but that seems like kind >>> of a ghetto solution, so I'd rather not have to if I don't have >>> to. >>> >>> thanks, >>> Klaus >>> >>> _______________________________________________ >>> Kickstart-list mailing list >>> Kickstart-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/kickstart-list >> >> >> >> From crhea at mayo.edu Mon Feb 4 21:43:32 2008 From: crhea at mayo.edu (Cris Rhea) Date: Mon, 4 Feb 2008 15:43:32 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? Message-ID: <20080204214332.GB26448@kaizen.mayo.edu> I'm trying to lay out disk partitions in a certain way-- This has worked for years, but recently broke due either to a change in the way Kickstart works or a bug in Anaconda. I use sfdisk to lay out the drive in the %pre section: %pre sfdisk -uM /dev/sda < References: <20080204214332.GB26448@kaizen.mayo.edu> Message-ID: <47A79BF4.5040900@redhat.com> Cris Rhea wrote: > I'm trying to lay out disk partitions in a certain way-- > This has worked for years, but recently broke due either to > a change in the way Kickstart works or a bug in Anaconda. > > I use sfdisk to lay out the drive in the %pre section: > Can't help you debug your sfdisk, but the magic cobbler (http://cobbler.et.redhat.com) uses by default is the following... %include /tmp/partinfo %pre # Determine how many drives we have set \$(list-harddrives) let numd=\$#/2 d1=\$1 d2=\$3 cat << EOF > /tmp/partinfo part / --fstype ext3 --size=1024 --grow --ondisk=\$d1 --asprimary part swap --size=1024 --ondisk=\$d1 --asprimary EOF This has been working very well. There shouldn't be any reason to call sfdisk that I'm aware of. (Thanks to Chip S. for the tip) --Michael From dwight at supercomputer.org Tue Feb 5 04:20:11 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Mon, 4 Feb 2008 20:20:11 -0800 Subject: Weird issue trying to boot Dell 2450 ... In-Reply-To: References: Message-ID: <200802042020.11540.dwight@supercomputer.org> Yes, if you can find a compatible newer drive, that is the easiest way to go. A firmware update, as Matt suggested, probably won't do it. The issue is the laser in the drive itself. It would be tuned for the CD-R wavelengths, and that's a hardware issue. That's why CD-RW disks don't work - they're made for a different type of laser. As for booting off of the hard disk, that's a different problem. It should be quite straightforward to track down, if you're interested in the effort. Assuming that you're using GRUB, the question would be how far do you get before you see any messages (or do you see any messages at all)? I.e. are you getting through the master boot block (which contains the partition table), through to the first stage bootstrap, or any of the subsequent stages? If you aren't seeing even a "GR", then you're not getting through the MBR, IIRC. GRUB has issues with older BIOSes, so it may be an incompatibility there. There are many alternatives available. syslinux is one. If you could get that going, you'd be able to boot up a kernel. Presumeably that would work, if you've been able to install Fedora on your system via some mechanism. syslinux, on an active FAT partition, is pretty simple and may be worth looking at. -dwight- On Monday 04 February 2008 12:13:54 pm Klaus Steden wrote: > Hmmm, I wonder ... it might be easier just to swap out the drive > for a different (newer) one, if I can find one lying around. > > However ... I'm still a little baffled as to why it won't boot the > hard drive directly, either. > > Klaus > > On 2/4/08 11:00 AM, "Matt Fahrner" did etch > on stone > > tablets: > > There might be a firmware image that you can use to upgrade the > > CD drive. I've been able to get CD readers to recognize > > additional media by getting the latest firmware in (or in some > > cases, hacked firmwares)... > > > > - Matt > > > > Klaus Steden wrote: > >> Yeah, the CD is fine (I can read it elsewhere), so I think > >> you?re right about it not being able to read CD-RWs. I wouldn?t > >> be surprised, this machine is quite old (Pentium 3!!) > >> > >> cheers, > >> Klaus > >> > >> On 2/1/08 11:31 PM, "dwight at supercomputer.org" > >> did etch on stone tablets: > >> > >> Many older systems have CD-ROM drives that only support > >> CD-R format. They won't recognize a CD-RW disk. That would be > >> the first thing that I'd check. I used to have a system like > >> that. > >> > >> The key question here is whether Fedora ir CentOS will even > >> read the CD that you have. If it is indeed a media > >> incompatibility, then no OS will be able to read it. If the CD > >> can indeed be read, then I'd suspect a BIOS issue. > >> > >> -dwight- > >> > >> On Thursday 31 January 2008 11:23:27 am Klaus Steden wrote: > >>> Hi there, > >>> > >>> I've encountered a strange issue trying to bootstrap an old > >>> Dell 2450. It will boot from floppy, it will boot from a Dell > >>> provisioning CD, but it won't boot from a CentOS disc (neither > >>> the LiveCD nor a CentOS 4.5 ISO), and it won't boot its own > >>> hard drive. > >>> > >>> I'm working around it using a boot floppy (!! how vintage) to > >>> bootstrap the OS into loading, but I'm baffled as to why it > >>> won't load itself by any other means. When Windows 2003 was > >>> installed, it booted off the hard drive fine, so I know it > >>> -can- do it ... but I have no clue why it won't load Linux on > >>> its own. > >>> > >>> Has anyone ever had an experience like this with a machine, > >>> Dell or otherwise? It's running Fedora Core 6, so it's a > >>> fairly current OS, and has no issues when the OS is actually > >>> loaded. > >>> > >>> I'm okay with it booting off the floppy but that seems like > >>> kind of a ghetto solution, so I'd rather not have to if I > >>> don't have to. > >>> > >>> thanks, > >>> Klaus > >>> > >>> _______________________________________________ > >>> 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 dwight at supercomputer.org Tue Feb 5 04:31:14 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Mon, 4 Feb 2008 20:31:14 -0800 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080204214332.GB26448@kaizen.mayo.edu> References: <20080204214332.GB26448@kaizen.mayo.edu> Message-ID: <200802042031.14927.dwight@supercomputer.org> You might try doing a `dd if=/dev/zero of=/dev/sda bs=1b count=1` and then rebooting, to zero out the partition table. It sounds like there's a conflict between the kernels' view of the partition table and the subsequent repartitioning of the drive. Hope that helps. -dwight- On Monday 04 February 2008 01:43:32 pm Cris Rhea wrote: > I'm trying to lay out disk partitions in a certain way-- > This has worked for years, but recently broke due either to > a change in the way Kickstart works or a bug in Anaconda. > > I use sfdisk to lay out the drive in the %pre section: > > %pre > > sfdisk -uM /dev/sda < ,150,L > ,,E > ; > ; > ,512,L > ,2048,S > ,6000,L > ,2048,L > ,10000,L > ,,L > EOF > > (this works correctly) > > If I specify what goes where in the KS file: > > part /boot --fstype ext3 --onpart sda1 > part /tmp --fstype ext3 --onpart sda9 > part /var --fstype ext3 --onpart sda8 > part swap --onpart sda6 > part /local1 --fstype ext3 --onpart sda10 > part /usr --fstype ext3 --onpart sda7 > part / --fstype ext3 --onpart sda5 > > Anaconda croaks with the following error: > > 15:27:41 CRITICAL: parted exception: Error: Error informing the > kernel about modifications to partition /dev/sda5 -- Device or > resource busy. This means Linux won't know about any changes you > made to /dev/sda5 until you reboot -- so you shouldn't mount it or > use it in any way before rebooting. > > > Googling this error shows it popping up at various time (back to > FC3 or so), but no solutions. > > This worked fine in CentOS 4.4 (RHEL4 u4), but broke in the next > update (and is still broken in CentOS 5.1 [RHEL5 u1]). > > I could work around this in CentOS 4 by loading systems against an > old 4.4 image, then applying all the updates to get to 4.6. I > can't do this in 5, since it is broken for all releases... > > Is there better way to achieve what I'm doing or is this just > an Anaconda bug that's making my life difficult? I don't see an > advantage to moving everything to LVM for simple single-disk > systems... > > > Thanks-- > > --- Cris > > PS: I have a couple Anaconda dumps if someone is interested... From debian at herakles.homelinux.org Tue Feb 5 11:05:07 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Tue, 05 Feb 2008 20:05:07 +0900 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080204214332.GB26448@kaizen.mayo.edu> References: <20080204214332.GB26448@kaizen.mayo.edu> Message-ID: <47A842E3.5040000@herakles.homelinux.org> Cris Rhea wrote: > I'm trying to lay out disk partitions in a certain way-- > This has worked for years, but recently broke due either to > a change in the way Kickstart works or a bug in Anaconda. > > I use sfdisk to lay out the drive in the %pre section: > > %pre > > sfdisk -uM /dev/sda < ,150,L > ,,E > ; > ; > ,512,L > ,2048,S > ,6000,L > ,2048,L > ,10000,L > ,,L > EOF > > (this works correctly) > > If I specify what goes where in the KS file: > > part /boot --fstype ext3 --onpart sda1 > part /tmp --fstype ext3 --onpart sda9 > part /var --fstype ext3 --onpart sda8 > part swap --onpart sda6 > part /local1 --fstype ext3 --onpart sda10 > part /usr --fstype ext3 --onpart sda7 > part / --fstype ext3 --onpart sda5 > > Anaconda croaks with the following error: > > 15:27:41 CRITICAL: parted exception: Error: Error informing the kernel > about modifications to partition /dev/sda5 -- Device or resource busy. > This means Linux won't know about any changes you made to /dev/sda5 until > you reboot -- so you shouldn't mount it or use it in any way before rebooting. > > > Googling this error shows it popping up at various time (back to FC3 or > so), but no solutions. > > This worked fine in CentOS 4.4 (RHEL4 u4), but broke in the next update > (and is still broken in CentOS 5.1 [RHEL5 u1]). > > I could work around this in CentOS 4 by loading systems against an old > 4.4 image, then applying all the updates to get to 4.6. I can't do this in > 5, since it is broken for all releases... > > Is there better way to achieve what I'm doing or is this just > an Anaconda bug that's making my life difficult? I don't see an advantage > to moving everything to LVM for simple single-disk systems... It's possible to drive fdisk in much the same way; if Michael's suggestion doesn't work then that might (different code may produce a different result). So might not creating sda5 in {,s}fdisk, and doing it in Anaconda instead. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From crhea at mayo.edu Tue Feb 5 12:02:52 2008 From: crhea at mayo.edu (Cris Rhea) Date: Tue, 5 Feb 2008 06:02:52 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <200802042031.14927.dwight@supercomputer.org> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802042031.14927.dwight@supercomputer.org> Message-ID: <20080205120252.GA29094@kaizen.mayo.edu> On Mon, Feb 04, 2008 at 08:31:14PM -0800, dwight at supercomputer.org wrote: > You might try doing a `dd if=/dev/zero of=/dev/sda bs=1b count=1` and > then rebooting, to zero out the partition table. It sounds like > there's a conflict between the kernels' view of the partition table > and the subsequent repartitioning of the drive. > > Hope that helps. > > -dwight- > I've tried with and without zeroing the first block (did this to nuke some systems that came from Sun with ZFS/GPT). No change... The sfdisk IS SUCCESSFUL (looking at the disk with the shell). I've even tried to partition it with one boot, then use those partitions on the next boot. Same issue. Thanks for the suggestion. -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From crhea at mayo.edu Tue Feb 5 12:13:28 2008 From: crhea at mayo.edu (Cris Rhea) Date: Tue, 5 Feb 2008 06:13:28 -0600 Subject: Weird issue trying to boot Dell 2450 ... Message-ID: <20080205121328.GA29133@kaizen.mayo.edu> I think you have something going on with that specific system... I've still got several 2450s running in production-- all were booted and/or loaded from CD-R (didn't try CD-RW). I'm also running CentOS-- my latest machine is running 4.4 (i.e., haven't tried to load a 2450 from a 4.5 or 4.6 CD). If memory serves me (I'm at home at the moment), a 2450 uses a laptop CDROM drive-- I've had *many* of these drives go flaky from this vintage. --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From crhea at mayo.edu Tue Feb 5 12:33:16 2008 From: crhea at mayo.edu (Cris Rhea) Date: Tue, 5 Feb 2008 06:33:16 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? Message-ID: <20080205123316.GA29157@kaizen.mayo.edu> Michael DeHaan wrote: > Can't help you debug your sfdisk, but the magic cobbler > (http://cobbler.et.redhat.com) uses by default is the following... > > %include /tmp/partinfo > > %pre > # Determine how many drives we have > set \$(list-harddrives) > let numd=\$#/2 > d1=\$1 > d2=\$3 > > cat << EOF > /tmp/partinfo > part / --fstype ext3 --size=1024 --grow --ondisk=\$d1 --asprimary > part swap --size=1024 --ondisk=\$d1 --asprimary > EOF > > This has been working very well. > > There shouldn't be any reason to call sfdisk that I'm aware of. The difference is that in my case, I have an external program (sfdisk) writing the partition table-- in the above, Anaconda is writing the partition table with its internal routines. Sfdisk writes the partition table correctly (as viewed by a shell window), but Anaconda gets confused using it. I don't want to use something like the above because: -- I want to have more granular file systems (separating /, /usr, /var, /tmp, etc.). -- I want to have the same layout across many systems (e.g., cluster nodes). From what I've seen, having Anaconda/Disk Druid lay out partitions can result in them being in a different order (e.g., Disk Druid arranges them by size). Controlling disk layout compared to having everything under "/" is a subjective, administrative decision. I'm not trying to convince anybody that "my way is better", just trying to get it to work (and it has always worked fine in everything pre-RHEL4u5). Thanks-- --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From Pablo.Iranzo at redhat.com Tue Feb 5 12:58:09 2008 From: Pablo.Iranzo at redhat.com (Pablo Iranzo =?ISO-8859-1?Q?G=F3mez?=) Date: Tue, 05 Feb 2008 12:58:09 +0000 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080205123316.GA29157@kaizen.mayo.edu> References: <20080205123316.GA29157@kaizen.mayo.edu> Message-ID: <1202216289.5467.51.camel@iranzo.users.redhat.com> Isn't LVM an option? sfdisk --force /dev/hda < From crhea at mayo.edu Tue Feb 5 17:24:59 2008 From: crhea at mayo.edu (Cris Rhea) Date: Tue, 5 Feb 2008 11:24:59 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080205170017.467CC735B1@hormel.redhat.com> References: <20080205170017.467CC735B1@hormel.redhat.com> Message-ID: <20080205172459.GA29957@kaizen.mayo.edu> > From: Pablo Iranzo G?mez > > Isn't LVM an option? > > sfdisk --force /dev/hda < 0,200,83,* > 201,,8e > EOF > > Then: > > part /boot --fstype ext3 --onpart=hda1 > part pv.6 --onpart=hda3 > volgroup THEBIGONE pv.6 > logvol / --fstype ext3 --name=root --vgname=THEBIGONE --size=1024 > logvol /home --fstype ext3 --name=home --vgname=THEBIGONE --size=150 > logvol /opt --fstype ext3 --name=opt --vgname=THEBIGONE --size=2000 > logvol /tmp --fstype ext3 --name=tmp --vgname=THEBIGONE --size=1024 > logvol /usr --fstype ext3 --name=usr --vgname=THEBIGONE --size=1500 > logvol swap --fstype swap --vgname=THEBIGONE --name=Swap --size=256 > > So just have two partitions, always hda1 as /boot, and the rest for a > LVM PV ;) > > Regards > Pablo > > PD: I have to dd first sectors of LVM in order to get it working for a > reinstallation If I can't figure out what's broken and get it fixed, I may have to go this route... to me, adding the complexity of LVM to a system that could (and did in the past) use simple partitions is adding bloat to Linux. I love LVM in multi-disk situations where you need to be able to span drives or easily resize volumes... BUT, I also like the simplicity of popping in a rescue disk and mounting /dev/sda5 to access and fix a problem in "/". I'm also curious-- if your suggestion above works (where you partition the disk with sfdisk and then use LVM), I'm even more curious why my version doesn't work-- I'm doing the same thing, just more statements of the form: "part /boot --fstype ext3 --onpart=hda1". Thanks for the suggestion... --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From Pablo.Iranzo at redhat.com Tue Feb 5 17:43:31 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Tue, 5 Feb 2008 18:43:31 +0100 (CET) Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080205172459.GA29957@kaizen.mayo.edu> References: <20080205170017.467CC735B1@hormel.redhat.com> <20080205172459.GA29957@kaizen.mayo.edu> Message-ID: On Tue, 5 Feb 2008, Cris Rhea wrote: > > From: Pablo Iranzo G?mez > If I can't figure out what's broken and get it fixed, I may have to > go this route... to me, adding the complexity of LVM to a system that could > (and did in the past) use simple partitions is adding bloat to Linux. LVM will allow you to redimension FS on demand without having to switch to single runlevel > I love LVM in multi-disk situations where you need to be able to span drives > or easily resize volumes... > > BUT, I also like the simplicity of popping in a rescue disk and mounting > /dev/sda5 to access and fix a problem in "/". Use the install media kernel and initrd.img and boot in "rescue" mode... you'll have lvm to activate vg's and lv's and you can mount them ;)... but yes, it's easier with plain partitions > I'm also curious-- if your suggestion above works (where you > partition the disk with sfdisk and then use LVM), I'm even more > curious why my version doesn't work-- I'm doing the same thing, just > more statements of the form: "part /boot --fstype ext3 --onpart=hda1". Mine is working with another layer... I'm kickstarting 64 mb machines (heavily unsuported), so I create a third partiton which a swapfile that I'm using during kickstart in order to avoid anaconda from aborting, in %post, I disable that swapfile, change partitition type with sfdisk (just edit type), I convert that partition to a PV and extending VG over that PV in order to not wasting space... ;) Regards Pablo From Tim.Mooney at ndsu.edu Tue Feb 5 18:07:34 2008 From: Tim.Mooney at ndsu.edu (Tim Mooney) Date: Tue, 5 Feb 2008 12:07:34 -0600 (CST) Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080205172459.GA29957@kaizen.mayo.edu> References: <20080205170017.467CC735B1@hormel.redhat.com> <20080205172459.GA29957@kaizen.mayo.edu> Message-ID: In regard to: Re: Old style fixed partitions in KS-- Anaconda bug??, Cris...: > I'm also curious-- if your suggestion above works (where you > partition the disk with sfdisk and then use LVM), I'm even more > curious why my version doesn't work-- I'm doing the same thing, just > more statements of the form: "part /boot --fstype ext3 --onpart=hda1". I used this just recently on RHEL 4.6, and it worked fine. I did partitioning manually in %pre (I didn't have time to get sfdisk right, though your examples will undoubtedly help me down the road...), and also created the filesystems and the basic LVM in %pre too. zerombr yes # # TVM: don't do anything to the disk label. # clearpart --none #/boot part /boot --noformat --onpart=hda1 # swap part swap --noformat --onpart=hda2 # / part / --noformat --onpart=hda3 # remainder for LVM part pv.01 --noformat --onpart=hda4 volgroup localvg1 pv.01 --noformat logvol /usr/local --noformat --name=local --vgname=localvg1 logvol /tmp --noformat --name=tmp --vgname=localvg1 logvol /var --noformat --name=var --vgname=localvg1 logvol /home --noformat --name=home --vgname=localvg1 logvol /u01 --noformat --name=oracle --vgname=localvg1 Tim -- Tim Mooney Tim.Mooney at ndsu.edu Information Technology Services (701) 231-1076 (Voice) Room 242-J6, IACC Building (701) 231-8541 (Fax) North Dakota State University, Fargo, ND 58105-5164 From mdokter at xs4all.nl Tue Feb 5 18:50:48 2008 From: mdokter at xs4all.nl (Mark C. Dokter) Date: Tue, 05 Feb 2008 19:50:48 +0100 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: References: <20080205170017.467CC735B1@hormel.redhat.com> <20080205172459.GA29957@kaizen.mayo.edu> Message-ID: <47A8B008.9020302@xs4all.nl> How can I unsubscribe from this list. redhat does not accept my email, no way to contact them (It is illegal what they are doing) Tim Mooney wrote: > In regard to: Re: Old style fixed partitions in KS-- Anaconda bug??, > Cris...: > >> I'm also curious-- if your suggestion above works (where you >> partition the disk with sfdisk and then use LVM), I'm even more >> curious why my version doesn't work-- I'm doing the same thing, just >> more statements of the form: "part /boot --fstype ext3 --onpart=hda1". > > I used this just recently on RHEL 4.6, and it worked fine. I did > partitioning manually in %pre (I didn't have time to get sfdisk right, > though your examples will undoubtedly help me down the road...), and > also created the filesystems and the basic LVM in %pre too. > > zerombr yes > # > # TVM: don't do anything to the disk label. > # > clearpart --none > > #/boot > part /boot --noformat --onpart=hda1 > > # swap > part swap --noformat --onpart=hda2 > > # / > part / --noformat --onpart=hda3 > > # remainder for LVM > part pv.01 --noformat --onpart=hda4 > > volgroup localvg1 pv.01 --noformat > > logvol /usr/local --noformat --name=local --vgname=localvg1 > logvol /tmp --noformat --name=tmp --vgname=localvg1 > logvol /var --noformat --name=var --vgname=localvg1 > logvol /home --noformat --name=home --vgname=localvg1 > logvol /u01 --noformat --name=oracle --vgname=localvg1 > > > > > Tim From Pablo.Iranzo at redhat.com Tue Feb 5 18:55:45 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Tue, 5 Feb 2008 19:55:45 +0100 (CET) Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <47A8B008.9020302@xs4all.nl> References: <20080205170017.467CC735B1@hormel.redhat.com> <20080205172459.GA29957@kaizen.mayo.edu> <47A8B008.9020302@xs4all.nl> Message-ID: Hi On Tue, 5 Feb 2008, Mark C. Dokter wrote: > How can I unsubscribe from this list. redhat does not accept my email, > no way to contact them (It is illegal what they are doing) This has remembered me this story: http://www.centos.org/modules/news/article.php?storyid=127 :D As with any mailing list, there is information at the end of each mail explaining what to do: > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list So go that webpage, enter your data (email, etc) and you can set what you want to do, even if you don't know your password, you can reset it using the form that will send you a new password in order to unsubscribe. If you can't remember from which email you subscribed, just check your email headers to learn which one was used. Regards Pablo From dwight at supercomputer.org Wed Feb 6 17:14:59 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Wed, 6 Feb 2008 09:14:59 -0800 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080205120252.GA29094@kaizen.mayo.edu> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802042031.14927.dwight@supercomputer.org> <20080205120252.GA29094@kaizen.mayo.edu> Message-ID: <200802060915.00288.dwight@supercomputer.org> On Tuesday 05 February 2008 04:02:52 am Cris Rhea wrote: > I've tried with and without zeroing the first block (did this to > nuke some systems that came from Sun with ZFS/GPT). No change... > > The sfdisk IS SUCCESSFUL (looking at the disk with the > shell). One has to be careful. There's an in-core copy of the partition table (used by the kernel) and an on-disk copy. The two are different, and can be out of sync. Looking at the changes by hand can give you the in-core copy, when in fact it hasn't been written to disk. This has been an annoying problem across all versions of UNIX and Linux ever since UNIX was first put on the PC back in the mid 1980's. And this is, in fact, what anaconda/parted appear to be running into: > 15:27:41 CRITICAL: parted exception: Error: Error informing the > kernel about modifications to partition /dev/sda5 -- Device or > resource busy. This means Linux won't know about any changes you > made to /dev/sda5 until you reboot -- so you shouldn't mount it or > use it in any way before rebooting. It would seem that anaconda is somehow leaving open /dev/sda, and preventing the new table from being sync'd. But this only happens if you do modifications in the %pre section. Clearly this is a failure on anaconda's part. The options aren't pleasant. One quick workaround might be to kickstart twice. The first time through, do the partitioning as you have it, and then reboot. The second time would do the rest of your kickstart file. This isn't pleasant, but it might work. Barring that, I think one would have to dig into anaconda, and figureout who is keeping that device open between the %pre section, and when parted is run. I hope that doesn't sound discouraging, because what you are doing is pretty important. The lack of a reliable way to define partitions in the %pre section is IMHO the greatest weakness of Kickstart. Anaconda's preordained ability to overwrite what the user wants, with no alternative, based upon arbitrary and undocumented defaults, is very limiting in a production environment. And your solution here is the best that I've seen. It should part of a FAQ somewhere. -dwight- From crhea at mayo.edu Wed Feb 6 18:41:20 2008 From: crhea at mayo.edu (Cris Rhea) Date: Wed, 6 Feb 2008 12:41:20 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <200802060915.00288.dwight@supercomputer.org> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802042031.14927.dwight@supercomputer.org> <20080205120252.GA29094@kaizen.mayo.edu> <200802060915.00288.dwight@supercomputer.org> Message-ID: <20080206184120.GB1707@kaizen.mayo.edu> On Wed, Feb 06, 2008 at 09:14:59AM -0800, dwight at supercomputer.org wrote: > One has to be careful. There's an in-core copy of the partition table > (used by the kernel) and an on-disk copy. The two are different, and > can be out of sync. Looking at the changes by hand can give you the > in-core copy, when in fact it hasn't been written to disk. I've verified that it was correctly written to disk-- both within the Kickstart session ( shell) and after the fact with a rescue disk. I've used this method successfully since the RedHat 7 days... > This has been an annoying problem across all versions of UNIX and > Linux ever since UNIX was first put on the PC back in the mid 1980's. sync; sync; sync :) > It would seem that anaconda is somehow leaving open /dev/sda, and > preventing the new table from being sync'd. But this only happens if > you do modifications in the %pre section. > > Clearly this is a failure on anaconda's part. The options aren't > pleasant. This appears to be exactly what is happening. Googling the error, there are other postings suggesting an open file descriptor in Anaconda. I was fine and didn't trip over this until CentOS4.5 (RHEL4.5). > One quick workaround might be to kickstart twice. The first time > through, do the partitioning as you have it, and then reboot. The > second time would do the rest of your kickstart file. This isn't > pleasant, but it might work. I've tried that-- strangely, it fails the same way. > Barring that, I think one would have to dig into anaconda, and > figureout who is keeping that device open between the %pre section, > and when parted is run. If I could figure out how to set up a dev/test environment for Anaconda, I would have gladly fixed this by now... been battling his for 6+ months. I just don't know enough (and can't find the info) on loading Anaconda/stage 2 boot, etc. > I hope that doesn't sound discouraging, because what you are doing is > pretty important. The lack of a reliable way to define partitions in > the %pre section is IMHO the greatest weakness of Kickstart. > Anaconda's preordained ability to overwrite what the user wants, > with no alternative, based upon arbitrary and undocumented defaults, > is very limiting in a production environment. IMHO, the various distributions are fighting two battles- one camp wants to make loading Linux so simple anybody can do it (the distributions where there's /boot and / as the only file systems). That versus the camps where they are trying to add every advanced feature from all the different versions of Unix (LVM, software RAID, journaling file systems, etc.). I've been doing Unix/Linux admin for 20+ years... I want to make the decisions on how my systems are set up-- not some idiot-proof admin tool like Anaconda. Even using Anaconda manually (no Kickstart), it's getting harder and harder to manually decide/partition disks as I wish -- rather than "just trusting Anaconda and Disk Druid". I much preferred the days where a radio button allowed users to manually partition with fdisk... > And your solution here is the best that I've seen. It should part > of a FAQ somewhere. Thanks-- I've seen some really cool ideas presented in this forum. > > -dwight- -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From klandreth at theplanet.com Wed Feb 6 19:58:27 2008 From: klandreth at theplanet.com (Landreth, Kevin) Date: Wed, 6 Feb 2008 13:58:27 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <200802060915.00288.dwight@supercomputer.org> References: <20080204214332.GB26448@kaizen.mayo.edu><200802042031.14927.dwight@supercomputer.org><20080205120252.GA29094@kaizen.mayo.edu> <200802060915.00288.dwight@supercomputer.org> Message-ID: <69F7A3B9C49D9D409AF0F06F8AC5076704666854@HOUEXCH01.PLANET.LOCAL> After sfdisk, do blockdev --rereadpt or partprobe to force the kernel to understand that the disk layout has changed. Once for each disk. Let us know if that helps :) -- Thanks, Kevin Landreth, RHCE Technology Architect -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of dwight at supercomputer.org Sent: Wednesday, February 06, 2008 11:15 AM To: kickstart-list at redhat.com Cc: Cris Rhea Subject: Re: Old style fixed partitions in KS-- Anaconda bug?? On Tuesday 05 February 2008 04:02:52 am Cris Rhea wrote: > I've tried with and without zeroing the first block (did this to > nuke some systems that came from Sun with ZFS/GPT). No change... > > The sfdisk IS SUCCESSFUL (looking at the disk with the > shell). One has to be careful. There's an in-core copy of the partition table (used by the kernel) and an on-disk copy. The two are different, and can be out of sync. Looking at the changes by hand can give you the in-core copy, when in fact it hasn't been written to disk. This has been an annoying problem across all versions of UNIX and Linux ever since UNIX was first put on the PC back in the mid 1980's. And this is, in fact, what anaconda/parted appear to be running into: > 15:27:41 CRITICAL: parted exception: Error: Error informing the > kernel about modifications to partition /dev/sda5 -- Device or > resource busy. This means Linux won't know about any changes you > made to /dev/sda5 until you reboot -- so you shouldn't mount it or > use it in any way before rebooting. It would seem that anaconda is somehow leaving open /dev/sda, and preventing the new table from being sync'd. But this only happens if you do modifications in the %pre section. Clearly this is a failure on anaconda's part. The options aren't pleasant. One quick workaround might be to kickstart twice. The first time through, do the partitioning as you have it, and then reboot. The second time would do the rest of your kickstart file. This isn't pleasant, but it might work. Barring that, I think one would have to dig into anaconda, and figureout who is keeping that device open between the %pre section, and when parted is run. I hope that doesn't sound discouraging, because what you are doing is pretty important. The lack of a reliable way to define partitions in the %pre section is IMHO the greatest weakness of Kickstart. Anaconda's preordained ability to overwrite what the user wants, with no alternative, based upon arbitrary and undocumented defaults, is very limiting in a production environment. And your solution here is the best that I've seen. It should part of a FAQ somewhere. -dwight- _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From dwight at supercomputer.org Thu Feb 7 17:24:53 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Thu, 7 Feb 2008 09:24:53 -0800 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080206184120.GB1707@kaizen.mayo.edu> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802060915.00288.dwight@supercomputer.org> <20080206184120.GB1707@kaizen.mayo.edu> Message-ID: <200802070924.53497.dwight@supercomputer.org> On Wednesday 06 February 2008 10:41:20 am Cris Rhea wrote: > On Wed, Feb 06, 2008 at 09:14:59AM -0800, dwight at supercomputer.org wrote: > > > This has been an annoying problem across all versions of UNIX > > and Linux ever since UNIX was first put on the PC back in the > > mid 1980's. > > sync; sync; sync :) Heh. Ok, only people who've been around a bit seem to know that piece of ancient UNIX lore these days. I'm running into a lot of young, bright and talented Linux people who don't know about the three syncs.. Of course, Kevin Landreth kindly pointed out the better obscure commands to do this particular deed. Thank you, Kevin. This thread is full of good wisdom. Here's an obscure bit of trivia. Why is it not one, not two, not four, but three syncs? > > One quick workaround might be to kickstart twice. The first time > > through, do the partitioning as you have it, and then reboot. > > The second time would do the rest of your kickstart file. This > > isn't pleasant, but it might work. > > I've tried that-- strangely, it fails the same way. Hang on now. If the partition table is fully predefined, and you're using the onpart arg, and this is happening? Without anything in the %pre section? > IMHO, the various distributions are fighting two battles- one camp > wants to make loading Linux so simple anybody can do it (the > distributions where there's /boot and / as the only file systems). > ... Ditto that. Currently I'm working with a major systems manufacturer to get CentOS/REL imaging going, and this has been quite annoying. I can see the issues which get in the way, but these could be handled. It's "just" a question of effort. As far as a debug environment goes (and this is just off the top of my head), wouldn't the image loaded down from a PXE install be suitable? minstage.img IIRC, on the CD ISO. Combine that on a USB stick with a VMWare image, and that might be an ideal debug environment. But that's just a quick thought. -dwight- From crhea at mayo.edu Thu Feb 7 17:43:04 2008 From: crhea at mayo.edu (Cris Rhea) Date: Thu, 7 Feb 2008 11:43:04 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080207170015.487A67337F@hormel.redhat.com> References: <20080207170015.487A67337F@hormel.redhat.com> Message-ID: <20080207174304.GB5715@kaizen.mayo.edu> > From: "Landreth, Kevin" > Date: Wed, 6 Feb 2008 13:58:27 -0600 > Subject: RE: Old style fixed partitions in KS-- Anaconda bug?? > > After sfdisk, do blockdev --rereadpt or partprobe to force > the kernel to understand that the disk layout has changed. Once for > each disk. > > Let us know if that helps :) > > -- > Thanks, > Kevin Landreth, RHCE > Technology Architect Same result... I don't think there is really an issue between the disks and the kernel-- I think Anaconda itself is confused... Thanks for the suggestion. (Hadn't seen the "blockdev" command before... cool...) --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From crhea at mayo.edu Thu Feb 7 17:58:52 2008 From: crhea at mayo.edu (Cris Rhea) Date: Thu, 7 Feb 2008 11:58:52 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <200802070924.53497.dwight@supercomputer.org> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802060915.00288.dwight@supercomputer.org> <20080206184120.GB1707@kaizen.mayo.edu> <200802070924.53497.dwight@supercomputer.org> Message-ID: <20080207175852.GD5715@kaizen.mayo.edu> On Thu, Feb 07, 2008 at 09:24:53AM -0800, dwight at supercomputer.org wrote: > Here's an obscure bit of trivia. Why is it not one, not two, not > four, but three syncs? Because the system will have done the first sync by the time you finished typing the other two. > Hang on now. If the partition table is fully predefined, and you're > using the onpart arg, and this is happening? Without anything in > the %pre section? Yup... perhaps anaconda is getting confused by the number of "onpart" file systems I'm defining or something equally as bizarre. One of the suggestions that came across the list was to define two partitions with sfdisk-- One for /boot and one for LVM. Since you use the "onpart" tag for the /boot partition, clearly Anaconda can A.) Deal with a disk that has been "sfdisk"ed. and B.) Deal with a partition using "onpart". > As far as a debug environment goes (and this is just off the top of > my head), wouldn't the image loaded down from a PXE install be > suitable? minstage.img IIRC, on the CD ISO. Combine that on a USB > stick with a VMWare image, and that might be an ideal debug > environment. But that's just a quick thought. How does one build a minstg2.img file? Can I just change *that* one file or do I also have to change more (i.e., doesn't Anaconda call all sorts of Python libs located elsewhere)? Yes, I could do the PXE environment (but not the VMWare), or I could use the diskboot.img (USB pendrive image). I've never played around with unrolling/re-rolling those images-- spent a couple hours looking one time, but never found much in the way of docs... > > -dwight- -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From eric at trueblade.com Fri Feb 8 01:05:07 2008 From: eric at trueblade.com (Eric Smith) Date: Thu, 07 Feb 2008 20:05:07 -0500 Subject: Upgrading kernel used in PXE boot Message-ID: <47ABAAC3.4080209@trueblade.com> I'm installing FC6 using kickstart from PXE. It works great, and I've used this (and a similar setup with FC2) for years. Now, I want to use a new kernel during the installation. FC6 comes with 2.6.18-1.2798.fc6, and I want the boot to use some version of 2.6.20 or greater. What I've been doing is installing with the default kernel, then upgrading to 2.6.20-1.2962.fc6 after the installation is complete. But now I have a need to run 2.6.20 during installation, because of some hardware that's supported in 2.6.20 but not in 2.6.18. So my question is: How do I create a new initrd.img and vmlinuz (that came in the isolinux directory) to be a new kernel? It's apparently not as simple as copying these files from a running system. I've spent hours googling. I've seen all sorts of info on adding drivers to initrd.img, but no info on updating the kernel. Any info would be great! Thanks. Eric. From eric at trueblade.com Fri Feb 8 01:50:09 2008 From: eric at trueblade.com (Eric Smith) Date: Thu, 07 Feb 2008 20:50:09 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47ABAAC3.4080209@trueblade.com> References: <47ABAAC3.4080209@trueblade.com> Message-ID: <47ABB551.1020509@trueblade.com> Eric Smith wrote: > So my question is: How do I create a new initrd.img and vmlinuz (that > came in the isolinux directory) to be a new kernel? It's apparently not > as simple as copying these files from a running system. And yes, I realize this might not be the best list for this question, but it's the best I could think of, since people here are no doubt familiar with the issue. If you know of a better venue for this question, please let me know. Eric. From kanarip at kanarip.com Fri Feb 8 10:00:43 2008 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Fri, 08 Feb 2008 11:00:43 +0100 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47ABAAC3.4080209@trueblade.com> References: <47ABAAC3.4080209@trueblade.com> Message-ID: <47AC284B.90305@kanarip.com> Eric Smith wrote: > I'm installing FC6 using kickstart from PXE. It works great, and I've > used this (and a similar setup with FC2) for years. > > Now, I want to use a new kernel during the installation. FC6 comes with > 2.6.18-1.2798.fc6, and I want the boot to use some version of 2.6.20 or > greater. > > What I've been doing is installing with the default kernel, then > upgrading to 2.6.20-1.2962.fc6 after the installation is complete. But > now I have a need to run 2.6.20 during installation, because of some > hardware that's supported in 2.6.20 but not in 2.6.18. > > So my question is: How do I create a new initrd.img and vmlinuz (that > came in the isolinux directory) to be a new kernel? It's apparently not > as simple as copying these files from a running system. > > I've spent hours googling. I've seen all sorts of info on adding > drivers to initrd.img, but no info on updating the kernel. > > Any info would be great! > Hi Eric, -mandatory note about Fedora Core 6 being End Of Life, you should upgrade or use some Enterprise Linux (derivative)- you might want to take a look at Revisor. We've been doing successful Fedora Core 6 spins in the past although it really started off with Fedora 7... To indicate what is successful, here's the test matrix that we commonly use (a Fedora 8 version of it): http://spins.fedoraunity.org/Members/kanarip/fedora-8-test-matrix-template Revisor you can find in the Fedora repositories (if you run Fedora 7 or 8), and there's a package built for EL5 as well. See: http://revisor.fedoraunity.org and it's mailing list at http://lists.fedoraunity.org/mailman/listinfo/revisor-users Kind regards, Jeroen van Meeuwen -kanarip From jgranado at redhat.com Fri Feb 8 10:07:42 2008 From: jgranado at redhat.com (Joel Andres Granados) Date: Fri, 08 Feb 2008 11:07:42 +0100 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47AC284B.90305@kanarip.com> References: <47ABAAC3.4080209@trueblade.com> <47AC284B.90305@kanarip.com> Message-ID: <47AC29EE.9080802@redhat.com> Jeroen van Meeuwen wrote: > Eric Smith wrote: >> I'm installing FC6 using kickstart from PXE. It works great, and I've >> used this (and a similar setup with FC2) for years. >> >> Now, I want to use a new kernel during the installation. FC6 comes >> with 2.6.18-1.2798.fc6, and I want the boot to use some version of >> 2.6.20 or greater. >> >> What I've been doing is installing with the default kernel, then >> upgrading to 2.6.20-1.2962.fc6 after the installation is complete. >> But now I have a need to run 2.6.20 during installation, because of >> some hardware that's supported in 2.6.20 but not in 2.6.18. >> >> So my question is: How do I create a new initrd.img and vmlinuz (that >> came in the isolinux directory) to be a new kernel? It's apparently >> not as simple as copying these files from a running system. >> >> I've spent hours googling. I've seen all sorts of info on adding >> drivers to initrd.img, but no info on updating the kernel. >> >> Any info would be great! >> > > Hi Eric, > > -mandatory note about Fedora Core 6 being End Of Life, you should > upgrade or use some Enterprise Linux (derivative)- > > you might want to take a look at Revisor. We've been doing successful > Fedora Core 6 spins in the past although it really started off with > Fedora 7... To indicate what is successful, here's the test matrix that > we commonly use (a Fedora 8 version of it): > > http://spins.fedoraunity.org/Members/kanarip/fedora-8-test-matrix-template > > Revisor you can find in the Fedora repositories (if you run Fedora 7 or > 8), and there's a package built for EL5 as well. > > See: http://revisor.fedoraunity.org and it's mailing list at > http://lists.fedoraunity.org/mailman/listinfo/revisor-users > > Kind regards, > > Jeroen van Meeuwen > -kanarip > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list + Just wanted to mention the different tools 1. revisor by Jeroen Meeuwen 2. pungi by Jessy Keating 3. buildinstall by anaconda team AFAK they are in that order. (revisor uses pung and pungi uses buildinstall). I think if you want an easy way to do this your best bet is revisor, If you want to get into the core of things you might want to give buildinstall a try. It a very scarry shell script :) Hope it helps. -- Joel Andres Granados Red Hat / Brno, Czech Republic From eric at trueblade.com Fri Feb 8 10:14:01 2008 From: eric at trueblade.com (Eric Smith) Date: Fri, 08 Feb 2008 05:14:01 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47AC284B.90305@kanarip.com> References: <47ABAAC3.4080209@trueblade.com> <47AC284B.90305@kanarip.com> Message-ID: <47AC2B69.2060407@trueblade.com> Jeroen van Meeuwen wrote: > Eric Smith wrote: >> I'm installing FC6 using kickstart from PXE. It works great, and I've >> used this (and a similar setup with FC2) for years. >> >> Now, I want to use a new kernel during the installation. FC6 comes >> with 2.6.18-1.2798.fc6, and I want the boot to use some version of >> 2.6.20 or greater. >> >> What I've been doing is installing with the default kernel, then >> upgrading to 2.6.20-1.2962.fc6 after the installation is complete. >> But now I have a need to run 2.6.20 during installation, because of >> some hardware that's supported in 2.6.20 but not in 2.6.18. >> >> So my question is: How do I create a new initrd.img and vmlinuz (that >> came in the isolinux directory) to be a new kernel? It's apparently >> not as simple as copying these files from a running system. >> >> I've spent hours googling. I've seen all sorts of info on adding >> drivers to initrd.img, but no info on updating the kernel. >> >> Any info would be great! >> > > Hi Eric, > > -mandatory note about Fedora Core 6 being End Of Life, you should > upgrade or use some Enterprise Linux (derivative)- Yes, I know. But I have a lot of FC6 boxes already deployed, and switching just for this one piece of hardware probably isn't worth doing. I'm okay with supporting it myself, for now. > you might want to take a look at Revisor. We've been doing successful > Fedora Core 6 spins in the past although it really started off with > Fedora 7... To indicate what is successful, here's the test matrix that > we commonly use (a Fedora 8 version of it): > > http://spins.fedoraunity.org/Members/kanarip/fedora-8-test-matrix-template > > Revisor you can find in the Fedora repositories (if you run Fedora 7 or > 8), and there's a package built for EL5 as well. > > See: http://revisor.fedoraunity.org and it's mailing list at > http://lists.fedoraunity.org/mailman/listinfo/revisor-users Okay, I'll check it out. Thanks for the pointer. Eric. From kanarip at kanarip.com Fri Feb 8 10:21:26 2008 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Fri, 08 Feb 2008 11:21:26 +0100 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47AC29EE.9080802@redhat.com> References: <47ABAAC3.4080209@trueblade.com> <47AC284B.90305@kanarip.com> <47AC29EE.9080802@redhat.com> Message-ID: <47AC2D26.4050304@kanarip.com> Joel Andres Granados wrote: > Jeroen van Meeuwen wrote: >> Eric Smith wrote: >>> I'm installing FC6 using kickstart from PXE. It works great, and >>> I've used this (and a similar setup with FC2) for years. >>> >>> Now, I want to use a new kernel during the installation. FC6 comes >>> with 2.6.18-1.2798.fc6, and I want the boot to use some version of >>> 2.6.20 or greater. >>> >>> What I've been doing is installing with the default kernel, then >>> upgrading to 2.6.20-1.2962.fc6 after the installation is complete. >>> But now I have a need to run 2.6.20 during installation, because of >>> some hardware that's supported in 2.6.20 but not in 2.6.18. >>> >>> So my question is: How do I create a new initrd.img and vmlinuz >>> (that came in the isolinux directory) to be a new kernel? It's >>> apparently not as simple as copying these files from a running system. >>> >>> I've spent hours googling. I've seen all sorts of info on adding >>> drivers to initrd.img, but no info on updating the kernel. >>> >>> Any info would be great! >>> >> >> Hi Eric, >> >> -mandatory note about Fedora Core 6 being End Of Life, you should >> upgrade or use some Enterprise Linux (derivative)- >> >> you might want to take a look at Revisor. We've been doing successful >> Fedora Core 6 spins in the past although it really started off with >> Fedora 7... To indicate what is successful, here's the test matrix >> that we commonly use (a Fedora 8 version of it): >> >> http://spins.fedoraunity.org/Members/kanarip/fedora-8-test-matrix-template >> >> >> Revisor you can find in the Fedora repositories (if you run Fedora 7 >> or 8), and there's a package built for EL5 as well. >> >> See: http://revisor.fedoraunity.org and it's mailing list at >> http://lists.fedoraunity.org/mailman/listinfo/revisor-users >> >> Kind regards, >> >> Jeroen van Meeuwen >> -kanarip >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list > > + > Just wanted to mention the different tools > > 1. revisor by Jeroen Meeuwen > 2. pungi by Jessy Keating 3. buildinstall by anaconda team > > AFAK they are in that order. (revisor uses pung and pungi uses > buildinstall). I think if you want an easy way to do this your best bet > is revisor, If you want to get into the core of things you might want > to give buildinstall a try. It a very scarry shell script :) > > Hope it helps. > I don't think pungi did anything for Fedora 7, nor from before Fedora 7. Also, Revisor does not use pungi. Kind regards, Jeroen van Meeuwen -kanarip From jgranado at redhat.com Fri Feb 8 10:31:54 2008 From: jgranado at redhat.com (Joel Andres Granados) Date: Fri, 08 Feb 2008 11:31:54 +0100 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47AC2D26.4050304@kanarip.com> References: <47ABAAC3.4080209@trueblade.com> <47AC284B.90305@kanarip.com> <47AC29EE.9080802@redhat.com> <47AC2D26.4050304@kanarip.com> Message-ID: <47AC2F9A.1060900@redhat.com> Jeroen van Meeuwen wrote: > Joel Andres Granados wrote: >> Jeroen van Meeuwen wrote: >>> Eric Smith wrote: >>>> I'm installing FC6 using kickstart from PXE. It works great, and >>>> I've used this (and a similar setup with FC2) for years. >>>> >>>> Now, I want to use a new kernel during the installation. FC6 comes >>>> with 2.6.18-1.2798.fc6, and I want the boot to use some version of >>>> 2.6.20 or greater. >>>> >>>> What I've been doing is installing with the default kernel, then >>>> upgrading to 2.6.20-1.2962.fc6 after the installation is complete. >>>> But now I have a need to run 2.6.20 during installation, because of >>>> some hardware that's supported in 2.6.20 but not in 2.6.18. >>>> >>>> So my question is: How do I create a new initrd.img and vmlinuz >>>> (that came in the isolinux directory) to be a new kernel? It's >>>> apparently not as simple as copying these files from a running system. >>>> >>>> I've spent hours googling. I've seen all sorts of info on adding >>>> drivers to initrd.img, but no info on updating the kernel. >>>> >>>> Any info would be great! >>>> >>> >>> Hi Eric, >>> >>> -mandatory note about Fedora Core 6 being End Of Life, you should >>> upgrade or use some Enterprise Linux (derivative)- >>> >>> you might want to take a look at Revisor. We've been doing successful >>> Fedora Core 6 spins in the past although it really started off with >>> Fedora 7... To indicate what is successful, here's the test matrix >>> that we commonly use (a Fedora 8 version of it): >>> >>> http://spins.fedoraunity.org/Members/kanarip/fedora-8-test-matrix-template >>> >>> >>> Revisor you can find in the Fedora repositories (if you run Fedora 7 >>> or 8), and there's a package built for EL5 as well. >>> >>> See: http://revisor.fedoraunity.org and it's mailing list at >>> http://lists.fedoraunity.org/mailman/listinfo/revisor-users >>> >>> Kind regards, >>> >>> Jeroen van Meeuwen >>> -kanarip >>> >>> _______________________________________________ >>> Kickstart-list mailing list >>> Kickstart-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/kickstart-list >> >> + >> Just wanted to mention the different tools >> >> 1. revisor by Jeroen Meeuwen >> 2. pungi by Jessy Keating 3. buildinstall by anaconda team >> >> AFAK they are in that order. (revisor uses pung and pungi uses >> buildinstall). I think if you want an easy way to do this your best >> bet is revisor, If you want to get into the core of things you might >> want to give buildinstall a try. It a very scarry shell script :) >> >> Hope it helps. >> > > I don't think pungi did anything for Fedora 7, nor from before Fedora 7. I'm pretty sure pungi began in F7. not sure if it was used to compose f7 or f8 though. have to ask jessey. > > Also, Revisor does not use pungi. mmm. my bad. where you using pungi and changed or did you never use pungi at all? I was convinced that how it worked. learn something new every day. So do you call buildinstall directly to build the images or do you have your own version of buildinstall? > > Kind regards, > > Jeroen van Meeuwen > -kanarip > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- Joel Andres Granados Red Hat / Brno, Czech Republic From kanarip at kanarip.com Fri Feb 8 11:25:17 2008 From: kanarip at kanarip.com (Jeroen van Meeuwen) Date: Fri, 08 Feb 2008 12:25:17 +0100 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47AC2F9A.1060900@redhat.com> References: <47ABAAC3.4080209@trueblade.com> <47AC284B.90305@kanarip.com> <47AC29EE.9080802@redhat.com> <47AC2D26.4050304@kanarip.com> <47AC2F9A.1060900@redhat.com> Message-ID: <47AC3C1D.7010301@kanarip.com> Joel Andres Granados wrote: > Jeroen van Meeuwen wrote: >> I don't think pungi did anything for Fedora 7, nor from before Fedora 7. > > I'm pretty sure pungi began in F7. not sure if it was used to compose f7 > or f8 though. > have to ask jessey. > pungi was used to compose F7 and F8 and will be used to compose F9 as well, but Jesse is convinced he can't get away with composing Fedora N+1 or Fedora N-1 on a Fedora N machine. Whether it is a support-, release engineering- or technical argument I do not know. >> >> Also, Revisor does not use pungi. > > mmm. my bad. where you using pungi and changed or did you never use > pungi at all? > I was convinced that how it worked. learn something new every day. > So do you call buildinstall directly to build the images or do you have > your own version of buildinstall? > We use anaconda-runtime's buildinstall alright. As much as we look like we're going renegade we do value upstream (and so it seems we value upstream more then they do downstream). Basically where we doing what pungi has been doing for Fedora 7, but by the time we started needing callbacks to be performed to keep our GUI responsive -applies to CLI as well, tweak the media being composed and needed to ship Revisor for EL5, we did not get the patches accepted upstream (out-of-scope, zero confidence, whatever other reason) although sometimes the functionality we requested has been applied later -anyway we didn't need to depend on pungi anymore. Kind regards, Jeroen van Meeuwen -kanarip From crhea at mayo.edu Fri Feb 8 14:52:58 2008 From: crhea at mayo.edu (Cris Rhea) Date: Fri, 8 Feb 2008 08:52:58 -0600 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208101444.7965E7325E@hormel.redhat.com> References: <20080208101444.7965E7325E@hormel.redhat.com> Message-ID: <20080208145258.GC9375@kaizen.mayo.edu> > From: Eric Smith > Date: Thu, 07 Feb 2008 20:50:09 -0500 > Subject: Re: Upgrading kernel used in PXE boot > > Eric Smith wrote: > >So my question is: How do I create a new initrd.img and vmlinuz (that > >came in the isolinux directory) to be a new kernel? It's apparently not > >as simple as copying these files from a running system. > > And yes, I realize this might not be the best list for this question, > but it's the best I could think of, since people here are no doubt > familiar with the issue. If you know of a better venue for this > question, please let me know. > > Eric. I haven't used the tools the others have suggested, but since you're in a PXEboot environment, yes, it can be as simple as using the vmlinux and initrd.img file from a patched/running system... The tough part is always making the ramdisk image-- getting all the correct drivers included... you seem to have figured this part out already... For example, in my situation (fighting an Anaconda bug), I boot an older release, then upgrade at the end of the install process: default linux label linux kernel vmlinuz-4.4vault append initrd=initrd.img-4.4vault load_ramdisk=1 ksdevice=eth0 ks=ftp://clustersw/ks/ks-wsnode-4.4vault.cfg I used to do this all the time when people built monolithic kernels to handle new/oddball hardware. Now, you just need to make sure the right modules are available in the ramdisk image. There may be other gotchas along the way (it may complain that the booted kernel doesn't match the distribution), but if your goal is to have a new kernel booted to start the install process, this should do it... HTH. --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From eric at trueblade.com Fri Feb 8 15:13:25 2008 From: eric at trueblade.com (Eric Smith) Date: Fri, 08 Feb 2008 10:13:25 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208145258.GC9375@kaizen.mayo.edu> References: <20080208101444.7965E7325E@hormel.redhat.com> <20080208145258.GC9375@kaizen.mayo.edu> Message-ID: <47AC7195.2080106@trueblade.com> Cris Rhea wrote: >> From: Eric Smith >> Date: Thu, 07 Feb 2008 20:50:09 -0500 >> Subject: Re: Upgrading kernel used in PXE boot >> >> Eric Smith wrote: >>> So my question is: How do I create a new initrd.img and vmlinuz (that >>> came in the isolinux directory) to be a new kernel? It's apparently not >>> as simple as copying these files from a running system. >> And yes, I realize this might not be the best list for this question, >> but it's the best I could think of, since people here are no doubt >> familiar with the issue. If you know of a better venue for this >> question, please let me know. >> >> Eric. > > I haven't used the tools the others have suggested, but since you're in a > PXEboot environment, yes, it can be as simple as using the vmlinux and > initrd.img file from a patched/running system... > > The tough part is always making the ramdisk image-- getting all the correct > drivers included... you seem to have figured this part out already... Actually, the hard part is getting all of the drivers together! But with the information below, I think I can slog through it. > For example, in my situation (fighting an Anaconda bug), I boot an older release, > then upgrade at the end of the install process: > > default linux > label linux > kernel vmlinuz-4.4vault > append initrd=initrd.img-4.4vault load_ramdisk=1 ksdevice=eth0 ks=ftp://clustersw/ks/ks-wsnode-4.4vault.cfg > > I used to do this all the time when people built monolithic kernels to handle new/oddball > hardware. Now, you just need to make sure the right modules are available in the > ramdisk image. > > There may be other gotchas along the way (it may complain that the booted kernel doesn't > match the distribution), but if your goal is to have a new kernel booted to start > the install process, this should do it... > > HTH. That does help, Cris. I've been unable to find a FC6 respin (which I'd rather use, as I think the default driver set with the new kernel will work for me), so if I can get this procedure working, I'll be all set. Thanks! Eric. From crhea at mayo.edu Fri Feb 8 17:30:34 2008 From: crhea at mayo.edu (Cris Rhea) Date: Fri, 8 Feb 2008 11:30:34 -0600 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208170016.8C8E073557@hormel.redhat.com> References: <20080208170016.8C8E073557@hormel.redhat.com> Message-ID: <20080208173034.GC9809@kaizen.mayo.edu> > From: Eric Smith > Subject: Re: Upgrading kernel used in PXE boot > > That does help, Cris. I've been unable to find a FC6 respin (which I'd > rather use, as I think the default driver set with the new kernel will > work for me), so if I can get this procedure working, I'll be all set. > Thanks! > > Eric. I've given up using Fedora for anything server-related. Once they hit the next release, they never look back (and the groups who were keeping up the patches for the older releases ran out of volunteers). I have FC8 on a laptop in our computer room (using it as a console/display). Every time I go in there, it has at least a dozen updates that need to be applied. IMHO, FC8 isn't what I would call "stable", yet they are already talking about Alpha releases of FC9 on the lists. FC6 is very stable, but as you've noticed, you get flak as it is EOL. Fedora is great if you want to play with cutting-edge packages, but the lifespan of any release is only about 6 months. I'm not trying to start a war on which distribution is better, but I found a lot more stability by using RHEL/CentOS for servers. I still use Fedora for desktops -- where re-imaging the system isn't that big of a deal and I may need the cutting-edge versions to handle new hardware. Just my $0.02. --- Cris BTW: What hardware are you trying to get running under FC6? -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From dwight at supercomputer.org Fri Feb 8 17:26:05 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Fri, 8 Feb 2008 09:26:05 -0800 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <20080207175852.GD5715@kaizen.mayo.edu> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802070924.53497.dwight@supercomputer.org> <20080207175852.GD5715@kaizen.mayo.edu> Message-ID: <200802080926.05503.dwight@supercomputer.org> On Thursday 07 February 2008 09:58:52 am Cris Rhea wrote: > On Thu, Feb 07, 2008 at 09:24:53AM -0800, dwight at supercomputer.org wrote: > > Here's an obscure bit of trivia. Why is it not one, not two, not > > four, but three syncs? > > Because the system will have done the first sync by the time you > finished typing the other two. What if you're a fast typer? As in using an alias or shell function? In over 30 years of dealing with *NIX, I've never met anyone who could give an empirical answer (though many sound reasonable), or who would (note I didn't say "could") show via the code as to why it's three syncs, not two, or four+. IMHO, it's lore based upon actual experience, because it has matched my experience as well. Even with modern Linux kernels and very slow IO devices (though the Linux kernel is certainly far better in this regard). Anyway, pardon the digression. Hopefully it was amusing. > Yup... perhaps anaconda is getting confused by the number of > "onpart" file systems I'm defining or something equally as > bizarre. I'm surprised to hear that anaconda isn't handling the onpart directive correctly, with a pre-existing partition table. I would think that more people would see it. One has to wonder if anaconda's defaults are somehow kicking in and getting in the way? > How does one build a minstg2.img file? Can I just change *that* > one file or do I also have to change more (i.e., doesn't Anaconda > call all sorts of Python libs located elsewhere)? Well, since no one else is answering, I'll supply what I know about it. If someone with more expertise in this area cares to comment or correct, please do so. It would be appreciated. >From what I've grokked from looking at the images and source code, the first starting point would be the (say) PXE boot initrd that you've set up, which came from the CD/DVD image under the images/pxeboot directory. If you unroll the initrd image, say via: mkdir root; cd root; zcat .../initrd.img | cpio -iv then you'll see a few things of interest. The first is that there's not much there. Mostly it's just /sbin/init, /sbin/loader and several symlinks to the latter. I assume that init calls loader directly, as strings(1) explicitly shows "/sbin/loader". The loader program appears to come from the loader2 directory in the anaconda source code package itself. This is the next thing that you want to look at, and it's full of all sorts of goodies. Just grab the anaconda source rpm, install it, build it via `rpmbuild -bp` and you'll have the source code. One interesting thing, as per your question, is that it loads in a second image, the minstg2 image. Presumably this is where anaconda is kept, as /sbin/loader will fork(2)/exec(2) anaconda. Now here's an interesting point. The loader code does a waitpid(2) after the fork(2) without closing any file descriptors. So, if a disk has somehow been opened, then the child process (anaconda) won't be able to do a final close on the device. I didn't see anything obvious, but some of that code appears to call kudzu-related functions(!), and I have to wonder if something there is keeping any devices open in the parent process. To muck with the minstg2 image, you need to be running on a squashfs-capable kernel. Here's a link: http://squashfs.sourceforge.net/ With that, you should be able to mount up the minstg2 image on a loopback device and play around with it. Alternatively, if you look at the loader source code, you'll see that it just expects to find /usr/bin/anaconda, so I would think that you could keep it all on one single initrd image. Note that the minstg2 image is 31 MB, and there used to be a problem with loading large initrd images. But that doesn't seem to be the case anymore. I too am a little surprised that this isn't documented better. Would there be any interest in a Wiki page on this? It might help getting anaconda bugs fixed faster, if people knew what to do. -dwight- From Matt.Fahrner at coat.com Fri Feb 8 18:24:43 2008 From: Matt.Fahrner at coat.com (Matt Fahrner) Date: Fri, 08 Feb 2008 13:24:43 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208173034.GC9809@kaizen.mayo.edu> References: <20080208170016.8C8E073557@hormel.redhat.com> <20080208173034.GC9809@kaizen.mayo.edu> Message-ID: <47AC9E6B.8020202@coat.com> Personally I just the last 2 days upgraded my best friend and mine's 1and1 server from FC6 to FC8 via "yum" (it started at FC4). I did this all remotely with very little pain. We've found it (Fedora) to be extremely stable as a server and quite easy to keep up to date with "yum". I've only once or twice had an issue which required tweaking after a "yum" update, and even then they were relatively minor. I don't know if I'd recommend it for a server farm (at Burlington Coat Factory where I work, we use SuSE mostly), but I've been pretty impressed for an Internet facing web server. - Matt Cris Rhea wrote: >> From: Eric Smith >> Subject: Re: Upgrading kernel used in PXE boot >> >> That does help, Cris. I've been unable to find a FC6 respin (which I'd >> rather use, as I think the default driver set with the new kernel will >> work for me), so if I can get this procedure working, I'll be all set. >> Thanks! >> >> Eric. >> > > I've given up using Fedora for anything server-related. Once they hit > the next release, they never look back (and the groups who were keeping up > the patches for the older releases ran out of volunteers). > > I have FC8 on a laptop in our computer room (using it as a console/display). > Every time I go in there, it has at least a dozen updates that need to > be applied. IMHO, FC8 isn't what I would call "stable", yet they are > already talking about Alpha releases of FC9 on the lists. > > FC6 is very stable, but as you've noticed, you get flak as it is EOL. > > Fedora is great if you want to play with cutting-edge packages, but the > lifespan of any release is only about 6 months. > > I'm not trying to start a war on which distribution is better, but I found > a lot more stability by using RHEL/CentOS for servers. I still use Fedora > for desktops -- where re-imaging the system isn't that big of a deal > and I may need the cutting-edge versions to handle new hardware. > > Just my $0.02. > > --- Cris > > BTW: What hardware are you trying to get running under FC6? > > -- > Cristopher J. Rhea > Mayo Clinic - Research Computing Facility > 200 First St SW, Rochester, MN 55905 > crhea at Mayo.EDU > (507) 284-0587 > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- --------------------------------------------------------------------- Matt Fahrner 2 South Park St. Chief Systems Architect Willis House Burlington Coat Factory Warehouse Lebanon, N.H. 03766 Tel: (603) 448-4100 x5150 USA Fax: (603) 443-6190 Matt.Fahrner at COAT.COM --------------------------------------------------------------------- -------------- next part -------------- An HTML attachment was scrubbed... URL: From douglas at hubler.us Fri Feb 8 19:00:00 2008 From: douglas at hubler.us (Douglas Hubler) Date: Fri, 8 Feb 2008 19:00:00 +0000 (UTC) Subject: error using kickstart file References: Message-ID: Douglas Hubler hubler.us> writes: > I'm getting this error from anaconda 11.3.0.50 doing a kickstart install ... > > Stack Trace > > File "/usr/bin/anaconda", line 917, in > kickstart.processKickstartFile(anaconda, opts.ksfile) > > ... (omitted) > > File "/usr/lib/python2.5/site-packages/pykickstart/commands/method.py", line > 81, in parse > (opts, extra) = op.parse_args(args=args) > AttributeError: 'NoneType' object has no attribute 'parse_args' > install exited abnormally [1/1] ... > I get this error with every incarnation of kickstart files. Here's > the simplest one I have > ----------------------------- > install > cdrom > lang en_US.UTF-8 > keyboard us > > %packages > @core > anaconda-runtime > iscsi-initiator-utils > memtest86+ > vnc-server > %end I sort of found the problem, in the kickstart code that get's bundled into CD respin simply cannot read the line cdrom so I created an updates.img from the kickstart code i get when i update my own f8 workstation by doing the following dd if=/dev/zero of=updates.img bs=1k count=2880 /sbin/mke2fs -F updates.img sudo mount -o loop updates.img updates cd updates cp -r /usr/lib/python2.5/site-packages/pykickstart . Then copy updates.img into CD respin into [cdrom]/images/updates.img I included the entire kickstart code, because I don't know where the exact error is and even if I did, who knows what other errors I'd introduce if i picked one source code file from one rev and mixed it with source code from a different rev Questions 1.) Anyone see any flaws in my logic? 2.) This bug was obviously fixed in kickstart, if someone knows the patch or even the bug db and source code repository so I can look it up myself, I'd appreciate it. Thanks From mdehaan at redhat.com Fri Feb 8 19:10:21 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 08 Feb 2008 14:10:21 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208173034.GC9809@kaizen.mayo.edu> References: <20080208170016.8C8E073557@hormel.redhat.com> <20080208173034.GC9809@kaizen.mayo.edu> Message-ID: <47ACA91D.6030908@redhat.com> Cris Rhea wrote: >> From: Eric Smith >> Subject: Re: Upgrading kernel used in PXE boot >> >> That does help, Cris. I've been unable to find a FC6 respin (which I'd >> rather use, as I think the default driver set with the new kernel will >> work for me), so if I can get this procedure working, I'll be all set. >> Thanks! >> >> Eric. >> > > I've given up using Fedora for anything server-related. Once they hit > the next release, they never look back (and the groups who were keeping up > the patches for the older releases ran out of volunteers). > > I have FC8 on a laptop in our computer room (using it as a console/display). > Every time I go in there, it has at least a dozen updates that need to > be applied. IMHO, FC8 isn't what I would call "stable", yet they are > already talking about Alpha releases of FC9 on the lists. > > FC6 is very stable, but as you've noticed, you get flak as it is EOL. > > Fedora is great if you want to play with cutting-edge packages, but the > lifespan of any release is only about 6 months. > > I'm not trying to start a war on which distribution is better, but I found > a lot more stability by using RHEL/CentOS for servers. I still use Fedora > for desktops -- where re-imaging the system isn't that big of a deal > and I may need the cutting-edge versions to handle new hardware. > > Just my $0.02. > > --- Cris > > BTW: What hardware are you trying to get running under FC6? > > Don't confuse stable with "updates frequently". "Updates frequently" is how we get work done -- which all of that work obviously rolls into long-term-supported distros like EL4/5. From crhea at mayo.edu Fri Feb 8 19:41:13 2008 From: crhea at mayo.edu (Cris Rhea) Date: Fri, 8 Feb 2008 13:41:13 -0600 Subject: Old style fixed partitions in KS-- Anaconda bug?? In-Reply-To: <200802080926.05503.dwight@supercomputer.org> References: <20080204214332.GB26448@kaizen.mayo.edu> <200802070924.53497.dwight@supercomputer.org> <20080207175852.GD5715@kaizen.mayo.edu> <200802080926.05503.dwight@supercomputer.org> Message-ID: <20080208194113.GB10326@kaizen.mayo.edu> On Fri, Feb 08, 2008 at 09:26:05AM -0800, dwight at supercomputer.org wrote: > I too am a little surprised that this isn't documented better. > Would there be any interest in a Wiki page on this? It might > help getting anaconda bugs fixed faster, if people knew what to do. > > -dwight- It gets even more messy... just looking at the trace shows that it's bombing coming back from partedmodule (which is a Python Shared Object Library). That takes it back to 7 C programs (about 2200 lines). You'd almost have to compare all the revisions in CVS to find what changed between 4.4 and 4.5 (and figure out exactly what versions of Anaconda, Python and the Shared Library were in each release). Doable, but looks like lots of hours... It would sure be nice to find one of the Anaconda developers-- they could probably go right to the place in the code... Sigh! --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From crhea at mayo.edu Fri Feb 8 20:11:30 2008 From: crhea at mayo.edu (Cris Rhea) Date: Fri, 8 Feb 2008 14:11:30 -0600 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208194151.C66A372E22@hormel.redhat.com> References: <20080208194151.C66A372E22@hormel.redhat.com> Message-ID: <20080208201130.GA14310@kaizen.mayo.edu> > From: Michael DeHaan > Date: Fri, 08 Feb 2008 14:10:21 -0500 > Subject: Re: Upgrading kernel used in PXE boot > > Don't confuse stable with "updates frequently". > > "Updates frequently" is how we get work done -- which all of that work > obviously rolls into long-term-supported distros like EL4/5. What I was suggesting was that "updates frequently" (whatever term you would like) isn't what's needed in most cases for *servers*. For a *server*, I don't want to apply updates every few days-- sometimes needing a reboot. I'm not bashing Fedora (nor trying to pick a fight with folks who have "redhat.com" on the right side of their email address)-- but I do believe it is correct to say that Fedora is a very fast-paced environment with relatively short-lived releases. If you only wish to reload the OS on your servers every couple years, then (IMHO) Fedora is a poor choice-- RHEL/CentOS would be a much better choice because of the "long-term-supported distro[s]". --- Cris -- Cristopher J. Rhea Mayo Clinic - Research Computing Facility 200 First St SW, Rochester, MN 55905 crhea at Mayo.EDU (507) 284-0587 From Matt.Fahrner at coat.com Fri Feb 8 20:40:29 2008 From: Matt.Fahrner at coat.com (Matt Fahrner) Date: Fri, 08 Feb 2008 15:40:29 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208201130.GA14310@kaizen.mayo.edu> References: <20080208194151.C66A372E22@hormel.redhat.com> <20080208201130.GA14310@kaizen.mayo.edu> Message-ID: <47ACBE3D.4010306@coat.com> Cris Rhea wrote: > If you only wish to reload the OS on your servers every couple years, then > (IMHO) Fedora is a poor choice-- RHEL/CentOS would be a much better > choice because of the "long-term-supported distro[s]". > While I agree Fedora is incredibly fast moving - with updates daily and very short cycles between "official" releases, in some ways I think you're comparing apples to oranges. Unlike RHE, SuSE, and others, you don't have to reinstall the server from scratch or do a CD based single user install to upgrade from release to release. It can be upgraded inline in multi-user via yum and done so in an a relatively non-intrusive way. I have no doubt in a "cookie cutter" installation it could be automated. In the end the version number of Fedora is a bit arbitrary and it can almost be seen as a continuum of a single constantly updated OS (again, I have a server that's gone from Fedora Core 4 to Fedora 8 very seamlessly). I see it as being more incremental than forklift upgrade. On the other hand, while I think it isn't a bad choice for a rapid development model where you want the latest and greatest, I wouldn't recommend Fedora for most server environments. If nothing else this is true because vendors like Oracle require quantifiable releases for support and Fedora is too fluid and not certified. That said, I have seen long waits for important fixes or dependencies out of the "stable" but supposedly "supported" releases that can be quite frustrating. I can't name the number of times we've been roadblocked on fixing a major issue by the fact that something like Oracle requires us to stay with an older but supposedly "stable" OS configuration. - Matt -- --------------------------------------------------------------------- Matt Fahrner 2 South Park St. Chief Systems Architect Willis House Burlington Coat Factory Warehouse Lebanon, N.H. 03766 Tel: (603) 448-4100 x5150 USA Fax: (603) 443-6190 Matt.Fahrner at COAT.COM --------------------------------------------------------------------- From mdehaan at redhat.com Fri Feb 8 22:14:23 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Fri, 08 Feb 2008 17:14:23 -0500 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208201130.GA14310@kaizen.mayo.edu> References: <20080208194151.C66A372E22@hormel.redhat.com> <20080208201130.GA14310@kaizen.mayo.edu> Message-ID: <47ACD43F.9060802@redhat.com> > I'm not bashing Fedora (nor trying to pick a fight with folks who > have "redhat.com" on the right side of their email address)-- but I do > believe it is correct to say that Fedora is a very fast-paced environment > with relatively short-lived releases. > mdehaan at fedoraproject.org works too :) From sergio at sergiomb.no-ip.org Sat Feb 9 02:55:44 2008 From: sergio at sergiomb.no-ip.org (Sergio Monteiro Basto) Date: Sat, 09 Feb 2008 02:55:44 +0000 Subject: Upgrading kernel used in PXE boot In-Reply-To: <20080208173034.GC9809@kaizen.mayo.edu> References: <20080208170016.8C8E073557@hormel.redhat.com> <20080208173034.GC9809@kaizen.mayo.edu> Message-ID: <1202525744.4408.33.camel@localhost.localdomain> On Fri, 2008-02-08 at 11:30 -0600, Cris Rhea wrote: > I have FC8 on a laptop in our computer room (using it as a > console/display). > Every time I go in there, it has at least a dozen updates that need to > be applied. you don't need and shouldn't update system every day, it only waste of time waiting to update the system. I update my fedoras 2, 3 times a month, When I build my respin iso with pungi. > IMHO, FC8 isn't what I would call "stable", yet they are > already talking about Alpha releases of FC9 on the lists. IMHO it is very stable, about life time of a release be 6 months and EOL 1 year. I also sometimes thinks that I would prefer a little longer versions. but sometimes not. I enjoy very much, the new NetworkManager, the new java and without a new version that wasn't possible. you a new spin of F7 made by fedoraunity team you may try: http://spins.fedoraunity.org/unity/fedora-unity-re-spin-f7-20080118 is the oldest one that I found. But you always have the change to do your FC6 re-spin with pungi, I think :) Regards, -- S?rgio M. B. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 2192 bytes Desc: not available URL: From sergio at sergiomb.no-ip.org Sat Feb 9 03:05:02 2008 From: sergio at sergiomb.no-ip.org (Sergio Monteiro Basto) Date: Sat, 09 Feb 2008 03:05:02 +0000 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47ABB551.1020509@trueblade.com> References: <47ABAAC3.4080209@trueblade.com> <47ABB551.1020509@trueblade.com> Message-ID: <1202526302.4408.38.camel@localhost.localdomain> On Thu, 2008-02-07 at 20:50 -0500, Eric Smith wrote: > And yes, I realize this might not be the best list for this question, yes ! , we are in kickstart list , I though we are on Discussion of Fedora build system I think that lis is more appropriated. I am going post there configuration for building re-spins on f7 and f8 , could work on FC 6 regards, -- S?rgio M. B. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 2192 bytes Desc: not available URL: From sergio at sergiomb.no-ip.org Fri Feb 8 02:40:39 2008 From: sergio at sergiomb.no-ip.org (Sergio Monteiro Basto) Date: Fri, 08 Feb 2008 02:40:39 +0000 Subject: Upgrading kernel used in PXE boot In-Reply-To: <47ABAAC3.4080209@trueblade.com> References: <47ABAAC3.4080209@trueblade.com> Message-ID: <1202438439.2848.8.camel@monteirov> On Thu, 2008-02-07 at 20:05 -0500, Eric Smith wrote: > So my question is: How do I create a new initrd.img and vmlinuz > (that > came in the isolinux directory) to be a new kernel? It's apparently > not > as simple as copying these files from a running system. Pungi do that ... isolinux/vmlinuz and isolinux/initrd.img Regards, -- S?rgio M.B. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 2192 bytes Desc: not available URL: From arifsaha at yahoo.com Mon Feb 11 23:09:58 2008 From: arifsaha at yahoo.com (S P Arif Sahari Wibowo) Date: Mon, 11 Feb 2008 18:09:58 -0500 (EST) Subject: Kickstart Installation using Existing LVM Logical Volume In-Reply-To: <47A1EB44.8080207@redhat.com> References: <47A1EB44.8080207@redhat.com> Message-ID: On Thu, 31 Jan 2008, Michael DeHaan wrote: > Have you looked at generating a sample kickstart using > system-config-kickstart and using something derived from that > one? Does not seem can be done. The GUI have no option to setup LVM volumes, and when I opened an existing file with LVM volumes and tried to edit it, the GUI crashed. -- (stephan paul) Arif Sahari Wibowo _____ _____ _____ _____ /____ /____/ /____/ /____ _____/ / / / _____/ http://www.arifsaha.com/ Xinnian Kuaile! ???? Gongxi Facai ???? From jgranado at redhat.com Tue Feb 12 09:51:59 2008 From: jgranado at redhat.com (Joel Andres Granados) Date: Tue, 12 Feb 2008 10:51:59 +0100 Subject: Ignore swap devices on install Message-ID: <47B16C3F.1080802@redhat.com> Hi list: It has been a problem to tell anaconda to not use some swap devices at install time. I have created a solution (if you can call it that) that adds a argument for kickstart in RHEL5. It basically tells anaconda : "Do not touch this list of swap devices". And anaconda does not use them at install time and does not add them to /etc/fstab at the end of installation. The solution involves code for anaconda and for pykickstart. Patch is attached. comments greatly appreciated. -- Joel Andres Granados Red Hat / Brno, Czech Republic -------------- next part -------------- A non-text attachment was scrubbed... Name: ignoredevA.diff Type: text/x-patch Size: 2693 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: ignoredevK.diff Type: text/x-patch Size: 2721 bytes Desc: not available URL: From jgranado at redhat.com Tue Feb 12 12:14:17 2008 From: jgranado at redhat.com (Joel Andres Granados) Date: Tue, 12 Feb 2008 13:14:17 +0100 Subject: Ignore swap devices on install In-Reply-To: <47B16C3F.1080802@redhat.com> References: <47B16C3F.1080802@redhat.com> Message-ID: <47B18D99.2040901@redhat.com> Joel Andres Granados wrote: > Hi list: > > It has been a problem to tell anaconda to not use some swap devices at > install time. I have created a solution (if you can call it that) that > adds a argument for kickstart in RHEL5. It basically tells anaconda : > "Do not touch this list of swap devices". And anaconda does not use > them at install time and does not add them to /etc/fstab at the end of > installation. > The solution involves code for anaconda and for pykickstart. > > Patch is attached. comments greatly appreciated. > > > ------------------------------------------------------------------------ > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list mmm, forgot to mention: bug number (218929) or (70477) -- Joel Andres Granados Red Hat / Brno, Czech Republic From arifsaha at yahoo.com Wed Feb 13 13:43:01 2008 From: arifsaha at yahoo.com (S P Arif Sahari Wibowo) Date: Wed, 13 Feb 2008 08:43:01 -0500 (EST) Subject: Kickstart Installation using Existing LVM Logical Volume In-Reply-To: References: Message-ID: Hi! I guess in my previous posting people have difficulties find the fact, so here I try again. I am working on a Fedora 8 kickstart installation on multiple (actually just 2) drive. It should use existing LVM logical volume on one drive as home partition, while it should clear and repartition the other drive. However, it keep failing with an error like this: > Could not allocate requested partitions: > > Adding this partition would not leave enough disk space > for already allocated logical volumes in 2008AF_data.. > > Press 'OK' to exit the installer. * 2008AF_data is the name of volume group The kickstart code dealing with partition is quoted below: clearpart --all --drives=sda part /boot --fstype ext3 --size=128 part pv.1 --size=1536 --grow --ondisk=sda volgroup 2008A6_work pv.1 logvol swap --fstype swap --name=2008A6_swap --vgname=2008A6_work --size=512 logvol /var --fstype ext3 --name=2008A6_var --vgname=2008A6_work --size=512 logvol / --fstype ext3 --name=2008A6_root --vgname=2008A6_work --size=3904 --grow part pv.2 --noformat --onpart sdb1 volgroup 2008AF_data pv.2 --noformat --useexisting logvol /home --noformat --useexisting --name=2008AF_home --vgname=2008AF_data --fstype ext3 --size-1024 --grow * 2008AF_home is the name of logical volume to be reused I already tried without --useexisting, without size and --grow, without --fstype, etc. same error. BTW, using kickstart GUI does not help, since apparently the GUI does not support LVM volumes. Any idea what's wrong? Thanks! -- (stephan paul) Arif Sahari Wibowo _____ _____ _____ _____ /____ /____/ /____/ /____ _____/ / / / _____/ http://www.arifsaha.com/ Xinnian Kuaile! ???? Gongxi Facai ???? From schwenk at math.udel.edu Wed Feb 13 14:55:13 2008 From: schwenk at math.udel.edu (Peter Schwenk) Date: Wed, 13 Feb 2008 09:55:13 -0500 Subject: Unable to eject boot cd during kickstart install Message-ID: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> I've been using kickstart installs with Fedora and CentOS for quite a while now. I boot the system with a cd and enter the appropriate command at the BOOT: prompt and away it goes. I've gotten used to being able to eject the cd from the system (using the button on the drive) at the point where it is installing packages. For some reason, with Fedora 8 kickstart installs the cd is locked in the drive and I cannot eject it. I've tried going to a virtual terminal and running 'eject /tmp/cdrom', but it returns "Input/output error". I'd like to be able to take the cd with me so that I don't have to go and retrieve it later. Does anyone know how this can be done? Thanks in advance for your help. ----- - Peter Schwenk - CITA-3, Systems Administrator - Mathematical Sciences - University of Delaware - schwenk _at_ math _dot_ udel _dot_ edu - http://www.math.udel.edu/~schwenk -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2425 bytes Desc: not available URL: From dwight at supercomputer.org Wed Feb 13 16:47:45 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Wed, 13 Feb 2008 08:47:45 -0800 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> Message-ID: <200802130847.45928.dwight@supercomputer.org> It sounds like there's some process left around which is keeping the device open. That would be the first thing I'd check. If nothing is obvious from a ps(1), then you'll have to dig deeper. lsof(1) and/or an `fuser -v` are typically used to track this down. The trouble is that these aren't available from the F2 virtual terminal. I'm going from memory at the moment, but I'm pretty certain this is the case. To use these utilities, you'll have to do something like the following: 1. chroot /mnt/system 2. mount up /sys (Check fstab, it would be something like mount -t sysfs /sys 3. mount up /proc 4. /sbin/start_udev You should then be able to use lsof and/or fuser. You can also now do all sorts of other nifty things. Note that to exit this environment cleanly, you have to clean up in the reverse order. Starting will killing the udevd daemon. Hope that helps. -dwight- On Wednesday 13 February 2008 06:55:13 am Peter Schwenk wrote: > I've been using kickstart installs with Fedora and CentOS for > quite a while now. I boot the system with a cd and enter the > appropriate command at the BOOT: prompt and away it goes. I've > gotten used to being able to eject the cd from the system (using > the button on the drive) at the point where it is installing > packages. For some reason, with Fedora 8 kickstart installs the > cd is locked in the drive and I cannot eject it. I've tried going > to a virtual terminal and running 'eject /tmp/cdrom', but it > returns "Input/output error". > > I'd like to be able to take the cd with me so that I don't have to > go and retrieve it later. Does anyone know how this can be done? > Thanks in advance for your help. > > > ----- > - Peter Schwenk > - CITA-3, Systems Administrator > - Mathematical Sciences > - University of Delaware > - schwenk _at_ math _dot_ udel _dot_ edu > - http://www.math.udel.edu/~schwenk From jgranado at redhat.com Wed Feb 13 17:27:26 2008 From: jgranado at redhat.com (Joel Andres Granados) Date: Wed, 13 Feb 2008 18:27:26 +0100 Subject: Ignore swap devices on install In-Reply-To: <200802130853.35261.dwight@supercomputer.org> References: <47B16C3F.1080802@redhat.com> <200802130853.35261.dwight@supercomputer.org> Message-ID: <47B3287E.9060102@redhat.com> dwight at supercomputer.org wrote: > This seems rather useful, and is cleanly done. > > What are the odds of getting it into the source tree? > > -dwight- > > On Tuesday 12 February 2008 01:51:59 am Joel Andres Granados wrote: >> Hi list: >> >> It has been a problem to tell anaconda to not use some swap >> devices at install time. I have created a solution (if you can >> call it that) that adds a argument for kickstart in RHEL5. It >> basically tells anaconda : "Do not touch this list of swap >> devices". And anaconda does not use them at install time and does >> not add them to /etc/fstab at the end of installation. The >> solution involves code for anaconda and for pykickstart. >> >> Patch is attached. comments greatly appreciated. > > Hey dwight: This is going in for RHEL, but I want to actually fix the bug in fedora. This would mean ignoring all the swaps that are not stated in the ks files. Anaconda would scream if there is no swap devs for installation time (swap is needed) and will not touch anything unless stated in a partition line. Regards -- Joel Andres Granados Red Hat / Brno, Czech Republic From lists at zadevalov.com Wed Feb 13 22:09:46 2008 From: lists at zadevalov.com (Eugeny Zadevalov) Date: Thu, 14 Feb 2008 00:09:46 +0200 Subject: problems with unpartitioned harddrives and automatic partitioning Message-ID: <47B36AAA.7010501@zadevalov.com> Hello All, I've been trying to setup automatic partitioning with Kickstart. Everything is okay until I try to run my Kickstart setup on machine with unpartitioned drives (like if partition table is broken or filled with zeros). I'm experiencing same issue with both Centos4.6 and Centos5.1. My goal is to get raid1 setup automatically through kickstart. Below is cut from ks.cfg: === zerombr clearpart --initlabel --all part raid.01 --size=2048 --onbiosdisk=80 --asprimary part raid.02 --size=8192 --onbiosdisk=80 --asprimary --grow part raid.03 --size=2048 --onbiosdisk=81 --asprimary part raid.04 --size=8192 --onbiosdisk=81 --asprimary --grow raid / --level=RAID1 --device=md0 --fstype=ext3 raid.02 raid.04 raid swap --level=RAID1 --device=md1 --fstype=swap raid.01 raid.03 === If I remove this part and run kickstarting machine over Kickstart with manual partitioning I get warning about drives needs to be initialized. That's normal I guess. But with my automatic partitioning I've "clearpart --initlabel --all" and that supposed to be initializing any/all unpartitioned drives but that's not happening. Kickstart just hangs giving bunch of errors. If I just go ahead and do something like: fdisk /dev/hda w ENTER fdisk /dev/hdc w ENTER And then try again that automatic partitioning scheme it works like expected. I've searched over the mailling list archive but wasn't able to find anything. For me it looks like a bug with "--initlabel" option. Any suggestions? Thanks! From Chip.Shabazian at bankofamerica.com Thu Feb 14 05:48:06 2008 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Wed, 13 Feb 2008 21:48:06 -0800 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: <47B36AAA.7010501@zadevalov.com> Message-ID: Not sure, but I do --all before --initlabel. Shouldn't make a difference, but I've seen stranger things. The other thing you could probably do is wipe the the partition table in %pre: set $(list-harddrives) cat << EOF >> /tmp/fdisk.input w EOF for i in $d1,$d2,$d3,...(better ways to do this, but this is worth starting with); do fdisk $i < /tmp/fdisk.input; done -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Eugeny Zadevalov Sent: Wednesday, February 13, 2008 2:10 PM To: kickstart-list at redhat.com Subject: problems with unpartitioned harddrives and automatic partitioning Hello All, I've been trying to setup automatic partitioning with Kickstart. Everything is okay until I try to run my Kickstart setup on machine with unpartitioned drives (like if partition table is broken or filled with zeros). I'm experiencing same issue with both Centos4.6 and Centos5.1. My goal is to get raid1 setup automatically through kickstart. Below is cut from ks.cfg: === zerombr clearpart --initlabel --all part raid.01 --size=2048 --onbiosdisk=80 --asprimary part raid.02 --size=8192 --onbiosdisk=80 --asprimary --grow part raid.03 --size=2048 --onbiosdisk=81 --asprimary part raid.04 --size=8192 --onbiosdisk=81 --asprimary --grow raid / --level=RAID1 --device=md0 --fstype=ext3 raid.02 raid.04 raid swap --level=RAID1 --device=md1 --fstype=swap raid.01 raid.03 === If I remove this part and run kickstarting machine over Kickstart with manual partitioning I get warning about drives needs to be initialized. That's normal I guess. But with my automatic partitioning I've "clearpart --initlabel --all" and that supposed to be initializing any/all unpartitioned drives but that's not happening. Kickstart just hangs giving bunch of errors. If I just go ahead and do something like: fdisk /dev/hda w ENTER fdisk /dev/hdc w ENTER And then try again that automatic partitioning scheme it works like expected. I've searched over the mailling list archive but wasn't able to find anything. For me it looks like a bug with "--initlabel" option. Any suggestions? Thanks! _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From jgranado at redhat.com Thu Feb 14 09:10:35 2008 From: jgranado at redhat.com (Joel Andres Granados) Date: Thu, 14 Feb 2008 10:10:35 +0100 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: References: Message-ID: <47B4058B.9040900@redhat.com> Shabazian, Chip wrote: > Not sure, but I do --all before --initlabel. Shouldn't make a > difference, but I've seen stranger things. > > The other thing you could probably do is wipe the the partition table in > %pre: > > set $(list-harddrives) > cat << EOF >> /tmp/fdisk.input > w > > EOF > > for i in $d1,$d2,$d3,...(better ways to do this, but this is worth > starting with); do fdisk $i < /tmp/fdisk.input; done > > > -----Original Message----- > From: kickstart-list-bounces at redhat.com > [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Eugeny Zadevalov > Sent: Wednesday, February 13, 2008 2:10 PM > To: kickstart-list at redhat.com > Subject: problems with unpartitioned harddrives and automatic > partitioning > > Hello All, > > I've been trying to setup automatic partitioning with Kickstart. > Everything is okay until I try to run my Kickstart setup on machine with > unpartitioned drives (like if partition table is broken or filled with > zeros). > > I'm experiencing same issue with both Centos4.6 and Centos5.1. > > My goal is to get raid1 setup automatically through kickstart. Below is > cut from ks.cfg: > === > zerombr > clearpart --initlabel --all > part raid.01 --size=2048 --onbiosdisk=80 --asprimary part raid.02 > --size=8192 --onbiosdisk=80 --asprimary --grow part raid.03 --size=2048 > --onbiosdisk=81 --asprimary part raid.04 --size=8192 --onbiosdisk=81 > --asprimary --grow > raid / --level=RAID1 --device=md0 --fstype=ext3 raid.02 raid.04 > raid swap --level=RAID1 --device=md1 --fstype=swap raid.01 raid.03 === > > If I remove this part and run kickstarting machine over Kickstart with > manual partitioning I get warning about drives needs to be initialized. > That's normal I guess. But with my automatic partitioning I've > "clearpart --initlabel --all" and that supposed to be initializing > any/all unpartitioned drives but that's not happening. Kickstart just > hangs giving bunch of errors. > > If I just go ahead and do something like: > fdisk /dev/hda > w > ENTER > fdisk /dev/hdc > w > ENTER > And then try again that automatic partitioning scheme it works like > expected. > > I've searched over the mailling list archive but wasn't able to find > anything. > > For me it looks like a bug with "--initlabel" option. > > Any suggestions? > > Thanks! > > _______________________________________________ > 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 Cant tell you how many times I hit this problem :) I solved it with a %pre section that initializes the drives. As my needs where very simple, the script is very simple. But you can add to it as your situation requires. %pre parted -s /dev/sdX mklabel msdos %end -- Joel Andres Granados Red Hat / Brno, Czech Republic From debian at herakles.homelinux.org Thu Feb 14 11:50:22 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Thu, 14 Feb 2008 20:50:22 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <200802130847.45928.dwight@supercomputer.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> Message-ID: <47B42AFE.3070504@herakles.homelinux.org> dwight at supercomputer.org wrote: > It sounds like there's some process left around which is keeping the > device open. That would be the first thing I'd check. If nothing is > obvious from a ps(1), then you'll have to dig deeper. I'm going to take this to the Anaconda list, because it illustrates something that's been on my mind for a while. When I boot CD to net install (which I suppose the OP is doing coz otherwise it doesn;t make sense to head for other parts, media in hand), I'd like it to spit the CD out as soon as it's not required. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From debian at herakles.homelinux.org Thu Feb 14 11:57:01 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Thu, 14 Feb 2008 20:57:01 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <200802130847.45928.dwight@supercomputer.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> Message-ID: <47B42C8D.4050500@herakles.homelinux.org> dwight at supercomputer.org wrote: I'm forking a thread on ks, because it illustrates something that's been on my mind for a while. When I boot CD to net install (which I suppose the OP is doing coz otherwise it doesn;t make sense to head for other parts, media in hand), I'd like it to spit the CD out as soon as it's not required. Normally, I use PXE but just sometimes, I might boot from CD and then install off a good ADSL connexion. Taking the CD ensures there won't be an install loop. Waiting for it's tedious. bts I've just been testing PX+KS (CentOS5.1) on a 3yo or so Pentium IV. Takes a little less than 15 minutes, about what RHHL 7.2-3 took on Pentium IIs fdr a similar install. In contrast, XP's equivalent takes around an hour and despite the .SIF file (equivalent to the ks) it insists asking more question. And then it's ready to install the applications such as office software. Good Work JK (who I think's been on it all that time) and crew. > It sounds like there's some process left around which is keeping the > device open. That would be the first thing I'd check. If nothing is > obvious from a ps(1), then you'll have to dig deeper. > > lsof(1) and/or an `fuser -v` are typically used to track this down. > The trouble is that these aren't available from the F2 virtual > terminal. I'm going from memory at the moment, but I'm pretty > certain this is the case. > > To use these utilities, you'll have to do something like the > following: > 1. chroot /mnt/system > 2. mount up /sys (Check fstab, it would be something like > mount -t sysfs /sys > 3. mount up /proc > 4. /sbin/start_udev > > You should then be able to use lsof and/or fuser. > > You can also now do all sorts of other nifty things. > > Note that to exit this environment cleanly, you have to clean up in > the reverse order. Starting will killing the udevd daemon. > > Hope that helps. > > -dwight- > > On Wednesday 13 February 2008 06:55:13 am Peter Schwenk wrote: >> I've been using kickstart installs with Fedora and CentOS for >> quite a while now. I boot the system with a cd and enter the >> appropriate command at the BOOT: prompt and away it goes. I've >> gotten used to being able to eject the cd from the system (using >> the button on the drive) at the point where it is installing >> packages. For some reason, with Fedora 8 kickstart installs the >> cd is locked in the drive and I cannot eject it. I've tried going >> to a virtual terminal and running 'eject /tmp/cdrom', but it >> returns "Input/output error". >> >> I'd like to be able to take the cd with me so that I don't have to >> go and retrieve it later. Does anyone know how this can be done? >> Thanks in advance for your help. >> >> >> ----- >> - Peter Schwenk >> - CITA-3, Systems Administrator >> - Mathematical Sciences >> - University of Delaware >> - schwenk _at_ math _dot_ udel _dot_ edu >> - http://www.math.udel.edu/~schwenk > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From debian at herakles.homelinux.org Thu Feb 14 11:59:08 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Thu, 14 Feb 2008 20:59:08 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B42C8D.4050500@herakles.homelinux.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42C8D.4050500@herakles.homelinux.org> Message-ID: <47B42D0C.3080003@herakles.homelinux.org> John Summerfield wrote: > dwight at supercomputer.org wrote: > > I'm forking a thread on ks, because it illustrates something that's been > on my mind for a while. So much for that plan:=\ -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From schwenk at math.udel.edu Thu Feb 14 14:25:06 2008 From: schwenk at math.udel.edu (Peter Schwenk) Date: Thu, 14 Feb 2008 09:25:06 -0500 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B42AFE.3070504@herakles.homelinux.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> Message-ID: <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> I, the OP, am in fact doing a network install. I get the ball rolling with the CD. I never tried Fedora 7, but Fedoras 6 and prior didn't lock the CD in the drive during the network install. I am reluctant to start killing processes during the install just so I can remove my CD as dwight at supercomputer.org suggested. I'm booting from the Rescue CD to do my installs, and one thing I noticed is that a message box pops up indicating that the install found local install media. This message hadn't appeared with the earlier Fedoras. I'll try to make a bare-bones boot CD somehow with no install files to see if that helps keep the CD from being locked in. On Feb 14, 2008, at 6:50 AM, John Summerfield wrote: > dwight at supercomputer.org wrote: >> It sounds like there's some process left around which is keeping >> the device open. That would be the first thing I'd check. If >> nothing is obvious from a ps(1), then you'll have to dig deeper. > > I'm going to take this to the Anaconda list, because it illustrates > something that's been on my mind for a while. > > When I boot CD to net install (which I suppose the OP is doing coz > otherwise it doesn;t make sense to head for other parts, media in > hand), I'd like it to spit the CD out as soon as it's not required. > > -- > > Cheers > John > > -- spambait > 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu > -- Advice > http://webfoot.com/advice/email.top.php > http://www.catb.org/~esr/faqs/smart-questions.html > http://support.microsoft.com/kb/555375 > > You cannot reply off-list:-) > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > ----- - Peter Schwenk - CITA-3, Systems Administrator - Mathematical Sciences - University of Delaware - schwenk _at_ math _dot_ udel _dot_ edu - http://www.math.udel.edu/~schwenk -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2425 bytes Desc: not available URL: From gary at mlbassoc.com Thu Feb 14 14:34:17 2008 From: gary at mlbassoc.com (Gary Thomas) Date: Thu, 14 Feb 2008 07:34:17 -0700 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> Message-ID: <47B45169.2040801@mlbassoc.com> Peter Schwenk wrote: > I, the OP, am in fact doing a network install. I get the ball rolling > with the CD. I never tried Fedora 7, but Fedoras 6 and prior didn't > lock the CD in the drive during the network install. I am reluctant to > start killing processes during the install just so I can remove my CD as > dwight at supercomputer.org suggested. I'm booting from the Rescue CD > to do my installs, and one thing I noticed is that a message box pops up > indicating that the install found local install media. This message > hadn't appeared with the earlier Fedoras. I'll try to make a bare-bones > boot CD somehow with no install files to see if that helps keep the CD > from being locked in. If you're doing a pure network install, then don't use the rescue disk :-) Boot from 'boot.iso' (found in the main distribution). Then you can select a network install (using kickstart or manually) and the CD can be ejected once the system starts (it only contains the kernel and initial ram disk) > On Feb 14, 2008, at 6:50 AM, John Summerfield wrote: > >> dwight at supercomputer.org wrote: >>> It sounds like there's some process left around which is keeping the >>> device open. That would be the first thing I'd check. If nothing is >>> obvious from a ps(1), then you'll have to dig deeper. >> >> I'm going to take this to the Anaconda list, because it illustrates >> something that's been on my mind for a while. >> >> When I boot CD to net install (which I suppose the OP is doing coz >> otherwise it doesn;t make sense to head for other parts, media in >> hand), I'd like it to spit the CD out as soon as it's not required. -- ------------------------------------------------------------ Gary Thomas | Consulting for the MLB Associates | Embedded world ------------------------------------------------------------ From rainer at ultra-secure.de Thu Feb 14 14:41:35 2008 From: rainer at ultra-secure.de (Rainer Duffner) Date: Thu, 14 Feb 2008 15:41:35 +0100 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> Message-ID: <47B4531F.8060203@ultra-secure.de> Peter Schwenk schrieb: > I, the OP, am in fact doing a network install. Would it be possible for you to setup cobbler? (DHCP, TFTP, etc.) It requires some time to get going, but after that, you never want to do it any other way again. cheers, Rainer From schwenk at math.udel.edu Thu Feb 14 14:40:38 2008 From: schwenk at math.udel.edu (Peter Schwenk) Date: Thu, 14 Feb 2008 09:40:38 -0500 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B45169.2040801@mlbassoc.com> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> <47B45169.2040801@mlbassoc.com> Message-ID: <7F6BD104-BA63-4DF2-97E4-46E0729BC68B@math.udel.edu> Thanks. Right after I sent that email, I poked around on my Fedora 8 DVD and found the very same boot.iso file. I made a CD and tried it, and it works as advertised. I used to use the first CD of the Fedora set for my boot disk, and it never was locked in there before. The behavior of anaconda has changed, but I know now that I should have been using the boot.iso. On Feb 14, 2008, at 9:34 AM, Gary Thomas wrote: > Peter Schwenk wrote: >> I, the OP, am in fact doing a network install. I get the ball >> rolling with the CD. I never tried Fedora 7, but Fedoras 6 and >> prior didn't lock the CD in the drive during the network install. >> I am reluctant to start killing processes during the install just >> so I can remove my CD as dwight at supercomputer.org suggested. >> I'm booting from the Rescue CD to do my installs, and one thing I >> noticed is that a message box pops up indicating that the install >> found local install media. This message hadn't appeared with the >> earlier Fedoras. I'll try to make a bare-bones boot CD somehow >> with no install files to see if that helps keep the CD from being >> locked in. > > If you're doing a pure network install, then don't use the rescue > disk :-) > Boot from 'boot.iso' (found in the main distribution). Then you can > select a network install (using kickstart or manually) and the CD can > be ejected once the system starts (it only contains the kernel and > initial > ram disk) > >> On Feb 14, 2008, at 6:50 AM, John Summerfield wrote: >>> dwight at supercomputer.org wrote: >>>> It sounds like there's some process left around which is keeping >>>> the device open. That would be the first thing I'd check. If >>>> nothing is obvious from a ps(1), then you'll have to dig deeper. >>> >>> I'm going to take this to the Anaconda list, because it >>> illustrates something that's been on my mind for a while. >>> >>> When I boot CD to net install (which I suppose the OP is doing coz >>> otherwise it doesn;t make sense to head for other parts, media in >>> hand), I'd like it to spit the CD out as soon as it's not required. > > -- > ------------------------------------------------------------ > Gary Thomas | Consulting for the > MLB Associates | Embedded world > ------------------------------------------------------------ > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > ----- - Peter Schwenk - CITA-3, Systems Administrator - Mathematical Sciences - University of Delaware - schwenk _at_ math _dot_ udel _dot_ edu - http://www.math.udel.edu/~schwenk -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2425 bytes Desc: not available URL: From andreas.fuchs at rtc.ch Thu Feb 14 15:19:04 2008 From: andreas.fuchs at rtc.ch (Fuchs Andreas) Date: Thu, 14 Feb 2008 16:19:04 +0100 Subject: New to RH Satellite server and Kickstart In-Reply-To: <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu><200802130847.45928.dwight@supercomputer.org><47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> Message-ID: Hi I'm pretty new to RHN Satellite server and Kickstart and have a couple of questions. We have the Satellite server setup with the needed channels and a kickstart profile ready for those channels. Bare Metall installation is running ok ending with a installed system with no hostname (registered as "unknown" in the satellite server) and ip from DHCP. We wonder now: - How does the server get his Hostname? - How do we configure the Networks (there are severals including some bonding)? What's best practice? - cloning a template kickstart profile for each server and adding the values? - create pre/post scripts who ask for the information during install? as we like this idea - is there some documentation/examples/helper scripts to get a this nicely done? - using config channels and also some kind of scripts to get hostname and ip? Also: - is cobbler supporting/integrating with the satellite server - or is there another way of doing this nicely Remarks: We need to install and configure rhel4 and 5 systems. There is no cmdb available who has the config stored. cheers Andreas From lists at zadevalov.com Thu Feb 14 06:56:26 2008 From: lists at zadevalov.com (Eugeny Zadevalov) Date: Thu, 14 Feb 2008 08:56:26 +0200 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: References: Message-ID: <47B3E61A.1070706@zadevalov.com> Hello Shabazian, Shabazian, Chip wrote: > Not sure, but I do --all before --initlabel. Shouldn't make a > difference, but I've seen stranger things. :-) Actually I do --all before --initlabel too, the example I've posted in my original e-mail was already from modified ks.cfg where I've tried to swap options. It didn't help through. :-( > > The other thing you could probably do is wipe the the partition table in > %pre: > > set $(list-harddrives) > cat << EOF >> /tmp/fdisk.input > w > > EOF > > for i in $d1,$d2,$d3,...(better ways to do this, but this is worth > starting with); do fdisk $i < /tmp/fdisk.input; done Well, I know about %pre but I thought here at kickstart maillist there are people who knows anaconda source and can maybe look into the code and try to fix this issue. From Pablo.Iranzo at redhat.com Thu Feb 14 15:26:06 2008 From: Pablo.Iranzo at redhat.com (Pablo Iranzo =?ISO-8859-1?Q?G=F3mez?=) Date: Thu, 14 Feb 2008 15:26:06 +0000 Subject: New to RH Satellite server and Kickstart In-Reply-To: References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> Message-ID: <1203002766.20816.116.camel@iranzo.users.redhat.com> Hi El jue, 14-02-2008 a las 16:19 +0100, Fuchs Andreas escribi?: > Hi > > I'm pretty new to RHN Satellite server and Kickstart and have a couple > of questions. > > We have the Satellite server setup with the needed channels and a > kickstart profile ready for those channels. > Bare Metall installation is running ok ending with a installed system > with no hostname (registered as "unknown" in the satellite server) and > ip from DHCP. > > We wonder now: > > - How does the server get his Hostname? From reverse IP resolution for your IP, the trick is to create an entry with your desired hostname and IP address in /etc/hosts or using API to edit shown name > - How do we configure the Networks (there are severals including some > bonding)? Create your custom /etc/sysconfig/network-scripts/ to play around with your desires, or do it the dirty way, putting them at /etc/rc.local > > What's best practice? > > - cloning a template kickstart profile for each server and adding the > values? > - create pre/post scripts who ask for the information during install? > as we like this idea This would be better > - is there some documentation/examples/helper scripts to get a this > nicely done? %post #!/bin/bash chvt 3 echo "ENTER CONFIG #" read valK > - using config channels and also some kind of scripts to get hostname > and ip? Check for config file's macros > > Also: > > - is cobbler supporting/integrating with the satellite server They're separate projects, probably could try at #cobbler on freenode -- Pablo Iranzo G?mez (Pablo.Iranzo at redhat.com) RHCE/Global Profesional Services Consultant Spain Phone: +34 645 01 01 49 (CET/CEST) GnuPG KeyID: 0xFAD3CF0D --- Direcci?n Comercial: C/Jose Bardasano Baos, 9, Edif. Gorbea 3, planta 3?D, 28016 Madrid, Spain Direcci?n Registrada: Red Hat S.L., C/ Velazquez 63, Madrid 28001, Spain Inscrita en el Reg. Mercantil de Madrid ? C.I.F. B82657941 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Esta parte del mensaje est? firmada digitalmente URL: From mdehaan at redhat.com Thu Feb 14 16:04:58 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 14 Feb 2008 11:04:58 -0500 Subject: New to RH Satellite server and Kickstart In-Reply-To: References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu><200802130847.45928.dwight@supercomputer.org><47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> Message-ID: <47B466AA.5030509@redhat.com> Fuchs Andreas wrote: > Hi > > I'm pretty new to RHN Satellite server and Kickstart and have a couple > of questions. > > We have the Satellite server setup with the needed channels and a > kickstart profile ready for those channels. > Bare Metall installation is running ok ending with a installed system > with no hostname (registered as "unknown" in the satellite server) and > ip from DHCP. > > We wonder now: > > - How does the server get his Hostname? > - How do we configure the Networks (there are severals including some > bonding)? > > What's best practice? > > - cloning a template kickstart profile for each server and adding the > values? > - create pre/post scripts who ask for the information during install? > as we like this idea > IMHO, you should be aiming for fully-automated installations whenever possible. Using a templating system, like Cobbler's, you can share one kickstart template among various systems, add the system network information into Cobbler, and cobbler will build out the per-system kickstarts for you automatically. In cobbler, this is as simple as: cobbler system add --name=foo --mac=AA:BB:CC:DD:EE:FF --ip=192.168.5.5 --subnet=.... and in the template you can reference things like $ip_address and so on automatically there is also a web interface available for setting this up if you don't like the command line. It can also help you manage your DHCP and PXE configurations if you want to do that. > - is there some documentation/examples/helper scripts to get a this > nicely done? > It's kind of spread out, but start here: http://cobbler.et.redhat.com https://fedorahosted.org/cobbler/ The templating system is described here: https://fedorahosted.org/cobbler/wiki/KickstartTemplating (warning, this reads a bit scarier than it actually is) > > Also: > > - is cobbler supporting/integrating with the satellite server > Cobbler plays nicely with Satellite. You can use Cobbler with Satellite/RHN very easily if you want to. All you have to do is call rhnreg_ks (or is it rhn_regks, I forget) in %post so that the systems you install get registered to RHN. From mdehaan at redhat.com Thu Feb 14 16:12:02 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 14 Feb 2008 11:12:02 -0500 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B42C8D.4050500@herakles.homelinux.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42C8D.4050500@herakles.homelinux.org> Message-ID: <47B46852.7040504@redhat.com> John Summerfield wrote: > Taking the CD ensures there won't be an install loop. Waiting for it's > tedious. > Sorry to take over the list (seriously), but cobbler has anti install loop prevention features that solve that problem. Just turn on pxe_just_once in settings and when a system finishes installing, it will toggle the netboot flag for that system. Should you want it to netboot again, you can do: cobbler system edit --name=foo --netboot-enabled=1 The alternative is to set the BIOS behavior to only PXE if there is no MBR, leaving PXE out of the normal boot order. --Michael From tom at ng23.net Thu Feb 14 17:21:57 2008 From: tom at ng23.net (Tom Brown) Date: Thu, 14 Feb 2008 17:21:57 +0000 Subject: New to RH Satellite server and Kickstart In-Reply-To: <47B466AA.5030509@redhat.com> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu><200802130847.45928.dwight@supercomputer.org><47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> <47B466AA.5030509@redhat.com> Message-ID: <47B478B5.10208@ng23.net> > > IMHO, you should be aiming for fully-automated installations whenever > possible. > > Using a templating system, like Cobbler's, you can share one kickstart > template among various systems, add the system > network information into Cobbler, and cobbler will build out the > per-system kickstarts for you automatically. > > In cobbler, this is as simple as: > cobbler system add --name=foo --mac=AA:BB:CC:DD:EE:FF --ip=192.168.5.5 > --subnet=.... > and in the template you can reference things like $ip_address and so > on automatically > there is also a web interface available for setting this up if you > don't like the command line. > > It can also help you manage your DHCP and PXE configurations if you > want to do that. > >> - is there some documentation/examples/helper scripts to get a this >> nicely done? >> > It's kind of spread out, but start here: > > http://cobbler.et.redhat.com > https://fedorahosted.org/cobbler/ > > The templating system is described here: > > https://fedorahosted.org/cobbler/wiki/KickstartTemplating (warning, > this reads a bit scarier than it actually is) > >> >> Also: >> >> - is cobbler supporting/integrating with the satellite server >> > > Cobbler plays nicely with Satellite. You can use Cobbler with > Satellite/RHN very easily if you want to. > > All you have to do is call rhnreg_ks (or is it rhn_regks, I forget) in > %post so that the systems you install get > registered to RHN. satellite is clunky, horrible to learn and just plain expensive - cobbler on the other hand is brilliant and once you get used to it adding new distros or profiles and syncing boxes to those profiles is a breeze. Users can also rebuild their own boxes using koan if so desired. From dwight at supercomputer.org Thu Feb 14 17:19:15 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Thu, 14 Feb 2008 09:19:15 -0800 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B42C8D.4050500@herakles.homelinux.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42C8D.4050500@herakles.homelinux.org> Message-ID: <200802140919.16201.dwight@supercomputer.org> Lest there be any confusion, I didn't write the stuff attributed to me (the first indented ">"); that belongs to John. No doubt due to a simple typing error. -dwight- On Thursday 14 February 2008 03:57:01 am John Summerfield wrote: > dwight at supercomputer.org wrote: > > I'm forking a thread on ks, because it illustrates something > that's been on my mind for a while. > > When I boot CD to net install (which I suppose the OP is doing coz > otherwise it doesn;t make sense to head for other parts, media in > hand), I'd like it to spit the CD out as soon as it's not > required. > > Normally, I use PXE but just sometimes, I might boot from CD and > then install off a good ADSL connexion. Taking the CD ensures > there won't be an install loop. Waiting for it's tedious. > > > bts I've just been testing PX+KS (CentOS5.1) on a 3yo or so > Pentium IV. Takes a little less than 15 minutes, about what RHHL > 7.2-3 took on Pentium IIs fdr a similar install. > > In contrast, XP's equivalent takes around an hour and despite the > .SIF file (equivalent to the ks) it insists asking more question. > And then it's ready to install the applications such as office > software. > > Good Work JK (who I think's been on it all that time) and crew. > > > It sounds like there's some process left around which is keeping > > the device open. That would be the first thing I'd check. If > > nothing is obvious from a ps(1), then you'll have to dig deeper. > > > > lsof(1) and/or an `fuser -v` are typically used to track this > > down. The trouble is that these aren't available from the F2 > > virtual terminal. I'm going from memory at the moment, but I'm > > pretty certain this is the case. > > > > To use these utilities, you'll have to do something like the > > following: > > 1. chroot /mnt/system > > 2. mount up /sys (Check fstab, it would be something like > > mount -t sysfs /sys > > 3. mount up /proc > > 4. /sbin/start_udev > > > > You should then be able to use lsof and/or fuser. > > > > You can also now do all sorts of other nifty things. > > > > Note that to exit this environment cleanly, you have to clean up > > in the reverse order. Starting will killing the udevd daemon. > > > > Hope that helps. > > > > -dwight- > > > > On Wednesday 13 February 2008 06:55:13 am Peter Schwenk wrote: > >> I've been using kickstart installs with Fedora and CentOS for > >> quite a while now. I boot the system with a cd and enter the > >> appropriate command at the BOOT: prompt and away it goes. I've > >> gotten used to being able to eject the cd from the system > >> (using the button on the drive) at the point where it is > >> installing packages. For some reason, with Fedora 8 kickstart > >> installs the cd is locked in the drive and I cannot eject it. > >> I've tried going to a virtual terminal and running 'eject > >> /tmp/cdrom', but it returns "Input/output error". > >> > >> I'd like to be able to take the cd with me so that I don't have > >> to go and retrieve it later. Does anyone know how this can be > >> done? Thanks in advance for your help. > >> > >> > >> ----- > >> - Peter Schwenk > >> - CITA-3, Systems Administrator > >> - Mathematical Sciences > >> - University of Delaware > >> - schwenk _at_ math _dot_ udel _dot_ edu > >> - http://www.math.udel.edu/~schwenk > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list From Greg.Caetano at hp.com Thu Feb 14 17:41:29 2008 From: Greg.Caetano at hp.com (Caetano, Greg) Date: Thu, 14 Feb 2008 17:41:29 +0000 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: <47B3E61A.1070706@zadevalov.com> References: <47B3E61A.1070706@zadevalov.com> Message-ID: Don't forget the "zerombr" kickstart directive is also available. zerombr (optional) If zerombr is specified, and yes is its sole argument, any invalid partition tables found on disks are initialized. This destroys all of the contents of disks with invalid partition tables. This command should be in the following format: zerombr yes No other format is effective. Greg Caetano HP TSG Linux Solutions Alliances Engineering greg.caetano at hp.com Red Hat Certified Engineer RHCE#805007310328754 -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Eugeny Zadevalov Sent: Thursday, February 14, 2008 12:56 AM To: Shabazian, Chip Subject: Re: problems with unpartitioned harddrives and automatic partitioning Hello Shabazian, Shabazian, Chip wrote: > Not sure, but I do --all before --initlabel. Shouldn't make a > difference, but I've seen stranger things. :-) Actually I do --all before --initlabel too, the example I've posted in my original e-mail was already from modified ks.cfg where I've tried to swap options. It didn't help through. :-( > > The other thing you could probably do is wipe the the partition table in > %pre: > > set $(list-harddrives) > cat << EOF >> /tmp/fdisk.input > w > > EOF > > for i in $d1,$d2,$d3,...(better ways to do this, but this is worth > starting with); do fdisk $i < /tmp/fdisk.input; done Well, I know about %pre but I thought here at kickstart maillist there are people who knows anaconda source and can maybe look into the code and try to fix this issue. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From debian at herakles.homelinux.org Thu Feb 14 23:31:59 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Fri, 15 Feb 2008 08:31:59 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> Message-ID: <47B4CF6F.1070900@herakles.homelinux.org> Peter Schwenk wrote: > I, the OP, am in fact doing a network install. I get the ball rolling > with the CD. I never tried Fedora 7, but Fedoras 6 and prior didn't > lock the CD in the drive during the network install. I am reluctant to > start killing processes during the install just so I can remove my CD as > dwight at supercomputer.org suggested. I'm booting from the Rescue CD > to do my installs, and one thing I noticed is that a message box pops up > indicating that the install found local install media. This message > hadn't appeared with the earlier Fedoras. I'll try to make a bare-bones > boot CD somehow with no install files to see if that helps keep the CD > from being locked in. > There's the boot.iso already available, that's what I use. With minor surgery it's possible to get a ks file into it and have it used. As you note, getting it out's the trick. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From debian at herakles.homelinux.org Thu Feb 14 23:37:55 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Fri, 15 Feb 2008 08:37:55 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B4531F.8060203@ultra-secure.de> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42AFE.3070504@herakles.homelinux.org> <21403246-39DD-4435-BBA1-4C16117C6782@math.udel.edu> <47B4531F.8060203@ultra-secure.de> Message-ID: <47B4D0D3.3070008@herakles.homelinux.org> Rainer Duffner wrote: > Peter Schwenk schrieb: >> I, the OP, am in fact doing a network install. > > > Would it be possible for you to setup cobbler? (DHCP, TFTP, etc.) > It requires some time to get going, but after that, you never want to do > it any other way again. _I_ already have the install infrastructure set up, it's evolved since RHL 7.2, and caters for Debian, Ubuntu and RHL-family. However, sometimes it's convenient not to use it. I'm contemplating extending it to do Windows too; I have one ISC DHCPD set up that does Windows, I'm contemplating choosing between Windows and The Others at PXE-boot time. There has also been discussion (maybe other fora) in the distant past about preventing install loops. Getting the CD out early would help. Also a note when it's safe to remove USB media. for those with that need. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From debian at herakles.homelinux.org Thu Feb 14 23:45:37 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Fri, 15 Feb 2008 08:45:37 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B46852.7040504@redhat.com> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42C8D.4050500@herakles.homelinux.org> <47B46852.7040504@redhat.com> Message-ID: <47B4D2A1.1030305@herakles.homelinux.org> Michael DeHaan wrote: > John Summerfield wrote: >> Taking the CD ensures there won't be an install loop. Waiting for it's >> tedious. >> > > Sorry to take over the list (seriously), but cobbler has anti install > loop prevention features that solve that problem. > > Just turn on pxe_just_once in settings and when a system finishes > installing, it will toggle the netboot flag for that system. Doesn't help, when one boots boot.iso. > > Should you want it to netboot again, you can do: > cobbler system edit --name=foo --netboot-enabled=1 I'm not sure I want to run round the offices or phone people to get this done. I often do network installs in another building from where the servers are, on a LAN that can't reach (by ssh and such) the servers. > > The alternative is to set the BIOS behavior to only PXE if there is no > MBR, leaving PXE out of the normal boot order. More recent systems I see have the ability to do a one-time boot from the network (and some from CD), but this doesn't help on systems that lack this facility. The question I put is not how to fix a problem for an individual, but about how to provide a solution for many users. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From debian at herakles.homelinux.org Thu Feb 14 23:51:47 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Fri, 15 Feb 2008 08:51:47 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <200802140919.16201.dwight@supercomputer.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org> <47B42C8D.4050500@herakles.homelinux.org> <200802140919.16201.dwight@supercomputer.org> Message-ID: <47B4D413.6050200@herakles.homelinux.org> dwight at supercomputer.org wrote: > Lest there be any confusion, I didn't write the stuff attributed to > me (the first indented ">"); that belongs to John. No doubt due to a > simple typing error. > > -dwight- Now I am confused. Please write directly below text you attribute to me and about which you see some confusion. I don't see anything I wrote that you'd be keen to disown, and I don't see anything I wrote that I'm especially keen to either claim or disclaim. Aside from spelling errors. I see "bts" where I intended "btw." > > On Thursday 14 February 2008 03:57:01 am John Summerfield wrote: >> dwight at supercomputer.org wrote: >> >> I'm forking a thread on ks, because it illustrates something >> that's been on my mind for a while. >> >> When I boot CD to net install (which I suppose the OP is doing coz >> otherwise it doesn;t make sense to head for other parts, media in >> hand), I'd like it to spit the CD out as soon as it's not >> required. >> >> Normally, I use PXE but just sometimes, I might boot from CD and >> then install off a good ADSL connexion. Taking the CD ensures >> there won't be an install loop. Waiting for it's tedious. >> >> >> bts I've just been testing PX+KS (CentOS5.1) on a 3yo or so >> Pentium IV. Takes a little less than 15 minutes, about what RHHL >> 7.2-3 took on Pentium IIs fdr a similar install. >> >> In contrast, XP's equivalent takes around an hour and despite the >> .SIF file (equivalent to the ks) it insists asking more question. >> And then it's ready to install the applications such as office >> software. >> >> Good Work JK (who I think's been on it all that time) and crew. >> >>> It sounds like there's some process left around which is keeping >>> the device open. That would be the first thing I'd check. If >>> nothing is obvious from a ps(1), then you'll have to dig deeper. >>> >>> lsof(1) and/or an `fuser -v` are typically used to track this >>> down. The trouble is that these aren't available from the F2 >>> virtual terminal. I'm going from memory at the moment, but I'm >>> pretty certain this is the case. >>> >>> To use these utilities, you'll have to do something like the >>> following: >>> 1. chroot /mnt/system >>> 2. mount up /sys (Check fstab, it would be something like >>> mount -t sysfs /sys >>> 3. mount up /proc >>> 4. /sbin/start_udev >>> >>> You should then be able to use lsof and/or fuser. >>> >>> You can also now do all sorts of other nifty things. >>> >>> Note that to exit this environment cleanly, you have to clean up >>> in the reverse order. Starting will killing the udevd daemon. >>> >>> Hope that helps. >>> >>> -dwight- >>> >>> On Wednesday 13 February 2008 06:55:13 am Peter Schwenk wrote: >>>> I've been using kickstart installs with Fedora and CentOS for >>>> quite a while now. I boot the system with a cd and enter the >>>> appropriate command at the BOOT: prompt and away it goes. I've >>>> gotten used to being able to eject the cd from the system >>>> (using the button on the drive) at the point where it is >>>> installing packages. For some reason, with Fedora 8 kickstart >>>> installs the cd is locked in the drive and I cannot eject it. >>>> I've tried going to a virtual terminal and running 'eject >>>> /tmp/cdrom', but it returns "Input/output error". >>>> >>>> I'd like to be able to take the cd with me so that I don't have >>>> to go and retrieve it later. Does anyone know how this can be >>>> done? Thanks in advance for your help. >>>> >>>> >>>> ----- >>>> - Peter Schwenk >>>> - CITA-3, Systems Administrator >>>> - Mathematical Sciences >>>> - University of Delaware >>>> - schwenk _at_ math _dot_ udel _dot_ edu >>>> - http://www.math.udel.edu/~schwenk >>> _______________________________________________ >>> 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 > -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From dwight at supercomputer.org Fri Feb 15 04:19:28 2008 From: dwight at supercomputer.org (dwight at supercomputer.org) Date: Thu, 14 Feb 2008 20:19:28 -0800 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B4D413.6050200@herakles.homelinux.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802140919.16201.dwight@supercomputer.org> <47B4D413.6050200@herakles.homelinux.org> Message-ID: <200802142019.28411.dwight@supercomputer.org> It showed up here as: dwight wrote: [Your words] > [My original words] Which is confusing. It would have been more clear to see this as: [Your words] dwight wrote: > [My original words] For more clarification, perhaps referring to the web view would help, as this is how it appeared to me: https://www.redhat.com/archives/kickstart-list/2008-February/msg00059.html I have no objection to your statements. It's just that I didn't make them, and wanted to make that clear. That's all I'll comment further on the list about this. I doubt anyone else is really interested in anything further. -dwight- On Thursday 14 February 2008 03:51:47 pm John Summerfield wrote: > dwight at supercomputer.org wrote: > > Lest there be any confusion, I didn't write the stuff attributed > > to me (the first indented ">"); that belongs to John. No doubt > > due to a simple typing error. > > > > -dwight- > > Now I am confused. Please write directly below text you attribute > to me and about which you see some confusion. I don't see anything > I wrote that you'd be keen to disown, and I don't see anything I > wrote that I'm especially keen to either claim or disclaim. Aside > from spelling errors. I see "bts" where I intended "btw." > > > On Thursday 14 February 2008 03:57:01 am John Summerfield wrote: > >> dwight at supercomputer.org wrote: > >> > >> I'm forking a thread on ks, because it illustrates something > >> that's been on my mind for a while. > >> > >> When I boot CD to net install (which I suppose the OP is doing > >> coz otherwise it doesn;t make sense to head for other parts, > >> media in hand), I'd like it to spit the CD out as soon as it's > >> not required. > >> > >> Normally, I use PXE but just sometimes, I might boot from CD > >> and then install off a good ADSL connexion. Taking the CD > >> ensures there won't be an install loop. Waiting for it's > >> tedious. > >> > >> > >> bts I've just been testing PX+KS (CentOS5.1) on a 3yo or so > >> Pentium IV. Takes a little less than 15 minutes, about what > >> RHHL 7.2-3 took on Pentium IIs fdr a similar install. > >> > >> In contrast, XP's equivalent takes around an hour and despite > >> the .SIF file (equivalent to the ks) it insists asking more > >> question. And then it's ready to install the applications such > >> as office software. > >> > >> Good Work JK (who I think's been on it all that time) and crew. > >> > >>> It sounds like there's some process left around which is > >>> keeping the device open. That would be the first thing I'd > >>> check. If nothing is obvious from a ps(1), then you'll have to > >>> dig deeper. > >>> > >>> lsof(1) and/or an `fuser -v` are typically used to track this > >>> down. The trouble is that these aren't available from the F2 > >>> virtual terminal. I'm going from memory at the moment, but I'm > >>> pretty certain this is the case. > >>> > >>> To use these utilities, you'll have to do something like the > >>> following: > >>> 1. chroot /mnt/system > >>> 2. mount up /sys (Check fstab, it would be something like > >>> mount -t sysfs /sys > >>> 3. mount up /proc > >>> 4. /sbin/start_udev > >>> > >>> You should then be able to use lsof and/or fuser. > >>> > >>> You can also now do all sorts of other nifty things. > >>> > >>> Note that to exit this environment cleanly, you have to clean > >>> up in the reverse order. Starting will killing the udevd > >>> daemon. > >>> > >>> Hope that helps. > >>> > >>> -dwight- > >>> > >>> On Wednesday 13 February 2008 06:55:13 am Peter Schwenk wrote: > >>>> I've been using kickstart installs with Fedora and CentOS for > >>>> quite a while now. I boot the system with a cd and enter the > >>>> appropriate command at the BOOT: prompt and away it goes. > >>>> I've gotten used to being able to eject the cd from the > >>>> system (using the button on the drive) at the point where it > >>>> is installing packages. For some reason, with Fedora 8 > >>>> kickstart installs the cd is locked in the drive and I cannot > >>>> eject it. I've tried going to a virtual terminal and running > >>>> 'eject /tmp/cdrom', but it returns "Input/output error". > >>>> > >>>> I'd like to be able to take the cd with me so that I don't > >>>> have to go and retrieve it later. Does anyone know how this > >>>> can be done? Thanks in advance for your help. > >>>> > >>>> > >>>> ----- > >>>> - Peter Schwenk > >>>> - CITA-3, Systems Administrator > >>>> - Mathematical Sciences > >>>> - University of Delaware > >>>> - schwenk _at_ math _dot_ udel _dot_ edu > >>>> - http://www.math.udel.edu/~schwenk > >>> > >>> _______________________________________________ > >>> 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 lists at zadevalov.com Fri Feb 15 07:50:48 2008 From: lists at zadevalov.com (Eugeny Zadevalov) Date: Fri, 15 Feb 2008 09:50:48 +0200 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: References: <47B3E61A.1070706@zadevalov.com> Message-ID: <47B54458.5030507@zadevalov.com> Hello Greg, Caetano, Greg wrote: > Don't forget the "zerombr" kickstart directive is also available. Thank you very much, below is cut&paste from my first e-mail to the list: === zerombr clearpart --initlabel --all part raid.01 --size=2048 --onbiosdisk=80 --asprimary part raid.02 --size=8192 --onbiosdisk=80 --asprimary --grow part raid.03 --size=2048 --onbiosdisk=81 --asprimary part raid.04 --size=8192 --onbiosdisk=81 --asprimary --grow raid / --level=RAID1 --device=md0 --fstype=ext3 raid.02 raid.04 raid swap --level=RAID1 --device=md1 --fstype=swap raid.01 raid.03 === From lists at zadevalov.com Fri Feb 15 08:10:47 2008 From: lists at zadevalov.com (Eugeny Zadevalov) Date: Fri, 15 Feb 2008 10:10:47 +0200 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: <47B36AAA.7010501@zadevalov.com> References: <47B36AAA.7010501@zadevalov.com> Message-ID: <47B54907.6070202@zadevalov.com> Hello, Here is what it's on the console when drives are unpartitioned: Traceback (most recent call last): File "/usr/bin/anaconda", line 938, in ? instClass.setInstallData(anaconda) File "/usr/lib/anaconda/kickstart.py", line 831, in setInstallData self.ksparser.readKickstart(self.file) File "/usr/lib/python2.4/site-packages/pykickstart/parser.py", line 1151, in readKickstart self.handleCommand(lineno, args) File "/usr/lib/anaconda/kickstart.py", line 719, in handleCommand self.handler.handlers[cmd](cmdArgs) File "/usr/lib/anaconda/kickstart.py", line 440, in doPartition if not hds.has_key(pd.disk) and hds.has_key('mapper/'+pd.disk): TypeError: cannot concatenate 'str' and 'NoneType' objects install exited abnormally [1/1] sending termination signals...done sending kill signals...done ... ... Does anyone from the anaconda development team actually read this list? Eugeny Zadevalov wrote: > Hello All, > > I've been trying to setup automatic partitioning with Kickstart. > Everything is okay until I try to run my Kickstart setup on machine with > unpartitioned drives (like if partition table is broken or filled with > zeros). > > I'm experiencing same issue with both Centos4.6 and Centos5.1. > > My goal is to get raid1 setup automatically through kickstart. Below is > cut from ks.cfg: > === > zerombr > clearpart --initlabel --all > part raid.01 --size=2048 --onbiosdisk=80 --asprimary > part raid.02 --size=8192 --onbiosdisk=80 --asprimary --grow > part raid.03 --size=2048 --onbiosdisk=81 --asprimary > part raid.04 --size=8192 --onbiosdisk=81 --asprimary --grow > raid / --level=RAID1 --device=md0 --fstype=ext3 raid.02 raid.04 > raid swap --level=RAID1 --device=md1 --fstype=swap raid.01 raid.03 > === > > If I remove this part and run kickstarting machine over Kickstart with > manual partitioning I get warning about drives needs to be initialized. > That's normal I guess. But with my automatic partitioning I've > "clearpart --initlabel --all" and that supposed to be initializing > any/all unpartitioned drives but that's not happening. Kickstart just > hangs giving bunch of errors. > > If I just go ahead and do something like: > fdisk /dev/hda > w > ENTER > fdisk /dev/hdc > w > ENTER > And then try again that automatic partitioning scheme it works like > expected. > > I've searched over the mailling list archive but wasn't able to find > anything. > > For me it looks like a bug with "--initlabel" option. > > Any suggestions? > > Thanks! > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From lists at zadevalov.com Sun Feb 17 11:47:18 2008 From: lists at zadevalov.com (Eugeny Zadevalov) Date: Sun, 17 Feb 2008 13:47:18 +0200 Subject: problems with unpartitioned harddrives and automatic partitioning In-Reply-To: References: <47B3E61A.1070706@zadevalov.com> <47B54458.5030507@zadevalov.com> Message-ID: <47B81EC6.6010309@zadevalov.com> Hello! I've modified "zerombr" to be "zerombr yes" and tested installation. This doesn't helps. Caetano, Greg wrote: > Eugeny: > > The "zerombr" directive needs a parameter "yes" to confirm you want to initialize the tables > > regards > Greg > > -----Original Message----- > From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Eugeny Zadevalov > Sent: Friday, February 15, 2008 1:51 AM > To: kickstart-list at redhat.com > Subject: Re: problems with unpartitioned harddrives and automatic partitioning > > Hello Greg, > > Caetano, Greg wrote: >> Don't forget the "zerombr" kickstart directive is also available. > > Thank you very much, below is cut&paste from my first e-mail to the list: > > === > zerombr > clearpart --initlabel --all > part raid.01 --size=2048 --onbiosdisk=80 --asprimary > part raid.02 --size=8192 --onbiosdisk=80 --asprimary --grow > part raid.03 --size=2048 --onbiosdisk=81 --asprimary > part raid.04 --size=8192 --onbiosdisk=81 --asprimary --grow > raid / --level=RAID1 --device=md0 --fstype=ext3 raid.02 raid.04 > raid swap --level=RAID1 --device=md1 --fstype=swap raid.01 raid.03 > === > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From jman at ablesky.com Mon Feb 18 08:02:18 2008 From: jman at ablesky.com (John Morris) Date: Mon, 18 Feb 2008 16:02:18 +0800 Subject: Installing onto existing LVM partition causes traceback Message-ID: <47B93B8A.7050104@ablesky.com> Hi list, CentOS 5.1, x86_64, while trying to reinstall a machine, trying to keep existing LVM vgs. The below disk section causes anaconda to stop with a traceback. The machine's 4 disks are divided thusly: sd[ab]1: RAID1, 100MB, for /boot; erase during install sd[ab]2: RAID1, 15GB, for 'dom0' volume group; don't erase during install dom0 volume group; don't erase during install root logvol; erase and reformat existing lv during install sd[ab]3: swap partitions sd[ab]4: RAID1, 735GB; sd[cd]1: RAID1, 750GB; don't erase during install vg1 volume group; don't erase during install many existing logical volumes that shouldn't be erased The traceback occurs both when creating both the dom0 and vg1 volume groups. If the 'volgroup' line is commented out or the '--noformat' option is removed, the kickstart will continue. I left the original, working lines in, commented out, for information. In the traceback, anaconda stops like this (I copied what I hope is the useful information): AttributeError: 'NoneType' object has no attribute 'size' /var/lib/anaconda/partRequests.py line 774, getDevice if (r.size > 0) or (r.device is not None): /var/lib/anaconda/partRequests.py line 192, toEntry device = self.getDevice (partitions) /var/lib/anaconda/partitioning.py line 80, partitioningComplete entry = request.toEntry(anaconda.id.partitions) What is wrong with the code? I haven't found anyone with this same error by googling. Plenty of folks have had trouble reusing existing devices in the past, but they didn't experience tracebacks. Thanks for looking. John # disk: for reinstall bootloader --location=mbr --driveorder=sda,sdb clearpart --none # /boot: raid1/md0 on sd[ab]1 (100MB) part raid.00 --onpart=sda1 --noformat part raid.01 --onpart=sdb1 --noformat raid /boot --fstype ext3 --level=1 --device=md0 raid.00 raid.01 # /: lvm/dom0 (5G) on raid1/md1 (15G) on sd[ab]2 part raid.10 --onpart=sda2 --noformat part raid.11 --onpart=sdb2 --noformat #raid pv.01 --level=1 --device=md1 raid.10 raid.11 # original line at first install raid pv.01 --level=1 --device=md1 raid.10 raid.11 --noformat #volgroup dom0 pv.01 # original line at first install volgroup dom0 pv.01 --noformat #logvol / --name=root --vgname=dom0 --size=5000 --fstype=ext3 # original line at first install logvol / --name=root --vgname=dom0 --useexisting --fstype=ext3 # swap part swap --onpart=sda3 part swap --onpart=sdb3 # vg1 part raid.20 --onpart=sda4 --noformat part raid.21 --onpart=sdb4 --noformat raid pv.02 --noformat --level=1 --device=md2 raid.20 raid.21 part raid.30 --onpart=sdc1 --noformat part raid.31 --onpart=sdd1 --noformat raid pv.03 --noformat --level=1 --device=md3 raid.30 raid.31 #volgroup vg1 pv.02 pv.03 --useexisting --noformat #logvol /d/distro/centos --fstype=ext3 --vgname=vg1 --name=dist-centos --useexisting --noformat #logvol /d/distro/fedora --fstype=ext3 --vgname=vg1 --name=dist-fedora --useexisting --noformat #logvol /d/distro/as --fstype=ext3 --vgname=vg1 --name=dist-as --useexisting --noformat From debian at herakles.homelinux.org Mon Feb 18 13:32:03 2008 From: debian at herakles.homelinux.org (John Summerfield) Date: Mon, 18 Feb 2008 22:32:03 +0900 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <200802142019.28411.dwight@supercomputer.org> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802140919.16201.dwight@supercomputer.org> <47B4D413.6050200@herakles.homelinux.org> <200802142019.28411.dwight@supercomputer.org> Message-ID: <47B988D3.9090201@herakles.homelinux.org> dwight at supercomputer.org wrote: > It showed up here as: > > dwight wrote: > [Your words] > > > [My original words] Since my words were obviously not quoted text, I didn't see any likelihood of confusion. I did it that way to make it clear that it wasn't my originally question, that it was being discussed elsewhere but that I'd like discussion in _this_ forum. -- Cheers John -- spambait 1aaaaaaa at coco.merseine.nu Z1aaaaaaa at coco.merseine.nu -- Advice http://webfoot.com/advice/email.top.php http://www.catb.org/~esr/faqs/smart-questions.html http://support.microsoft.com/kb/555375 You cannot reply off-list:-) From jman at ablesky.com Wed Feb 20 05:51:52 2008 From: jman at ablesky.com (John Morris) Date: Wed, 20 Feb 2008 13:51:52 +0800 Subject: How to restrict package installation to x86_64 Message-ID: <47BBBFF8.40905@ablesky.com> Not sure if this is a kickstart, anaconda or yum question, so please slap and redirect me if this is the wrong forum. Our server environment is all x86_64 architecture CentOS 5.1, and thus far, new kickstarts have had only a few i386 arch RPMs installed. However, now that the nss_ldap package is included in the ks.cfg file, it is installed in both x86_64 and i386 architectures. The i386 version brings a bevy of other supporting packages with it. I wouldn't care that much, except that we have a repo of customized packages, all x86_64, and don't build for the i386 architecture; however, one of the supporting packages that we've upgraded in our custom repo, openldap, breaks the install because the original CentOS openldap package of an earlier version has file conflicts with our newer version. In kickstart, is there a simple way to restrict package installs to x86_64 architecture only? Or is there a different, 'correct' way to go? The other solutions I've considered, such as rebuilding our repo for i386 or removing all i386 packages from the vendor repos, seem hackish and inconvenient. Thanks- John From Gerrard.Geldenhuis at datacash.com Wed Feb 20 17:10:51 2008 From: Gerrard.Geldenhuis at datacash.com (Gerrard Geldenhuis) Date: Wed, 20 Feb 2008 17:10:51 -0000 Subject: Unable to eject boot cd during kickstart install In-Reply-To: <47B46852.7040504@redhat.com> References: <2FDF8583-5523-4265-B313-D95211C19216@math.udel.edu> <200802130847.45928.dwight@supercomputer.org><47B42C8D.4050500@herakles.homelinux.org> <47B46852.7040504@redhat.com> Message-ID: > > The alternative is to set the BIOS behavior to only PXE if there is no > MBR, leaving PXE out of the normal boot order. > > --Michael we set pxe to by default boot from local disk. That way we can easily access pxe if need be and not re-install critical servers. Regards From klaus.steden at thomson.net Wed Feb 20 22:23:11 2008 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 20 Feb 2008 14:23:11 -0800 Subject: How to restrict package installation to x86_64 In-Reply-To: <47BBBFF8.40905@ablesky.com> Message-ID: Hello John, I don't know if there's a way to globally restrict packages the way you're looking for, but if it suits your purposes, you can write something like this in the '%packages' section of your ks.cfg: %packages @ Admin myfoopackage.x86_64 -myfoopackage.i386 The '-' prefix tells Kickstart not to install it. hth, Klaus On 2/19/08 9:51 PM, "John Morris" did etch on stone tablets: > Not sure if this is a kickstart, anaconda or yum question, so please > slap and redirect me if this is the wrong forum. > > Our server environment is all x86_64 architecture CentOS 5.1, and thus > far, new kickstarts have had only a few i386 arch RPMs installed. > However, now that the nss_ldap package is included in the ks.cfg file, > it is installed in both x86_64 and i386 architectures. The i386 version > brings a bevy of other supporting packages with it. I wouldn't care > that much, except that we have a repo of customized packages, all > x86_64, and don't build for the i386 architecture; however, one of the > supporting packages that we've upgraded in our custom repo, openldap, > breaks the install because the original CentOS openldap package of an > earlier version has file conflicts with our newer version. > > In kickstart, is there a simple way to restrict package installs to > x86_64 architecture only? Or is there a different, 'correct' way to > go? The other solutions I've considered, such as rebuilding our repo > for i386 or removing all i386 packages from the vendor repos, seem > hackish and inconvenient. > > Thanks- > > John > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list From jason at rampaginggeek.com Wed Feb 20 22:55:39 2008 From: jason at rampaginggeek.com (Jason Edgecombe) Date: Wed, 20 Feb 2008 17:55:39 -0500 Subject: How to restrict package installation to x86_64 In-Reply-To: References: Message-ID: <47BCAFEB.80506@rampaginggeek.com> I wonder if using the following would work: %packages glibc.x86_64 -glibc.i386 does everything require glibc as a dependency? If not could some other key package by excluded? What about removing all of the i386 rpms from your install source? Jason Klaus Steden wrote: > Hello John, > > I don't know if there's a way to globally restrict packages the way you're > looking for, but if it suits your purposes, you can write something like > this in the '%packages' section of your ks.cfg: > > %packages > @ Admin > myfoopackage.x86_64 > -myfoopackage.i386 > > The '-' prefix tells Kickstart not to install it. > > hth, > Klaus > > On 2/19/08 9:51 PM, "John Morris" did etch on stone > tablets: > > >> Not sure if this is a kickstart, anaconda or yum question, so please >> slap and redirect me if this is the wrong forum. >> >> Our server environment is all x86_64 architecture CentOS 5.1, and thus >> far, new kickstarts have had only a few i386 arch RPMs installed. >> However, now that the nss_ldap package is included in the ks.cfg file, >> it is installed in both x86_64 and i386 architectures. The i386 version >> brings a bevy of other supporting packages with it. I wouldn't care >> that much, except that we have a repo of customized packages, all >> x86_64, and don't build for the i386 architecture; however, one of the >> supporting packages that we've upgraded in our custom repo, openldap, >> breaks the install because the original CentOS openldap package of an >> earlier version has file conflicts with our newer version. >> >> In kickstart, is there a simple way to restrict package installs to >> x86_64 architecture only? Or is there a different, 'correct' way to >> go? The other solutions I've considered, such as rebuilding our repo >> for i386 or removing all i386 packages from the vendor repos, seem >> hackish and inconvenient. >> >> Thanks- >> >> John >> >> _______________________________________________ >> 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 klaus.steden at thomson.net Wed Feb 20 23:05:54 2008 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 20 Feb 2008 15:05:54 -0800 Subject: How to restrict package installation to x86_64 In-Reply-To: <47BCAFEB.80506@rampaginggeek.com> Message-ID: Hi Jason, Given the importance of libc.so to pretty much everything, it would stand to reason that excluding it would prevent any apps from installing ... but then you can tell Kickstart to pull dependencies during installation, so that could undo the exclusion. I've never tried that aggressive an approach to package exclusion; I would just explicitly reference the i386 packages I didn't want and reference the x86_64 versions that I did want ... it worked well enough, although it wasn't what I would call an elegant solution. Klaus On 2/20/08 2:55 PM, "Jason Edgecombe" did etch on stone tablets: > I wonder if using the following would work: > > %packages > glibc.x86_64 > -glibc.i386 > > does everything require glibc as a dependency? If not could some other > key package by excluded? > What about removing all of the i386 rpms from your install source? > > Jason > > Klaus Steden wrote: >> Hello John, >> >> I don't know if there's a way to globally restrict packages the way you're >> looking for, but if it suits your purposes, you can write something like >> this in the '%packages' section of your ks.cfg: >> >> %packages >> @ Admin >> myfoopackage.x86_64 >> -myfoopackage.i386 >> >> The '-' prefix tells Kickstart not to install it. >> >> hth, >> Klaus >> >> On 2/19/08 9:51 PM, "John Morris" did etch on stone >> tablets: >> >> >>> Not sure if this is a kickstart, anaconda or yum question, so please >>> slap and redirect me if this is the wrong forum. >>> >>> Our server environment is all x86_64 architecture CentOS 5.1, and thus >>> far, new kickstarts have had only a few i386 arch RPMs installed. >>> However, now that the nss_ldap package is included in the ks.cfg file, >>> it is installed in both x86_64 and i386 architectures. The i386 version >>> brings a bevy of other supporting packages with it. I wouldn't care >>> that much, except that we have a repo of customized packages, all >>> x86_64, and don't build for the i386 architecture; however, one of the >>> supporting packages that we've upgraded in our custom repo, openldap, >>> breaks the install because the original CentOS openldap package of an >>> earlier version has file conflicts with our newer version. >>> >>> In kickstart, is there a simple way to restrict package installs to >>> x86_64 architecture only? Or is there a different, 'correct' way to >>> go? The other solutions I've considered, such as rebuilding our repo >>> for i386 or removing all i386 packages from the vendor repos, seem >>> hackish and inconvenient. >>> >>> Thanks- >>> >>> John >>> >>> _______________________________________________ >>> 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 Vivek.Kalia at Euroclear.com Thu Feb 21 18:41:53 2008 From: Vivek.Kalia at Euroclear.com (Vivek.Kalia at Euroclear.com) Date: Thu, 21 Feb 2008 18:41:53 +0000 Subject: Vivek Kalia is out of the office. Message-ID: I will be out of the office starting 21/02/2008 and will not return until 25/02/2008. Please contact Mario Nevrides, Adam Burton or James Connolly in my abscence. This e-mail message, including any attachments transmitted with it, is CONFIDENTIAL and may contain legally privileged information. This message is intended solely for the use of the individual or entity to whom it is addressed. If you are not the intended recipient you should not read, copy, distribute, disclose or otherwise use this information. If you have received this message in error, please notify us immediately and delete it from your system. The integrity and security of this message cannot be guaranteed and it may be subject to data corruption and unauthorised amendment, for which we accept no liability. Euroclear reserves the right to retain email messages on its systems and to the extent and under circumstances permitted by applicable law, to monitor and intercept email messages to and from its systems. Euroclear is the marketing name for the Euroclear System, Euroclear plc, Euroclear SA/NV and their affiliates. Please visit our website: http://www.euroclear.com From jman at ablesky.com Sat Feb 23 17:08:56 2008 From: jman at ablesky.com (John Morris) Date: Sun, 24 Feb 2008 01:08:56 +0800 Subject: How to restrict package installation to x86_64 In-Reply-To: References: Message-ID: <47C05328.4080306@ablesky.com> Hi, folks, thanks for the good suggestions. I'm following up on them as time allows. > Given the importance of libc.so to pretty much everything, it would stand to > reason that excluding it would prevent any apps from installing ... but then > you can tell Kickstart to pull dependencies during installation, so that > could undo the exclusion. And pulling the dependencies back in is exactly what happens, unfortunately. I put this in my kickstart.cfg file under %packages: >> %packages >> glibc.x86_64 >> -glibc.i386 but that later i386 packages drag the glibc.i386 package back in: [root at sandbox ~]# rpm -qa --qf '%{name}-%{arch}\n' | grep "glibc-[xi]" glibc-x86_64 glibc-i686 >> does everything require glibc as a dependency? If not could some other >> key package by excluded? >> I think most non-'noarch' packages do depend on glibc. But unfortunately those are the ones that drag glibc.i386 back in. >>> I don't know if there's a way to globally restrict packages the way you're >>> looking for, but if it suits your purposes, you can write something like >>> this in the '%packages' section of your ks.cfg: >>> >>> %packages >>> @ Admin >>> myfoopackage.x86_64 >>> -myfoopackage.i386 >>> Putting the specific unwanted packages into my ks.cfg file would certainly work, but would require me first to figure out which package is dragging in all the others, or else to put all i386 arch packages into my ks.cfg file. I'm trying to find a way that's more maintainable and less messy. >>> What about removing all of the i386 rpms from your install source? So, after saying I wouldn't take this approach in the previous mail, this may turn out to be the easiest thing to do. We do have our own mirror of the repos we use (the fastest non-local mirrors from Beijing are just painfully slow...). In our nightly cronjob that brings down the updates, we can just add an option to rsync to ignore all files ending with .i?86.rpm, and put in an extra 'createrepo' line. Thanks again for the suggestions. I'm a bit disappointed that kickstart doesn't have a simple mechanism to achieve this. It appears that yum's 'exclude' option may help with this after kickstart. I would think that some folks might have legitimate reason to want this. My 'legitimate' reason, again, is simple laziness. I've been getting my systems under configuration control with bcfg2, and when there are two packages of the same name but different architecture there, I have to type three extra lines per package to tell the system that that's OK. I'm definitely going to take this up with bcfg2's author, of course, but it seems that the problem stems from kickstart. John From jman at ablesky.com Sun Feb 24 02:14:43 2008 From: jman at ablesky.com (John Morris) Date: Sun, 24 Feb 2008 10:14:43 +0800 Subject: How to restrict package installation to x86_64 In-Reply-To: <47C05328.4080306@ablesky.com> References: <47C05328.4080306@ablesky.com> Message-ID: <47C0D313.9080308@ablesky.com> > It appears that yum's 'exclude' option may help with this after kickstart. Another incorrect statement of mine. I was unable to get the 'exclude' option to ignore i386 packages too, with the following line in /etc/yum.conf: exclude=*.i?86.rpm Ah well. John From jkeating at j2solutions.net Sun Feb 24 14:58:20 2008 From: jkeating at j2solutions.net (Jesse Keating) Date: Sun, 24 Feb 2008 09:58:20 -0500 Subject: How to restrict package installation to x86_64 In-Reply-To: <47C0D313.9080308@ablesky.com> References: <47C05328.4080306@ablesky.com> <47C0D313.9080308@ablesky.com> Message-ID: <20080224095820.2572ea32@j2solutions.net> On Sun, 24 Feb 2008 10:14:43 +0800 John Morris wrote: > exclude=*.i?86.rpm Drop the .rpm -- Jesse Keating RHCE (jkeating.livejournal.com) Fedora Project (fedoraproject.org/wiki/JesseKeating) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: not available URL: From rainer at ultra-secure.de Sun Feb 24 17:47:14 2008 From: rainer at ultra-secure.de (Rainer Duffner) Date: Sun, 24 Feb 2008 18:47:14 +0100 Subject: How to restrict package installation to x86_64 In-Reply-To: <47BBBFF8.40905@ablesky.com> References: <47BBBFF8.40905@ablesky.com> Message-ID: <47C1ADA2.4010107@ultra-secure.de> John Morris schrieb: > Not sure if this is a kickstart, anaconda or yum question, so please > slap and redirect me if this is the wrong forum. I have got a related question. Why was the naming-system chosen like it is anyway? Why weren't the i386 libraries named as "compat-something"? When I do a rpm -qi _package_ it isn't even obvious, which one is which! Why are there so many i386-rpms in the x86_64 release anyway? Wouldn't it be better to make everything 64bit and just leave out the few things that can't yet compile? Who needs OpenOffice on a server anyway? cheers, Rainer From jman at ablesky.com Mon Feb 25 03:13:21 2008 From: jman at ablesky.com (John Morris) Date: Mon, 25 Feb 2008 11:13:21 +0800 Subject: How to restrict package installation to x86_64 In-Reply-To: <47C1ADA2.4010107@ultra-secure.de> References: <47BBBFF8.40905@ablesky.com> <47C1ADA2.4010107@ultra-secure.de> Message-ID: <47C23251.2030001@ablesky.com> Rainer Duffner wrote: > Why are there so many i386-rpms in the x86_64 release anyway? > > Wouldn't it be better to make everything 64bit and just leave out the > few things that can't yet compile? > Who needs OpenOffice on a server anyway? Rainer, I can't speak to your other questions, but I can tell you why our company needs i386 libs on x86_64. Most of our servers run OSS in RPM form only, and on those machines we don't need the i386 libs. But we have a few machines that run proprietary, vendor-supplied binaries that are compiled for the i386 architecture only. These software, though important, don't warrant their own i386 architecture servers. In fact, we're running a lot of things in virtual servers, and with the vendor-supplied xen software, running i386 guests on an x86_64 host hasn't been very stable. So, this is one usage case for you. John From jkeating at j2solutions.net Mon Feb 25 15:40:25 2008 From: jkeating at j2solutions.net (Jesse Keating) Date: Mon, 25 Feb 2008 10:40:25 -0500 Subject: How to restrict package installation to x86_64 In-Reply-To: <47C1ADA2.4010107@ultra-secure.de> References: <47BBBFF8.40905@ablesky.com> <47C1ADA2.4010107@ultra-secure.de> Message-ID: <20080225104025.1b279a24@j2solutions.net> On Sun, 24 Feb 2008 18:47:14 +0100 Rainer Duffner wrote: > Why weren't the i386 libraries named as "compat-something"? Because that wouldn't make any sense on i386. Those are the exact same packages from the i386 tree just copied into the x86_64 tree. > When I do a rpm -qi _package_ it isn't even obvious, which one is > which! Use rpm -qi package.arch -- Jesse Keating RHCE (jkeating.livejournal.com) Fedora Project (fedoraproject.org/wiki/JesseKeating) GPG Public Key (geek.j2solutions.net/jkeating.j2solutions.pub) -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: not available URL: From Vivek.Kalia at Euroclear.com Mon Feb 25 22:02:43 2008 From: Vivek.Kalia at Euroclear.com (Vivek.Kalia at Euroclear.com) Date: Mon, 25 Feb 2008 22:02:43 +0000 Subject: Vivek Kalia is out of the office (working from home) Message-ID: I will be out of the office starting 25/02/2008 and will not return until 03/03/2008. I will respond to your message, but if you need quicker response, please contact Mario Nevrides, Adam Burton or James Connolly in my abscence. This e-mail message, including any attachments transmitted with it, is CONFIDENTIAL and may contain legally privileged information. This message is intended solely for the use of the individual or entity to whom it is addressed. If you are not the intended recipient you should not read, copy, distribute, disclose or otherwise use this information. If you have received this message in error, please notify us immediately and delete it from your system. The integrity and security of this message cannot be guaranteed and it may be subject to data corruption and unauthorised amendment, for which we accept no liability. Euroclear reserves the right to retain email messages on its systems and to the extent and under circumstances permitted by applicable law, to monitor and intercept email messages to and from its systems. Euroclear is the marketing name for the Euroclear System, Euroclear plc, Euroclear SA/NV and their affiliates. Please visit our website: http://www.euroclear.com From jman at ablesky.com Tue Feb 26 03:49:04 2008 From: jman at ablesky.com (John Morris) Date: Tue, 26 Feb 2008 11:49:04 +0800 Subject: How to restrict package installation to x86_64 In-Reply-To: <20080225104025.1b279a24@j2solutions.net> References: <47BBBFF8.40905@ablesky.com> <47C1ADA2.4010107@ultra-secure.de> <20080225104025.1b279a24@j2solutions.net> Message-ID: <47C38C30.1060608@ablesky.com> Jesse Keating wrote: > On Sun, 24 Feb 2008 18:47:14 +0100 > Rainer Duffner wrote: > >> When I do a rpm -qi _package_ it isn't even obvious, which one is >> which! >> > > Use rpm -qi package.arch > The '-i' option is for 'info'. The info should be exactly the same for both i?86 and x86_64 arch packages. If you do have a reason to want this information, though, you can use the following query, in this case for the libgcc package: rpm -qi --qf '%{name}-%{version}-%{release}-%{arch}\n' libgcc John From rbrown at BallistaSec.com Tue Feb 26 22:36:52 2008 From: rbrown at BallistaSec.com (Rodrick Brown) Date: Tue, 26 Feb 2008 17:36:52 -0500 Subject: Getting started Message-ID: I'm very new to kick start and was wondering if there was a comprehensive set of documents available that one could easily follow to build a custom kick start profile and server. I found many inconsistencies with the slew of documents referenced by Google thanks. --- Rodrick R. Brown -------------- next part -------------- An HTML attachment was scrubbed... URL: From Chip.Shabazian at bankofamerica.com Tue Feb 26 22:48:39 2008 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Tue, 26 Feb 2008 14:48:39 -0800 Subject: Getting started In-Reply-To: Message-ID: You could start with the presentation I gave at last years LinuxWorld: http://www.shabazian.com/lw2007.pdf ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Rodrick Brown Sent: Tuesday, February 26, 2008 2:37 PM To: kickstart-list at redhat.com Subject: Getting started I'm very new to kick start and was wondering if there was a comprehensive set of documents available that one could easily follow to build a custom kick start profile and server. I found many inconsistencies with the slew of documents referenced by Google thanks. --- Rodrick R. Brown -------------- next part -------------- An HTML attachment was scrubbed... URL: From herrold at owlriver.com Tue Feb 26 22:56:32 2008 From: herrold at owlriver.com (R P Herrold) Date: Tue, 26 Feb 2008 17:56:32 -0500 (EST) Subject: Getting started In-Reply-To: References: Message-ID: On Tue, 26 Feb 2008, Shabazian, Chip wrote: > You could start with the presentation I gave at last years LinuxWorld: > http://www.shabazian.com/lw2007.pdf Well recognized as the best, most recent work in the field; see also the older step-by-step walkthrough materials at: http://www.owlriver.com/tips/ hi, Chip ;) -- Russ herrold From saetaes at gmail.com Wed Feb 27 00:23:03 2008 From: saetaes at gmail.com (Mike M) Date: Tue, 26 Feb 2008 19:23:03 -0500 Subject: Getting started In-Reply-To: References: Message-ID: <8552c3a30802261623l704f4ec5l12d5cc1503f4d412@mail.gmail.com> On Tue, Feb 26, 2008 at 5:36 PM, Rodrick Brown wrote: > I'm very new to kick start and was wondering if there was a comprehensive > set of documents available that one could easily follow to build a custom > kick start profile and server. You should take a look at Cobbler: http://cobbler.et.redhat.com/ It makes getting kickstart up and running dead easy. Mike From mdehaan at redhat.com Wed Feb 27 15:15:14 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 10:15:14 -0500 Subject: Getting started In-Reply-To: References: Message-ID: <47C57E82.3090904@redhat.com> Shabazian, Chip wrote: > You could start with the presentation I gave at last years LinuxWorld: > http://www.shabazian.com/lw2007.pdf > +1. My biggest tip is to "yum install system-config-kickstart" (which I believe Chip mentions). It can help guide you through the install questions and builds a kickstart along the way. That's a good way to see what the various options are (and is a bit easier than reading some of the docs for the first couple of times -- eventually you probably won't need the tool). From Rudi at SoftDux.com Wed Feb 27 15:20:48 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 17:20:48 +0200 Subject: Getting started In-Reply-To: <47C57E82.3090904@redhat.com> References: <47C57E82.3090904@redhat.com> Message-ID: <47C57FD0.6040508@SoftDux.com> Michael DeHaan wrote: > Shabazian, Chip wrote: >> You could start with the presentation I gave at last years LinuxWorld: >> http://www.shabazian.com/lw2007.pdf >> > +1. > > My biggest tip is to "yum install system-config-kickstart" (which I > believe Chip mentions). > > It can help guide you through the install questions and builds a > kickstart along the way. That's a good way to see what > the various options are (and is a bit easier than reading some of the > docs for the first couple of times -- eventually you > probably won't need the tool). > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list That's a good tip, but for people who don't have X installed (for example setting up servers), the GUI based kickstart config tool won't work. I also had to go through this exercise but soon found a lot of sites on the net with good sample kickstart files. One thing I want to add though. If it's possible, setup sometimes like XEN / VMWare / Virtualbox / QMU / etc to test the CD ISO. It's a pain in the neck to first create the CD ISO with makeisofs, then burn it to CD, and then putting it into my test machine to test it. Kind Regards Rudi Ahlers Web: http://www.SoftDux.com Forum: http://Forum.SoftDux.com Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za From mdehaan at redhat.com Wed Feb 27 15:28:34 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 10:28:34 -0500 Subject: Getting started In-Reply-To: <47C57FD0.6040508@SoftDux.com> References: <47C57E82.3090904@redhat.com> <47C57FD0.6040508@SoftDux.com> Message-ID: <47C581A2.1080309@redhat.com> Rudi Ahlers wrote: > > > Michael DeHaan wrote: >> Shabazian, Chip wrote: >>> You could start with the presentation I gave at last years LinuxWorld: >>> http://www.shabazian.com/lw2007.pdf >>> >> +1. >> >> My biggest tip is to "yum install system-config-kickstart" (which I >> believe Chip mentions). >> >> It can help guide you through the install questions and builds a >> kickstart along the way. That's a good way to see what >> the various options are (and is a bit easier than reading some of the >> docs for the first couple of times -- eventually you >> probably won't need the tool). >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list > > That's a good tip, but for people who don't have X installed (for > example setting up servers), the GUI based kickstart config tool won't > work. I also had to go through this exercise but soon found a lot of > sites on the net with good sample kickstart files. One thing I want to > add though. If it's possible, setup sometimes like XEN / VMWare / > Virtualbox / QMU / etc to test the CD ISO. It's a pain in the neck to > first create the CD ISO with makeisofs, then burn it to CD, and then > putting it into my test machine to test it. True! You can of course run system-config-kickstart on any box though, so as long as you have at least one Linux box running X, you can use that :) With respect to the other suggestions, cobbler happens to both make installing Xen/KVM very easy for repeated installs as you tweak kickstarts and also includes sample base kickstarts :) --Michael From Gerrard.Geldenhuis at datacash.com Wed Feb 27 15:32:42 2008 From: Gerrard.Geldenhuis at datacash.com (Gerrard Geldenhuis) Date: Wed, 27 Feb 2008 15:32:42 -0000 Subject: Getting started In-Reply-To: References: Message-ID: I must say this presentation notes is really great. I wish the Redhat docs could be more up to this standard. Regards ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Shabazian, Chip Sent: 26 February 2008 22:49 To: Discussion list about Kickstart Subject: RE: Getting started You could start with the presentation I gave at last years LinuxWorld: http://www.shabazian.com/lw2007.pdf ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Rodrick Brown Sent: Tuesday, February 26, 2008 2:37 PM To: kickstart-list at redhat.com Subject: Getting started I'm very new to kick start and was wondering if there was a comprehensive set of documents available that one could easily follow to build a custom kick start profile and server. I found many inconsistencies with the slew of documents referenced by Google thanks. --- Rodrick R. Brown -------------- next part -------------- An HTML attachment was scrubbed... URL: From Rudi at SoftDux.com Wed Feb 27 15:33:11 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 17:33:11 +0200 Subject: Getting started In-Reply-To: <47C581A2.1080309@redhat.com> References: <47C57E82.3090904@redhat.com> <47C57FD0.6040508@SoftDux.com> <47C581A2.1080309@redhat.com> Message-ID: <47C582B7.6050101@SoftDux.com> Michael DeHaan wrote: > Rudi Ahlers wrote: >> >> >> Michael DeHaan wrote: >>> Shabazian, Chip wrote: >>>> You could start with the presentation I gave at last years LinuxWorld: >>>> http://www.shabazian.com/lw2007.pdf >>>> >>> +1. >>> >>> My biggest tip is to "yum install system-config-kickstart" (which I >>> believe Chip mentions). >>> >>> It can help guide you through the install questions and builds a >>> kickstart along the way. That's a good way to see what >>> the various options are (and is a bit easier than reading some of >>> the docs for the first couple of times -- eventually you >>> probably won't need the tool). >>> >>> _______________________________________________ >>> Kickstart-list mailing list >>> Kickstart-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/kickstart-list >> >> That's a good tip, but for people who don't have X installed (for >> example setting up servers), the GUI based kickstart config tool >> won't work. I also had to go through this exercise but soon found a >> lot of sites on the net with good sample kickstart files. One thing I >> want to add though. If it's possible, setup sometimes like XEN / >> VMWare / Virtualbox / QMU / etc to test the CD ISO. It's a pain in >> the neck to first create the CD ISO with makeisofs, then burn it to >> CD, and then putting it into my test machine to test it. > > True! You can of course run system-config-kickstart on any box > though, so as long as you have at least one Linux box running X, you > can use that :) > > With respect to the other suggestions, cobbler happens to both make > installing Xen/KVM very easy for repeated installs as you tweak > kickstarts and also includes sample base kickstarts :) > > --Michael > > _______________________________________________ Does it? I tried using the Fedora Core 7 kickstart GUI, but the package listing was different, so I didn't want to add / remove stuff which CentOS didn't have Kind Regards Rudi Ahlers Web: http://www.SoftDux.com Forum: http://Forum.SoftDux.com Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za From Rudi at SoftDux.com Wed Feb 27 15:36:26 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 17:36:26 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? Message-ID: <47C5837A.4060604@SoftDux.com> Hi all I'm busy creating a custom CentOS installation CD to be used with cPanel, which basically needs a minimum installation of Linux, but I would also like to use Xen for VPS' on the server, and thought I could do both at the same time - automatically install my server to my needs, and install Xen (without needing to run "yum install xen kernel-xen virt-manager" every time). So, my question is. How would I add those RPM's to the custom CentOS CD, and install them, and then setup grub.conf to use the Xen kernel upon reboot as well? -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Forum: http://Forum.SoftDux.com Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za From mdehaan at redhat.com Wed Feb 27 15:37:10 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 10:37:10 -0500 Subject: Getting started In-Reply-To: <47C582B7.6050101@SoftDux.com> References: <47C57E82.3090904@redhat.com> <47C57FD0.6040508@SoftDux.com> <47C581A2.1080309@redhat.com> <47C582B7.6050101@SoftDux.com> Message-ID: <47C583A6.6080803@redhat.com> Rudi Ahlers wrote: > Michael DeHaan wrote: >> Rudi Ahlers wrote: >>> >>> >>> Michael DeHaan wrote: >>>> Shabazian, Chip wrote: >>>>> You could start with the presentation I gave at last years >>>>> LinuxWorld: >>>>> http://www.shabazian.com/lw2007.pdf >>>>> >>>> +1. >>>> >>>> My biggest tip is to "yum install system-config-kickstart" (which I >>>> believe Chip mentions). >>>> >>>> It can help guide you through the install questions and builds a >>>> kickstart along the way. That's a good way to see what >>>> the various options are (and is a bit easier than reading some of >>>> the docs for the first couple of times -- eventually you >>>> probably won't need the tool). >>>> >>>> _______________________________________________ >>>> Kickstart-list mailing list >>>> Kickstart-list at redhat.com >>>> https://www.redhat.com/mailman/listinfo/kickstart-list >>> >>> That's a good tip, but for people who don't have X installed (for >>> example setting up servers), the GUI based kickstart config tool >>> won't work. I also had to go through this exercise but soon found a >>> lot of sites on the net with good sample kickstart files. One thing >>> I want to add though. If it's possible, setup sometimes like XEN / >>> VMWare / Virtualbox / QMU / etc to test the CD ISO. It's a pain in >>> the neck to first create the CD ISO with makeisofs, then burn it to >>> CD, and then putting it into my test machine to test it. >> >> True! You can of course run system-config-kickstart on any box >> though, so as long as you have at least one Linux box running X, you >> can use that :) >> >> With respect to the other suggestions, cobbler happens to both make >> installing Xen/KVM very easy for repeated installs as you tweak >> kickstarts and also includes sample base kickstarts :) >> >> --Michael >> >> _______________________________________________ > > > Does it? I tried using the Fedora Core 7 kickstart GUI, but the > package listing was different, so I didn't want to add / remove stuff > which CentOS didn't have > > Kind Regards > Rudi Ahlers It's close enough. Just do a base package-set and add what you need later. The main change in kickstart between versions is the addition of being able to use yum repos at install time -- which is availabe in FC6/F7 (which is "Fedora 7" no longer "Fedora Core", BTW) and later, and therefore not in EL4. From Pablo.Iranzo at redhat.com Wed Feb 27 15:40:55 2008 From: Pablo.Iranzo at redhat.com (Pablo Iranzo =?ISO-8859-1?Q?G=F3mez?=) Date: Wed, 27 Feb 2008 16:40:55 +0100 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C5837A.4060604@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> Message-ID: <1204126855.11610.3.camel@iranzo.users.redhat.com> Have you tried: @ Virtualization at %packages? For me, that installs Xen kernel and utils as default for system where it is installed. Regards Pablo El mi?, 27-02-2008 a las 17:36 +0200, Rudi Ahlers escribi?: > Hi all > > I'm busy creating a custom CentOS installation CD to be used with > cPanel, which basically needs a minimum installation of Linux, but I > would also like to use Xen for VPS' on the server, and thought I could > do both at the same time - automatically install my server to my needs, > and install Xen (without needing to run "yum install xen kernel-xen > virt-manager" every time). > > So, my question is. How would I add those RPM's to the custom CentOS CD, > and install them, and then setup grub.conf to use the Xen kernel upon > reboot as well? > -- Pablo Iranzo G?mez (Pablo.Iranzo at redhat.com) RHCE/Global Profesional Services Consultant Spain Phone: +34 645 01 01 49 (CET/CEST) GnuPG KeyID: 0xFAD3CF0D --- Direcci?n Comercial: C/Jose Bardasano Baos, 9, Edif. Gorbea 3, planta 3?D, 28016 Madrid, Spain Direcci?n Registrada: Red Hat S.L., C/ Velazquez 63, Madrid 28001, Spain Inscrita en el Reg. Mercantil de Madrid ? C.I.F. B82657941 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Esta parte del mensaje est? firmada digitalmente URL: From mdehaan at redhat.com Wed Feb 27 15:42:17 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 10:42:17 -0500 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C5837A.4060604@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> Message-ID: <47C584D9.4080101@redhat.com> Rudi Ahlers wrote: > Hi all > > I'm busy creating a custom CentOS installation CD to be used with > cPanel, which basically needs a minimum installation of Linux, but I > would also like to use Xen for VPS' on the server, and thought I could > do both at the same time - automatically install my server to my > needs, and install Xen (without needing to run "yum install xen > kernel-xen virt-manager" every time). > > So, my question is. How would I add those RPM's to the custom CentOS > CD, and install them, and then setup grub.conf to use the Xen kernel > upon reboot as well? > All you have to do is add the list of packages you regularly yum-install to your kickstart file: For your dom0 (host OS): %packages kernel-xen xen virsh koan virt-manager etc... From Rudi at SoftDux.com Wed Feb 27 15:43:14 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 17:43:14 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <1204126855.11610.3.camel@iranzo.users.redhat.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> Message-ID: <47C58512.5010103@SoftDux.com> Pablo Iranzo G?mez wrote: > Have you tried: > > @ Virtualization at %packages? > > For me, that installs Xen kernel and utils as default for system where > it is installed. > > Regards > Pablo > > El mi??, 27-02-2008 a las 17:36 +0200, Rudi Ahlers escribi??: > >> Hi all >> >> I'm busy creating a custom CentOS installation CD to be used with >> cPanel, which basically needs a minimum installation of Linux, but I >> would also like to use Xen for VPS' on the server, and thought I could >> do both at the same time - automatically install my server to my needs, >> and install Xen (without needing to run "yum install xen kernel-xen >> virt-manager" every time). >> >> So, my question is. How would I add those RPM's to the custom CentOS CD, >> and install them, and then setup grub.conf to use the Xen kernel upon >> reboot as well? >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list I have, but since I only have the 1st CD, it's not on the CD, and I don't want to download all the other CD's to see on which CD it is. It comes to a point where it's asking for the 2nd CD. I need a bare minimal install for cPanel, so I didn't even include any packages Kind Regards Rudi Ahlers Web: http://www.SoftDux.com Forum: http://Forum.SoftDux.com Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za From mdehaan at redhat.com Wed Feb 27 15:45:33 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 10:45:33 -0500 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C584D9.4080101@redhat.com> References: <47C5837A.4060604@SoftDux.com> <47C584D9.4080101@redhat.com> Message-ID: <47C5859D.4010202@redhat.com> Michael DeHaan wrote: > Rudi Ahlers wrote: >> Hi all >> >> I'm busy creating a custom CentOS installation CD to be used with >> cPanel, which basically needs a minimum installation of Linux, but I >> would also like to use Xen for VPS' on the server, and thought I >> could do both at the same time - automatically install my server to >> my needs, and install Xen (without needing to run "yum install xen >> kernel-xen virt-manager" every time). >> >> So, my question is. How would I add those RPM's to the custom CentOS >> CD, and install them, and then setup grub.conf to use the Xen kernel >> upon reboot as well? >> > All you have to do is add the list of packages you regularly > yum-install to your kickstart file: > > For your dom0 (host OS): > > %packages > kernel-xen > xen > virsh > koan > virt-manager > > etc... > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list Ah, you said CD, my fault. In Fedora I'd use Pungi, though since you also mentioned cPanel it would be better to set up a network installation environment. Even if you can't PXE, koan has live CD capability to provide the same feature set... --Michael From Pablo.Iranzo at redhat.com Wed Feb 27 15:47:23 2008 From: Pablo.Iranzo at redhat.com (Pablo Iranzo =?ISO-8859-1?Q?G=F3mez?=) Date: Wed, 27 Feb 2008 16:47:23 +0100 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C58512.5010103@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> Message-ID: <1204127243.11610.6.camel@iranzo.users.redhat.com> Well, the easiest part would be to create a custom install CD with just the packages you need ;-) and then, rerun createrepo to update the packages list and do what Mr Dehaan said, listing individual packages at %packages instead of the "Virtualization" group If you already have them, you can always put a %post with an rpm -Uvh package1.rpm package2.rpm, etc if you don't want to do the createrepo stuff Regards Pablo El mi?, 27-02-2008 a las 17:43 +0200, Rudi Ahlers escribi?: > > Pablo Iranzo G?mez wrote: > > Have you tried: > > > > @ Virtualization at %packages? > > > > For me, that installs Xen kernel and utils as default for system where > > it is installed. > > > > Regards > > Pablo > > > > El mi??, 27-02-2008 a las 17:36 +0200, Rudi Ahlers escribi??: > > > >> Hi all > >> > >> I'm busy creating a custom CentOS installation CD to be used with > >> cPanel, which basically needs a minimum installation of Linux, but I > >> would also like to use Xen for VPS' on the server, and thought I could > >> do both at the same time - automatically install my server to my needs, > >> and install Xen (without needing to run "yum install xen kernel-xen > >> virt-manager" every time). > >> > >> So, my question is. How would I add those RPM's to the custom CentOS CD, > >> and install them, and then setup grub.conf to use the Xen kernel upon > >> reboot as well? > >> > >> > >> ------------------------------------------------------------------------ > >> > >> _______________________________________________ > >> Kickstart-list mailing list > >> Kickstart-list at redhat.com > >> https://www.redhat.com/mailman/listinfo/kickstart-list > > I have, but since I only have the 1st CD, it's not on the CD, and I > don't want to download all the other CD's to see on which CD it is. It > comes to a point where it's asking for the 2nd CD. > > I need a bare minimal install for cPanel, so I didn't even include any > packages > > Kind Regards > Rudi Ahlers > > Web: http://www.SoftDux.com > Forum: http://Forum.SoftDux.com > > Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- Pablo Iranzo G?mez (Pablo.Iranzo at redhat.com) RHCE/Global Profesional Services Consultant Spain Phone: +34 645 01 01 49 (CET/CEST) GnuPG KeyID: 0xFAD3CF0D --- Direcci?n Comercial: C/Jose Bardasano Baos, 9, Edif. Gorbea 3, planta 3?D, 28016 Madrid, Spain Direcci?n Registrada: Red Hat S.L., C/ Velazquez 63, Madrid 28001, Spain Inscrita en el Reg. Mercantil de Madrid ? C.I.F. B82657941 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Esta parte del mensaje est? firmada digitalmente URL: From Rudi at SoftDux.com Wed Feb 27 16:09:22 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 18:09:22 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C5859D.4010202@redhat.com> References: <47C5837A.4060604@SoftDux.com> <47C584D9.4080101@redhat.com> <47C5859D.4010202@redhat.com> Message-ID: <47C58B32.1080407@SoftDux.com> Michael DeHaan wrote: > Michael DeHaan wrote: >> Rudi Ahlers wrote: >>> Hi all >>> >>> I'm busy creating a custom CentOS installation CD to be used with >>> cPanel, which basically needs a minimum installation of Linux, but I >>> would also like to use Xen for VPS' on the server, and thought I >>> could do both at the same time - automatically install my server to >>> my needs, and install Xen (without needing to run "yum install xen >>> kernel-xen virt-manager" every time). >>> >>> So, my question is. How would I add those RPM's to the custom CentOS >>> CD, and install them, and then setup grub.conf to use the Xen kernel >>> upon reboot as well? >>> >> All you have to do is add the list of packages you regularly >> yum-install to your kickstart file: >> >> For your dom0 (host OS): >> >> %packages >> kernel-xen >> xen >> virsh >> koan >> virt-manager >> >> etc... >> >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list > > Ah, you said CD, my fault. > > In Fedora I'd use Pungi, though since you also mentioned cPanel it > would be better to set up a network installation environment. Even > if you can't PXE, koan has live CD capability to provide the same > feature set... > > --Michael > > > ________________________________________ I can't entirely setup a PXE environment, since the servers are being colocated, and we can only do CD / DVD installations, or directly off the internet. And since I pay for bandwidth per MB, I'd prefer to have a CD, which I can then just give to each technician and they can install / reinstall servers from there with a specific setup Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Forum: http://Forum.SoftDux.com Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za From Rudi at SoftDux.com Wed Feb 27 16:19:55 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 18:19:55 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <1204127243.11610.6.camel@iranzo.users.redhat.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> Message-ID: <47C58DAB.7060904@SoftDux.com> Pablo Iranzo G?mez wrote: > Well, the easiest part would be to create a custom install CD with just > the packages you need ;-) and then, rerun createrepo to update the > packages list and do what Mr Dehaan said, listing individual packages at > %packages instead of the "Virtualization" group > > If you already have them, you can always put a %post with an rpm -Uvh > package1.rpm package2.rpm, etc if you don't want to do the createrepo > stuff > > Regards > Pablo > > > > El mi??, 27-02-2008 a las 17:43 +0200, Rudi Ahlers escribi??: > >> Pablo Iranzo G??mez wrote: >> >>> Have you tried: >>> >>> @ Virtualization at %packages? >>> >>> For me, that installs Xen kernel and utils as default for system where >>> it is installed. >>> >>> Regards >>> Pablo >>> >>> El mi????, 27-02-2008 a las 17:36 +0200, Rudi Ahlers escribi????: >>> >>> >>>> Hi all >>>> >>>> I'm busy creating a custom CentOS installation CD to be used with >>>> cPanel, which basically needs a minimum installation of Linux, but I >>>> would also like to use Xen for VPS' on the server, and thought I could >>>> do both at the same time - automatically install my server to my needs, >>>> and install Xen (without needing to run "yum install xen kernel-xen >>>> virt-manager" every time). >>>> >>>> So, my question is. How would I add those RPM's to the custom CentOS CD, >>>> and install them, and then setup grub.conf to use the Xen kernel upon >>>> reboot as well? >>>> >>>> >>>> ------------------------------------------------------------------------ >>>> >>>> _______________________________________________ >>>> Kickstart-list mailing list >>>> Kickstart-list at redhat.com >>>> https://www.redhat.com/mailman/listinfo/kickstart-list >>>> >> I have, but since I only have the 1st CD, it's not on the CD, and I >> don't want to download all the other CD's to see on which CD it is. It >> comes to a point where it's asking for the 2nd CD. >> >> I need a bare minimal install for cPanel, so I didn't even include any >> packages >> >> Kind Regards >> Rudi Ahlers >> >> Web: http://www.SoftDux.com >> Forum: http://Forum.SoftDux.com >> >> Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za >> >> >> >> >> _______________________________________________ >> Well, I only started playing with kickstart yesterday And my brain hurts by now, but I'm learning a lot :) How does it work? Do I copy all the rpm's I want to use to the CentOS folder, and then run createrepo on that folder? Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Forum: http://Forum.SoftDux.com Join SA WebHostingTalk today, on http://www.WebHostingTalk.co.za From mdehaan at redhat.com Wed Feb 27 16:19:48 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 11:19:48 -0500 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C58B32.1080407@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <47C584D9.4080101@redhat.com> <47C5859D.4010202@redhat.com> <47C58B32.1080407@SoftDux.com> Message-ID: <47C58DA4.1080902@redhat.com> Rudi Ahlers wrote: > > Michael DeHaan wrote: >> Michael DeHaan wrote: >>> Rudi Ahlers wrote: >>>> Hi all >>>> >>>> I'm busy creating a custom CentOS installation CD to be used with >>>> cPanel, which basically needs a minimum installation of Linux, but >>>> I would also like to use Xen for VPS' on the server, and thought I >>>> could do both at the same time - automatically install my server to >>>> my needs, and install Xen (without needing to run "yum install xen >>>> kernel-xen virt-manager" every time). >>>> >>>> So, my question is. How would I add those RPM's to the custom >>>> CentOS CD, and install them, and then setup grub.conf to use the >>>> Xen kernel upon reboot as well? >>>> >>> All you have to do is add the list of packages you regularly >>> yum-install to your kickstart file: >>> >>> For your dom0 (host OS): >>> >>> %packages >>> kernel-xen >>> xen >>> virsh >>> koan >>> virt-manager >>> >>> etc... >>> >>> >>> _______________________________________________ >>> Kickstart-list mailing list >>> Kickstart-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/kickstart-list >> >> Ah, you said CD, my fault. >> >> In Fedora I'd use Pungi, though since you also mentioned cPanel it >> would be better to set up a network installation environment. Even >> if you can't PXE, koan has live CD capability to provide the same >> feature set... >> >> --Michael >> >> >> ________________________________________ > > > I can't entirely setup a PXE environment, since the servers are being > colocated, and we can only do CD / DVD installations, or directly off > the internet. If you can't do a PXE setup, that's not a problem. You can still do bare-metal net installs without having to burn custom media. Right, with the koan Live CD you run a cobbler server on one of your colo boxes, and insert the "magic" CD into each machine. The CD then autodiscovers what to install on each system based on the Mac address, which you can control and "preprogram" any time you want without having to generate new CD's. From the one magic CD, you can install F-8 on one box, Centos 5 on one box, and RHEL 4 on another... it autodetects what to install on each. If you don't know the MAC addresses, you can make one live CD for each profile, and still you can customize that profile (add new packages, etc) without having to respin another CD. The instructions on creating this setup are here: https://fedorahosted.org/cobbler/wiki/CobblerLiveCd --Michael From Rudi at SoftDux.com Wed Feb 27 16:22:40 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 18:22:40 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <1204126855.11610.3.camel@iranzo.users.redhat.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> Message-ID: <47C58E50.5070007@SoftDux.com> Pablo Iranzo G?mez wrote: > Have you tried: > > @ Virtualization at %packages? > > For me, that installs Xen kernel and utils as default for system where > it is installed. > > Regards > Pablo > > El mi??, 27-02-2008 a las 17:36 +0200, Rudi Ahlers escribi??: > >> Hi all >> >> I'm busy creating a custom CentOS installation CD to be used with >> cPanel, which basically needs a minimum installation of Linux, but I >> would also like to use Xen for VPS' on the server, and thought I could >> do both at the same time - automatically install my server to my needs, >> and install Xen (without needing to run "yum install xen kernel-xen >> virt-manager" every time). >> >> So, my question is. How would I add those RPM's to the custom CentOS CD, >> and install them, and then setup grub.conf to use the Xen kernel upon >> reboot as well? >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list Can I just download those file from http://mirror.nsc.liu.se/CentOS/5.1/os/x86_64/CentOS/ and copy them to the CentOS folder? -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From Pablo.Iranzo at redhat.com Wed Feb 27 17:23:28 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Wed, 27 Feb 2008 18:23:28 +0100 (CET) Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C58DAB.7060904@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> Message-ID: On Wed, 27 Feb 2008, Rudi Ahlers wrote: > Well, I only started playing with kickstart yesterday And my brain hurts > by now, but I'm learning a lot :) That's nice ;) > > How does it work? Do I copy all the rpm's I want to use to the CentOS > folder, and then run createrepo on that folder? Yes, but first, you need to test dependencies in order to have "ALL REQUIRED" packages ;) After that, you can remove not needed ones. Usually one way to do that is to try installing to a different folder, not your system so you can keep rpm complaining of missing dependencies... and when everything is fine, then you can "createrepo" and master a custom cd. Are you using CentOS 5? Regards Pablo From Pablo.Iranzo at redhat.com Wed Feb 27 17:49:09 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Wed, 27 Feb 2008 18:49:09 +0100 (CET) Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C58E50.5070007@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58E50.5070007@SoftDux.com> Message-ID: On Wed, 27 Feb 2008, Rudi Ahlers wrote: > Can I just download those file from > http://mirror.nsc.liu.se/CentOS/5.1/os/x86_64/CentOS/ and copy them to > the CentOS folder? Sure, but you'll need to "createrepo" in order to use them during install, or put a rpm -Uvh in your %post Regards Pablo From rbrown at BallistaSec.com Wed Feb 27 18:10:16 2008 From: rbrown at BallistaSec.com (Rodrick Brown) Date: Wed, 27 Feb 2008 13:10:16 -0500 Subject: Getting started In-Reply-To: <8552c3a30802261623l704f4ec5l12d5cc1503f4d412@mail.gmail.com> References: <8552c3a30802261623l704f4ec5l12d5cc1503f4d412@mail.gmail.com> Message-ID: Hello all when using cobbler everything went really at least getting it setup simple how ever after my host network books and begins the Linux installation I get an error that it cant load the default kickstart file http://myhost/clbr/kickstarts/default/ks.cfg I'm able to wget this file with no problem, however it seems my ks client is having issues I'm not sure what steps I could look at to trouble shoot this. devserver2 var # find . -name \ks.cfg ./lib/cobbler/kickstarts/ks.cfg ./www/cobbler/kickstarts/default/ks.cfg ./www/cobbler/kickstarts_sys/001E4F5126A8/ks.cfg -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Mike M Sent: Tuesday, February 26, 2008 7:23 PM To: Discussion list about Kickstart Subject: Re: Getting started On Tue, Feb 26, 2008 at 5:36 PM, Rodrick Brown wrote: > I'm very new to kick start and was wondering if there was a comprehensive > set of documents available that one could easily follow to build a custom > kick start profile and server. You should take a look at Cobbler: http://cobbler.et.redhat.com/ It makes getting kickstart up and running dead easy. Mike _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From mdehaan at redhat.com Wed Feb 27 18:27:44 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Wed, 27 Feb 2008 13:27:44 -0500 Subject: Getting started In-Reply-To: References: <8552c3a30802261623l704f4ec5l12d5cc1503f4d412@mail.gmail.com> Message-ID: <47C5ABA0.5080402@redhat.com> Rodrick Brown wrote: > Hello all when using cobbler everything went really at least getting it > setup simple how ever after my host network books and begins the Linux > installation I get an error that it cant load the default kickstart file > http://myhost/clbr/kickstarts/default/ks.cfg I'm able to wget this file > with no problem, however it seems my ks client is having issues I'm not > sure what steps I could look at to trouble shoot this. > > devserver2 var # find . -name \ks.cfg > ./lib/cobbler/kickstarts/ks.cfg > ./www/cobbler/kickstarts/default/ks.cfg > ./www/cobbler/kickstarts_sys/001E4F5126A8/ks.cfg > Join #cobbler on irc.freenode.net and someone can help debug your setup. As this isn't really the cobbler channel (more so the general use kickstart and kickstart syntax channel), that would be a better place. You can also try the cobbler mailing list. --Michael From Rudi at SoftDux.com Wed Feb 27 20:06:53 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Wed, 27 Feb 2008 22:06:53 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> Message-ID: <47C5C2DD.20506@SoftDux.com> Pablo Iranzo G?mez wrote: > On Wed, 27 Feb 2008, Rudi Ahlers wrote: > >> Well, I only started playing with kickstart yesterday And my brain hurts >> by now, but I'm learning a lot :) >> > > That's nice ;) > >> How does it work? Do I copy all the rpm's I want to use to the CentOS >> folder, and then run createrepo on that folder? >> > > Yes, but first, you need to test dependencies in order to have > "ALL REQUIRED" packages ;) > > After that, you can remove not needed ones. Usually one way to do > that is to try installing to a different folder, not your system so you > can keep rpm complaining of missing dependencies... and when everything is > fine, then you can "createrepo" and master a custom cd. > > Are you using CentOS 5? > > Regards > Pablo > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > This is a bit above me. I use SME 7.3 (which is built on CentOS 4.6) as inhouse production server, and then I have my (soon to colocate) test server, which currently runs CentOS 5.1 64. How would I test an rpm on an already running system? What I have done till now, is todo a base install, and then create only use the files installed (as per /installation.log) in my new CD. There's still quite a few programs that I don't need and could probably delete them from the CentOS folder on the CD, but I still need to figure out how these file are installed. I saw the repodata/comps.xml file, but it's LONG, and have a lot of different languages, not some something I can take on right now. It's almost easier to run something like "yum remove {package-name} from the kickstart file -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From Pablo.Iranzo at redhat.com Wed Feb 27 20:37:53 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Wed, 27 Feb 2008 21:37:53 +0100 (CET) Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C5C2DD.20506@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> Message-ID: On Wed, 27 Feb 2008, Rudi Ahlers wrote: > This is a bit above me. > > I use SME 7.3 (which is built on CentOS 4.6) as inhouse production > server, and then I have my (soon to colocate) test server, which > currently runs CentOS 5.1 64. > How would I test an rpm on an already running system? rpm --test and the proposed for testing dependencies: rpm --justdb --root /tmp/tempath/ package1, package2 > What I have done till now, is todo a base install, and then create only > use the files installed (as per /installation.log) in my new CD. There's > still quite a few programs that I don't need and could probably delete > them from the CentOS folder on the CD, but I still need to figure out > how these file are installed. I saw the repodata/comps.xml file, but > it's LONG, and have a lot of different languages, not some something I > can take on right now. It's almost easier to run something like "yum > remove {package-name} from the kickstart file You can put "-package" at %packages In order to have a minimal one just test with "%packages --nobase" and begin adding packages you find missing on the target system (ssh, joe, etc) Have a look at the --root trick to check all "really" required files ;) Regards Pablo From rbrown at BallistaSec.com Wed Feb 27 21:35:49 2008 From: rbrown at BallistaSec.com (Rodrick Brown) Date: Wed, 27 Feb 2008 16:35:49 -0500 Subject: adding drivers to initrd.img renders image unmountable Message-ID: I have a batch of systems that need and updated e1000 driver before the network device can been seen by the default RHEL5 images provided. I was able to get the files I needed and now I'm trying to add the initrd but when trying to do this the system pulls down the new image but is unable to read it correctly I get the following errors: Attempt to access beyond the end of device No file system could mount root, unable to mount rootfs on unknown block(0,1) This seems pretty trivial yet its giving me a major headache if anyone could help out I would appreciate it. Here are the steps I used to add the new modules to initrd, the modules I added overwrite the previous ones. ahci.ko, e1000.ko # mkdir /tmp/newinitrd # cd /tmp/newinitrd # gzip -d < /home/rbrown/RHEL5_64/images/pxeboot/initrd.img | cpio -i --make-directories # cd /tmp/newinitrd/modules # gzip -d < modules.cgz | cpio -i --make-directories # cd 2.6.18-8.el5/x86_64 # cp /home/rbrown/dell/*.ko . # find 2.6.18-8.el5 | cpio -o -H crc | gzip -9 > modules.cgz # rm -rf 2.6.18-8.el5 # cd /tmp/newinitrd # find . | cpio -o -H crc | gzip -9 > /tmp/initrd.img --- Rodrick R. Brown rbrown<@>ballista<.>com 120 Wall St. Suite 2400 C: 347 702 0012 | O: 646 307 4709 http://www.ballistasec.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From Rudi at SoftDux.com Thu Feb 28 19:08:25 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Thu, 28 Feb 2008 21:08:25 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> Message-ID: <47C706A9.5060307@SoftDux.com> Pablo Iranzo G?mez wrote: > On Wed, 27 Feb 2008, Rudi Ahlers wrote: > >> This is a bit above me. >> >> I use SME 7.3 (which is built on CentOS 4.6) as inhouse production >> server, and then I have my (soon to colocate) test server, which >> currently runs CentOS 5.1 64. >> How would I test an rpm on an already running system? >> > > rpm --test > > and the proposed for testing dependencies: rpm --justdb --root > /tmp/tempath/ package1, package2 > > >> What I have done till now, is todo a base install, and then create only >> use the files installed (as per /installation.log) in my new CD. There's >> still quite a few programs that I don't need and could probably delete >> them from the CentOS folder on the CD, but I still need to figure out >> how these file are installed. I saw the repodata/comps.xml file, but >> it's LONG, and have a lot of different languages, not some something I >> can take on right now. It's almost easier to run something like "yum >> remove {package-name} from the kickstart file >> > > You can put "-package" at %packages > > In order to have a minimal one just test with "%packages --nobase" > and begin adding packages you find missing on the target system (ssh, joe, > etc) > > Have a look at the --root trick to check all "really" required > files ;) > > Regards > Pablo > > _______________________________________________ > > Ok, this project didn't work as well as I thought. So, I installed an "empty" system, without any repositories as well. I'm going to take a new approach, and install xen, kernel-xen virt-manager & virsh via yum and then use that to build up my rpm list to work with. So, which repositories do I need on CentOS 5.1 x64 to install xen? -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From mdehaan at redhat.com Thu Feb 28 19:13:14 2008 From: mdehaan at redhat.com (Michael DeHaan) Date: Thu, 28 Feb 2008 14:13:14 -0500 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C706A9.5060307@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> <47C706A9.5060307@SoftDux.com> Message-ID: <47C707CA.1020203@redhat.com> Rudi Ahlers wrote: > Pablo Iranzo G?mez wrote: >> On Wed, 27 Feb 2008, Rudi Ahlers wrote: >> >>> This is a bit above me. >>> >>> I use SME 7.3 (which is built on CentOS 4.6) as inhouse production >>> server, and then I have my (soon to colocate) test server, which >>> currently runs CentOS 5.1 64. >>> How would I test an rpm on an already running system? >>> >> >> rpm --test >> >> and the proposed for testing dependencies: rpm --justdb --root >> /tmp/tempath/ package1, package2 >> >> >>> What I have done till now, is todo a base install, and then create only >>> use the files installed (as per /installation.log) in my new CD. >>> There's >>> still quite a few programs that I don't need and could probably delete >>> them from the CentOS folder on the CD, but I still need to figure out >>> how these file are installed. I saw the repodata/comps.xml file, but >>> it's LONG, and have a lot of different languages, not some something I >>> can take on right now. It's almost easier to run something like "yum >>> remove {package-name} from the kickstart file >>> >> >> You can put "-package" at %packages >> >> In order to have a minimal one just test with "%packages --nobase" >> and begin adding packages you find missing on the target system (ssh, >> joe, >> etc) >> >> Have a look at the --root trick to check all "really" required >> files ;) >> >> Regards >> Pablo >> >> _______________________________________________ >> >> > Ok, this project didn't work as well as I thought. So, I installed an > "empty" system, without any repositories as well. > I'm going to take a new approach, and install xen, kernel-xen > virt-manager & virsh via yum and then use that to build up my rpm list > to work with. > > So, which repositories do I need on CentOS 5.1 x64 to install xen? > All of that should be in the base repos for pretty much any relatively recent distro. Not sure what "without any repositories means". This isn't CentOS specific but is worth reading: http://fedoraproject.org/wiki/Docs/Fedora8VirtQuickStart --Michael From Rudi at SoftDux.com Thu Feb 28 19:17:55 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Thu, 28 Feb 2008 21:17:55 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C707CA.1020203@redhat.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> <47C706A9.5060307@SoftDux.com> <47C707CA.1020203@redhat.com> Message-ID: <47C708E3.8020505@SoftDux.com> Michael DeHaan wrote: > Rudi Ahlers wrote: >> Pablo Iranzo G?mez wrote: >>> On Wed, 27 Feb 2008, Rudi Ahlers wrote: >>> >>>> This is a bit above me. >>>> >>>> I use SME 7.3 (which is built on CentOS 4.6) as inhouse production >>>> server, and then I have my (soon to colocate) test server, which >>>> currently runs CentOS 5.1 64. >>>> How would I test an rpm on an already running system? >>>> >>> >>> rpm --test >>> >>> and the proposed for testing dependencies: rpm --justdb --root >>> /tmp/tempath/ package1, package2 >>> >>> >>>> What I have done till now, is todo a base install, and then create >>>> only >>>> use the files installed (as per /installation.log) in my new CD. >>>> There's >>>> still quite a few programs that I don't need and could probably delete >>>> them from the CentOS folder on the CD, but I still need to figure out >>>> how these file are installed. I saw the repodata/comps.xml file, but >>>> it's LONG, and have a lot of different languages, not some something I >>>> can take on right now. It's almost easier to run something like "yum >>>> remove {package-name} from the kickstart file >>>> >>> >>> You can put "-package" at %packages >>> >>> In order to have a minimal one just test with "%packages --nobase" >>> and begin adding packages you find missing on the target system >>> (ssh, joe, >>> etc) >>> >>> Have a look at the --root trick to check all "really" required >>> files ;) >>> >>> Regards >>> Pablo >>> >>> _______________________________________________ >>> >>> >> Ok, this project didn't work as well as I thought. So, I installed an >> "empty" system, without any repositories as well. >> I'm going to take a new approach, and install xen, kernel-xen >> virt-manager & virsh via yum and then use that to build up my rpm >> list to work with. >> >> So, which repositories do I need on CentOS 5.1 x64 to install xen? >> > > All of that should be in the base repos for pretty much any relatively > recent distro. > Not sure what "without any repositories means". > > This isn't CentOS specific but is worth reading: > http://fedoraproject.org/wiki/Docs/Fedora8VirtQuickStart > > --Michael > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > Thanx, I'll check out the post. For some reason /etc/yum.conf is empty: [root at gimbli ~]# more /etc/yum.conf [main] cachedir=/var/cache/yum keepcache=1 debuglevel=2 logfile=/var/log/yum.log pkgpolicy=newest distroverpkg=redhat-release tolerant=1 exactarch=1 obsoletes=1 gpgcheck=1 plugins=1 metadata_expire=1800 # PUT YOUR REPOS HERE OR IN separate files named file.repo # in /etc/yum.repos.d -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From Pablo.Iranzo at redhat.com Thu Feb 28 19:20:57 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Thu, 28 Feb 2008 20:20:57 +0100 (CET) Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C708E3.8020505@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> <47C706A9.5060307@SoftDux.com> <47C707CA.1020203@redhat.com> <47C708E3.8020505@SoftDux.com> Message-ID: Hi yum.conf is always empty, have a look at /etc/yum.repos.d/ There should be some files there defining repos Regards Pablo -- Pablo Iranzo G??mez (http://Alufis35.uv.es/~iranzo/) (PGPKey Available on http://www.uv.es/~iranzop/PGPKey.pgp) -- Postulado de Boling sobre la Ley de Murphy: Si se encuentra bien, no se preocupe. Se le pasar?? On Thu, 28 Feb 2008, Rudi Ahlers wrote: > Michael DeHaan wrote: > > Rudi Ahlers wrote: > >> Pablo Iranzo G?mez wrote: > >>> On Wed, 27 Feb 2008, Rudi Ahlers wrote: > >>> > >>>> This is a bit above me. > >>>> > >>>> I use SME 7.3 (which is built on CentOS 4.6) as inhouse production > >>>> server, and then I have my (soon to colocate) test server, which > >>>> currently runs CentOS 5.1 64. > >>>> How would I test an rpm on an already running system? > >>>> > >>> > >>> rpm --test > >>> > >>> and the proposed for testing dependencies: rpm --justdb --root > >>> /tmp/tempath/ package1, package2 > >>> > >>> > >>>> What I have done till now, is todo a base install, and then create > >>>> only > >>>> use the files installed (as per /installation.log) in my new CD. > >>>> There's > >>>> still quite a few programs that I don't need and could probably delete > >>>> them from the CentOS folder on the CD, but I still need to figure out > >>>> how these file are installed. I saw the repodata/comps.xml file, but > >>>> it's LONG, and have a lot of different languages, not some something I > >>>> can take on right now. It's almost easier to run something like "yum > >>>> remove {package-name} from the kickstart file > >>>> > >>> > >>> You can put "-package" at %packages > >>> > >>> In order to have a minimal one just test with "%packages --nobase" > >>> and begin adding packages you find missing on the target system > >>> (ssh, joe, > >>> etc) > >>> > >>> Have a look at the --root trick to check all "really" required > >>> files ;) > >>> > >>> Regards > >>> Pablo > >>> > >>> _______________________________________________ > >>> > >>> > >> Ok, this project didn't work as well as I thought. So, I installed an > >> "empty" system, without any repositories as well. > >> I'm going to take a new approach, and install xen, kernel-xen > >> virt-manager & virsh via yum and then use that to build up my rpm > >> list to work with. > >> > >> So, which repositories do I need on CentOS 5.1 x64 to install xen? > >> > > > > All of that should be in the base repos for pretty much any relatively > > recent distro. > > Not sure what "without any repositories means". > > > > This isn't CentOS specific but is worth reading: > > http://fedoraproject.org/wiki/Docs/Fedora8VirtQuickStart > > > > --Michael > > > > > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > Thanx, I'll check out the post. > > For some reason /etc/yum.conf is empty: > > > > [root at gimbli ~]# more /etc/yum.conf > [main] > cachedir=/var/cache/yum > keepcache=1 > debuglevel=2 > logfile=/var/log/yum.log > pkgpolicy=newest > distroverpkg=redhat-release > tolerant=1 > exactarch=1 > obsoletes=1 > gpgcheck=1 > plugins=1 > metadata_expire=1800 > > # PUT YOUR REPOS HERE OR IN separate files named file.repo > # in /etc/yum.repos.d > > > > -- > > Kind Regards > Rudi Ahlers > CEO, SoftDux > > Web: http://www.SoftDux.com > Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From Rudi at SoftDux.com Thu Feb 28 19:34:28 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Thu, 28 Feb 2008 21:34:28 +0200 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> <47C706A9.5060307@SoftDux.com> <47C707CA.1020203@redhat.com> <47C708E3.8020505@SoftDux.com> Message-ID: <47C70CC4.5010701@SoftDux.com> Pablo G?mez wrote: > Hi > yum.conf is always empty, have a look at /etc/yum.repos.d/ > > There should be some files there defining repos > > Regards > Pablo > > > > really? On SME 7.3 (which runs CentOS 4.6), the repositories are stored in /etc/yum.conf - I suppose that's how they do it. Anyway, I got it working. My LAN proxy was messing around, but it's downloading all the xen stuff now. Thanx :) -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From Pablo.Iranzo at redhat.com Thu Feb 28 19:41:05 2008 From: Pablo.Iranzo at redhat.com (=?UTF-8?Q?Pablo_Iranzo_G=C3=B3mez?=) Date: Thu, 28 Feb 2008 20:41:05 +0100 (CET) Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: <47C70CC4.5010701@SoftDux.com> References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> <47C706A9.5060307@SoftDux.com> <47C707CA.1020203@redhat.com> <47C708E3.8020505@SoftDux.com> <47C70CC4.5010701@SoftDux.com> Message-ID: On Thu, 28 Feb 2008, Rudi Ahlers wrote: > really? On SME 7.3 (which runs CentOS 4.6), the repositories are stored > in /etc/yum.conf - I suppose that's how they do it. My fault, forgot about you saying CentOS 4.6 :/ > Anyway, I got it working. My LAN proxy was messing around, but it's > downloading all the xen stuff now. Thanx :) Nice! Pablo From Rudi at SoftDux.com Thu Feb 28 20:26:06 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Thu, 28 Feb 2008 22:26:06 +0200 Subject: how do I detect the HDD's on the system? Message-ID: <47C718DE.5080206@SoftDux.com> Hi I want to setup different software RAID levels, according to the number & type of drives in the actual system. My problem is, I often have systems with mixed IDE & SATA drives. I found the following website, http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-preinstallconfig.html which has some neat tricks on detecting the IDE drives, but the same techniques doesn't apply for SATA drives ( I don't even know how SAS / SCSI drives operate on Linux, since I haven't worked with them yet) The basis of the script is as follows: for file in /proc/ide/h* do mymedia=`cat $file/media` if [ $mymedia == "disk" ] ; then hds="$hds `basename $file`" fi done set $hds numhd=`echo $#` drive1=`echo $hds | cut -d' ' -f1` drive2=`echo $hds | cut -d' ' -f2` It doesn't indicate which drive is first in the list, which could be a problem is the cdrom is on HDA for example (mine is, due to the 2U case layout and how the IDE cable runs) With SATA (which uses scsi emulation), the process is different. There's no /media folder in /proc/scsi/ - only a scsi file, which lists all the devices: cat /proc/scsi/scsi Attached devices: Host: scsi0 Channel: 00 Id: 00 Lun: 00 Vendor: ATA Model: ST3250410AS Rev: 3.AA Type: Direct-Access ANSI SCSI revision: 05 Host: scsi1 Channel: 00 Id: 00 Lun: 00 Vendor: ATA Model: ST3250410AS Rev: 3.AA Type: Direct-Access ANSI SCSI revision: 05 dmesg | grep sd SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) SCSI device sda: drive cache: write back SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) SCSI device sda: drive cache: write back sda: sda1 sda2 Attached scsi disk sda at scsi0, channel 0, id 0, lun 0 SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) SCSI device sdb: drive cache: write back SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) SCSI device sdb: drive cache: write back sdb: sdb1 sdb2 So, how does the system know that scsi0 = sata1 ? I'd like to find this out, cause it will make the RAID setup much easier for mixed systems -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From dan at half-asleep.com Thu Feb 28 20:34:35 2008 From: dan at half-asleep.com (Daniel Segall) Date: Thu, 28 Feb 2008 15:34:35 -0500 Subject: how do I detect the HDD's on the system? In-Reply-To: <47C718DE.5080206@SoftDux.com> References: <47C718DE.5080206@SoftDux.com> Message-ID: <4484c95febc084b4383b60d754ae4dcc@mail.half-asleep.com> The way I detect whether it's scsi/sas or ide is as follows. This assumes that if the disks are scsi or sas, then they have a hardware RAID (which has already been configured). If IDE create the software RAID. As far as number of drives/multiple formats, you'll have to do some additional logic based on what it finds. Hope it helps some. %pre if [ "`cat /proc/ide/hda/media`" = "disk" ] then echo "clearpart --all --initlabel" >/tmp/part-include echo "part raid.10 --size=100 --ondisk=hda --asprimary" >>/tmp/part-include echo "part raid.11 --size=100 --ondisk=hdb --asprimary" >>/tmp/part-include echo "part raid.20 --size 5120 --ondisk=hda --asprimary" >>/tmp/part-include echo "part raid.21 --size 5120 --ondisk=hdb --asprimary" >>/tmp/part-include echo "part raid.30 --size 5120 --ondisk=hda --asprimary" >>/tmp/part-include echo "part raid.31 --size 5120 --ondisk=hdb --asprimary" >>/tmp/part-include echo "part raid.40 --size 2048 --ondisk=hda" >>/tmp/part-include echo "part raid.41 --size 2048 --ondisk=hdb" >>/tmp/part-include echo "part raid.50 --size 10240 --ondisk=hda" >>/tmp/part-include echo "part raid.51 --size 10240 --ondisk=hdb" >>/tmp/part-include echo "raid / --fstype ext3 --device=md0 --level=RAID1 raid.20 raid.21" >>/tmp/part-include echo "raid /boot --fstype ext3 --device=md1 --level=RAID1 raid.10 raid.11" >>/tmp/part-include echo "raid /home --fstype ext3 --device=md2 --level=RAID1 raid.30 raid.31" >>/tmp/part-include echo "raid swap --fstype swap --device=md4 --level=RAID1 raid.40 raid.41" >>/tmp/part-include echo "raid /var --fstype ext3 --device=md3 --level=RAID1 raid.50 raid.51" >>/tmp/part-include else echo "clearpart --all --initlabel" >/tmp/part-include echo "part /boot --size 100 --fstype ext3 --asprimary" >>/tmp/part-include echo "part swap --size 2048 --fstype swap" >>/tmp/part-include echo "part / --size 5120 --fstype ext3 --asprimary" >>/tmp/part-include echo "part /home --size 5120 --fstype ext3 --asprimary" >>/tmp/part-include echo "part /var --size 10240 --fstype ext3" >>/tmp/part-include fi -Dan On Thu, 28 Feb 2008 22:26:06 +0200, Rudi Ahlers wrote: > Hi > > I want to setup different software RAID levels, according to the number > & type of drives in the actual system. My problem is, I often have > systems with mixed IDE & SATA drives. > > I found the following website, > http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-preinstallconfig.html > which has some neat tricks on detecting the IDE drives, but the same > techniques doesn't apply for SATA drives ( I don't even know how SAS / > SCSI drives operate on Linux, since I haven't worked with them yet) > > The basis of the script is as follows: > > > for file in /proc/ide/h* > do > mymedia=`cat $file/media` > if [ $mymedia == "disk" ] ; then > hds="$hds `basename $file`" > fi > done > > set $hds > numhd=`echo $#` > > drive1=`echo $hds | cut -d' ' -f1` > drive2=`echo $hds | cut -d' ' -f2` > > It doesn't indicate which drive is first in the list, which could be a > problem is the cdrom is on HDA for example (mine is, due to the 2U case > layout and how the IDE cable runs) > > With SATA (which uses scsi emulation), the process is different. There's > no /media folder in /proc/scsi/ - only a scsi file, which lists all the > devices: > > cat /proc/scsi/scsi > Attached devices: > Host: scsi0 Channel: 00 Id: 00 Lun: 00 > Vendor: ATA Model: ST3250410AS Rev: 3.AA > Type: Direct-Access ANSI SCSI revision: 05 > Host: scsi1 Channel: 00 Id: 00 Lun: 00 > Vendor: ATA Model: ST3250410AS Rev: 3.AA > Type: Direct-Access ANSI SCSI revision: 05 > > > dmesg | grep sd > SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) > SCSI device sda: drive cache: write back > SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) > SCSI device sda: drive cache: write back > sda: sda1 sda2 > Attached scsi disk sda at scsi0, channel 0, id 0, lun 0 > SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) > SCSI device sdb: drive cache: write back > SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) > SCSI device sdb: drive cache: write back > sdb: sdb1 sdb2 > > > > So, how does the system know that scsi0 = sata1 ? > > I'd like to find this out, cause it will make the RAID setup much easier > for mixed systems > > -- > > Kind Regards > Rudi Ahlers > CEO, SoftDux > > Web: http://www.SoftDux.com > Check out my technical blog, http://blog.softdux.com for Linux or other > technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting > stugg > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- From mail-lists at karan.org Thu Feb 28 20:38:09 2008 From: mail-lists at karan.org (Karanbir Singh) Date: Thu, 28 Feb 2008 20:38:09 +0000 Subject: include XEN kernel & tools in CentOS 5.1 kickstart? In-Reply-To: References: <47C5837A.4060604@SoftDux.com> <1204126855.11610.3.camel@iranzo.users.redhat.com> <47C58512.5010103@SoftDux.com> <1204127243.11610.6.camel@iranzo.users.redhat.com> <47C58DAB.7060904@SoftDux.com> <47C5C2DD.20506@SoftDux.com> <47C706A9.5060307@SoftDux.com> <47C707CA.1020203@redhat.com> <47C708E3.8020505@SoftDux.com> <47C70CC4.5010701@SoftDux.com> Message-ID: <47C71BB1.8000902@karan.org> Pablo Iranzo G?mez wrote: > On Thu, 28 Feb 2008, Rudi Ahlers wrote: >> really? On SME 7.3 (which runs CentOS 4.6), the repositories are stored >> in /etc/yum.conf - I suppose that's how they do it. > > My fault, forgot about you saying CentOS 4.6 :/ > CentOS-4 also puts all the repos in /etc/yum.repos.d/ and not in /etc/yum.conf - and it always has. CentOS-3 had the repo's setup in /etc/yum.conf -- Karanbir Singh : http://www.karan.org/ : 2522219 at icq From Chip.Shabazian at bankofamerica.com Thu Feb 28 21:27:23 2008 From: Chip.Shabazian at bankofamerica.com (Shabazian, Chip) Date: Thu, 28 Feb 2008 13:27:23 -0800 Subject: how do I detect the HDD's on the system? In-Reply-To: <47C718DE.5080206@SoftDux.com> Message-ID: Take a look at the presentation I sent out yesterday for another question for some tricks on detecting drives: http://www.shabazian.com/lw2007.pdf -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Rudi Ahlers Sent: Thursday, February 28, 2008 12:26 PM To: Discussion list about Kickstart Subject: how do I detect the HDD's on the system? Hi I want to setup different software RAID levels, according to the number & type of drives in the actual system. My problem is, I often have systems with mixed IDE & SATA drives. I found the following website, http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-gui de/s1-kickstart2-preinstallconfig.html which has some neat tricks on detecting the IDE drives, but the same techniques doesn't apply for SATA drives ( I don't even know how SAS / SCSI drives operate on Linux, since I haven't worked with them yet) The basis of the script is as follows: for file in /proc/ide/h* do mymedia=`cat $file/media` if [ $mymedia == "disk" ] ; then hds="$hds `basename $file`" fi done set $hds numhd=`echo $#` drive1=`echo $hds | cut -d' ' -f1` drive2=`echo $hds | cut -d' ' -f2` It doesn't indicate which drive is first in the list, which could be a problem is the cdrom is on HDA for example (mine is, due to the 2U case layout and how the IDE cable runs) With SATA (which uses scsi emulation), the process is different. There's no /media folder in /proc/scsi/ - only a scsi file, which lists all the devices: cat /proc/scsi/scsi Attached devices: Host: scsi0 Channel: 00 Id: 00 Lun: 00 Vendor: ATA Model: ST3250410AS Rev: 3.AA Type: Direct-Access ANSI SCSI revision: 05 Host: scsi1 Channel: 00 Id: 00 Lun: 00 Vendor: ATA Model: ST3250410AS Rev: 3.AA Type: Direct-Access ANSI SCSI revision: 05 dmesg | grep sd SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) SCSI device sda: drive cache: write back SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) SCSI device sda: drive cache: write back sda: sda1 sda2 Attached scsi disk sda at scsi0, channel 0, id 0, lun 0 SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) SCSI device sdb: drive cache: write back SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) SCSI device sdb: drive cache: write back sdb: sdb1 sdb2 So, how does the system know that scsi0 = sata1 ? I'd like to find this out, cause it will make the RAID setup much easier for mixed systems -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From Rudi at SoftDux.com Thu Feb 28 22:39:37 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Fri, 29 Feb 2008 00:39:37 +0200 Subject: how do I detect the HDD's on the system? In-Reply-To: <4484c95febc084b4383b60d754ae4dcc@mail.half-asleep.com> References: <47C718DE.5080206@SoftDux.com> <4484c95febc084b4383b60d754ae4dcc@mail.half-asleep.com> Message-ID: <47C73829.8040509@SoftDux.com> Daniel wrote: > The way I detect whether it's scsi/sas or ide is as follows. This assumes > that if the disks are scsi or sas, then they have a hardware RAID (which > has already been configured). If IDE create the software RAID. As far as > number of drives/multiple formats, you'll have to do some additional logic > based on what it finds. Hope it helps some. > > %pre > if [ "`cat /proc/ide/hda/media`" = "disk" ] > then > echo "clearpart --all --initlabel" >/tmp/part-include > echo "part raid.10 --size=100 --ondisk=hda --asprimary" >>/tmp/part-include > echo "part raid.11 --size=100 --ondisk=hdb --asprimary" >>/tmp/part-include > echo "part raid.20 --size 5120 --ondisk=hda --asprimary" > >>> /tmp/part-include >>> > echo "part raid.21 --size 5120 --ondisk=hdb --asprimary" > >>> /tmp/part-include >>> > echo "part raid.30 --size 5120 --ondisk=hda --asprimary" > >>> /tmp/part-include >>> > echo "part raid.31 --size 5120 --ondisk=hdb --asprimary" > >>> /tmp/part-include >>> > echo "part raid.40 --size 2048 --ondisk=hda" >>/tmp/part-include > echo "part raid.41 --size 2048 --ondisk=hdb" >>/tmp/part-include > echo "part raid.50 --size 10240 --ondisk=hda" >>/tmp/part-include > echo "part raid.51 --size 10240 --ondisk=hdb" >>/tmp/part-include > echo "raid / --fstype ext3 --device=md0 --level=RAID1 raid.20 raid.21" > >>> /tmp/part-include >>> > echo "raid /boot --fstype ext3 --device=md1 --level=RAID1 raid.10 raid.11" > >>> /tmp/part-include >>> > echo "raid /home --fstype ext3 --device=md2 --level=RAID1 raid.30 raid.31" > >>> /tmp/part-include >>> > echo "raid swap --fstype swap --device=md4 --level=RAID1 raid.40 raid.41" > >>> /tmp/part-include >>> > echo "raid /var --fstype ext3 --device=md3 --level=RAID1 raid.50 raid.51" > >>> /tmp/part-include >>> > else > echo "clearpart --all --initlabel" >/tmp/part-include > echo "part /boot --size 100 --fstype ext3 --asprimary" >>/tmp/part-include > echo "part swap --size 2048 --fstype swap" >>/tmp/part-include > echo "part / --size 5120 --fstype ext3 --asprimary" >>/tmp/part-include > echo "part /home --size 5120 --fstype ext3 --asprimary" >>/tmp/part-include > echo "part /var --size 10240 --fstype ext3" >>/tmp/part-include > fi > > -Dan > > > On Thu, 28 Feb 2008 22:26:06 +0200, Rudi Ahlers wrote: > >> Hi >> >> I want to setup different software RAID levels, according to the number >> & type of drives in the actual system. My problem is, I often have >> systems with mixed IDE & SATA drives. >> >> I found the following website, >> >> > http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-preinstallconfig.html > >> which has some neat tricks on detecting the IDE drives, but the same >> techniques doesn't apply for SATA drives ( I don't even know how SAS / >> SCSI drives operate on Linux, since I haven't worked with them yet) >> >> The basis of the script is as follows: >> >> >> for file in /proc/ide/h* >> do >> mymedia=`cat $file/media` >> if [ $mymedia == "disk" ] ; then >> hds="$hds `basename $file`" >> fi >> done >> >> set $hds >> numhd=`echo $#` >> >> drive1=`echo $hds | cut -d' ' -f1` >> drive2=`echo $hds | cut -d' ' -f2` >> >> It doesn't indicate which drive is first in the list, which could be a >> problem is the cdrom is on HDA for example (mine is, due to the 2U case >> layout and how the IDE cable runs) >> >> With SATA (which uses scsi emulation), the process is different. There's >> no /media folder in /proc/scsi/ - only a scsi file, which lists all the >> devices: >> >> cat /proc/scsi/scsi >> Attached devices: >> Host: scsi0 Channel: 00 Id: 00 Lun: 00 >> Vendor: ATA Model: ST3250410AS Rev: 3.AA >> Type: Direct-Access ANSI SCSI revision: 05 >> Host: scsi1 Channel: 00 Id: 00 Lun: 00 >> Vendor: ATA Model: ST3250410AS Rev: 3.AA >> Type: Direct-Access ANSI SCSI revision: 05 >> >> >> dmesg | grep sd >> SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) >> SCSI device sda: drive cache: write back >> SCSI device sda: 488395055 512-byte hdwr sectors (250058 MB) >> SCSI device sda: drive cache: write back >> sda: sda1 sda2 >> Attached scsi disk sda at scsi0, channel 0, id 0, lun 0 >> SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) >> SCSI device sdb: drive cache: write back >> SCSI device sdb: 488397168 512-byte hdwr sectors (250059 MB) >> SCSI device sdb: drive cache: write back >> sdb: sdb1 sdb2 >> >> >> >> So, how does the system know that scsi0 = sata1 ? >> >> I'd like to find this out, cause it will make the RAID setup much easier >> for mixed systems >> >> -- >> >> Kind Regards >> Rudi Ahlers >> CEO, SoftDux >> >> Web: http://www.SoftDux.com >> Check out my technical blog, http://blog.softdux.com for Linux or other >> technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting >> stugg >> >> _______________________________________________ >> Kickstart-list mailing list >> Kickstart-list at redhat.com >> https://www.redhat.com/mailman/listinfo/kickstart-list >> Hi Daniel Thank you for the info, but it doesn't really help me. You also check to see if the work "disk" appears in the /proc/ide/*/media files. My problem is, I want this to work for scenarios where I use software RAID. What makes it trickey, is that one some mobo's the SATA HDD's would be the first few drives, and then the IDE's, meaning that I may not necessarily have hda & hdb. I currently have hddc & hdd, so with this particular setup kickstart fails, since it wants to setup hda. My SATA HDD's work as sda, sdb, sdc, etc. As far as I can tell, that sequence will always be there, which could make it easier. -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From Pablo.Iranzo at redhat.com Fri Feb 29 07:28:17 2008 From: Pablo.Iranzo at redhat.com (Pablo Iranzo =?ISO-8859-1?Q?G=F3mez?=) Date: Fri, 29 Feb 2008 08:28:17 +0100 Subject: how do I detect the HDD's on the system? In-Reply-To: <47C73829.8040509@SoftDux.com> References: <47C718DE.5080206@SoftDux.com> <4484c95febc084b4383b60d754ae4dcc@mail.half-asleep.com> <47C73829.8040509@SoftDux.com> Message-ID: <1204270097.11610.43.camel@iranzo.users.redhat.com> Does this work fine for you? set $(list-harddrives) let numd=$#/2 # HDD NUM d1=$1 # 1st HDD Device S1=$2 # 1st HDD Size (and so on) Then you can start defining partitions based on the detection maade with a %pre script Regards Pablo El vie, 29-02-2008 a las 00:39 +0200, Rudi Ahlers escribi?: -- Pablo Iranzo G?mez (Pablo.Iranzo at redhat.com) RHCE/Global Profesional Services Consultant Spain Phone: +34 645 01 01 49 (CET/CEST) GnuPG KeyID: 0xFAD3CF0D --- Direcci?n Comercial: C/Jose Bardasano Baos, 9, Edif. Gorbea 3, planta 3?D, 28016 Madrid, Spain Direcci?n Registrada: Red Hat S.L., C/ Velazquez 63, Madrid 28001, Spain Inscrita en el Reg. Mercantil de Madrid ? C.I.F. B82657941 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Esta parte del mensaje est? firmada digitalmente URL: From Rudi at SoftDux.com Fri Feb 29 07:46:34 2008 From: Rudi at SoftDux.com (Rudi Ahlers) Date: Fri, 29 Feb 2008 09:46:34 +0200 Subject: how do I detect the HDD's on the system? In-Reply-To: <1204270097.11610.43.camel@iranzo.users.redhat.com> References: <47C718DE.5080206@SoftDux.com> <4484c95febc084b4383b60d754ae4dcc@mail.half-asleep.com> <47C73829.8040509@SoftDux.com> <1204270097.11610.43.camel@iranzo.users.redhat.com> Message-ID: <47C7B85A.4020007@SoftDux.com> Pablo Iranzo G?mez wrote: > Does this work fine for you? > > set $(list-harddrives) > let numd=$#/2 # HDD NUM > d1=$1 # 1st HDD Device > S1=$2 # 1st HDD Size (and so on) > > Then you can start defining partitions based on the detection maade > with a %pre script > > Regards > Pablo > > > > El vie, 29-02-2008 a las 00:39 +0200, Rudi Ahlers escribi??: > > ------------------------------------------------------------------------ > > _______________________________________________ > Where / how do I use that? -- Kind Regards Rudi Ahlers CEO, SoftDux Web: http://www.SoftDux.com Check out my technical blog, http://blog.softdux.com for Linux or other technical stuff, or visit http://www.WebHostingTalk.co.za for Web Hosting stugg From Pablo.Iranzo at redhat.com Fri Feb 29 07:51:06 2008 From: Pablo.Iranzo at redhat.com (Pablo Iranzo =?ISO-8859-1?Q?G=F3mez?=) Date: Fri, 29 Feb 2008 07:51:06 +0000 Subject: how do I detect the HDD's on the system? In-Reply-To: <47C7B85A.4020007@SoftDux.com> References: <47C718DE.5080206@SoftDux.com> <4484c95febc084b4383b60d754ae4dcc@mail.half-asleep.com> <47C73829.8040509@SoftDux.com> <1204270097.11610.43.camel@iranzo.users.redhat.com> <47C7B85A.4020007@SoftDux.com> Message-ID: <1204271466.11610.47.camel@iranzo.users.redhat.com> Hi! %include /tmp/part-include %pre set $(list-harddrives) let numd=$#/2 # N?mero de discos d1=$1 # 1st HDD drive S1=$2 # 1st HDD size DISCO=$d1 echo "clearpart --drives=$DISCO --all --initlabel" >> /tmp/part-include echo "part /boot --fstype ext3 --size=100 --ondisk=$DISCO" >> /tmp/part-include echo "part pv.100000 --size=1 --grow --ondisk=$DISCO" >> /tmp/part-include echo "volgroup VolGroup --pesize=32768 pv.100000" >> /tmp/part-include echo "logvol swap --fstype swap --name=Swap --vgname=VolGroup --size=2047" >> /tmp/part-include echo "logvol / --fstype ext3 --name=root --vgname=VolGroup --size=1024" >> /tmp/part-include echo "logvol /home --fstype ext3 --size=1024 --name=home --vgname=VolGroup" >> /tmp/part-include echo "logvol /tmp --fstype ext3 --size=1024 --name=tmp --vgname=VolGroup" >> /tmp/part-include echo "logvol /usr --fstype ext3 --size=3000 --name=usr --vgname=VolGroup" >> /tmp/part-include echo "logvol /opt --fstype ext3 --size=12000 --name=opt --vgname=VolGroup" >> /tmp/part-include echo "part /opt/iso --fstype vfat --size=4410 --ondisk=$DISCO" >> /tmp/part-include But with your intended partitioning layout and required "conditionals" to automatically setup raid as expressed in previous posts on this thread Regards Pablo El vie, 29-02-2008 a las 09:46 +0200, Rudi Ahlers escribi?: > Pablo Iranzo G?mez wrote: > > Does this work fine for you? > > > > set $(list-harddrives) > > let numd=$#/2 # HDD NUM > > d1=$1 # 1st HDD Device > > S1=$2 # 1st HDD Size (and so on) > > > > Then you can start defining partitions based on the detection maade > > with a %pre script > > > > Regards > > Pablo > > > > > > > > El vie, 29-02-2008 a las 00:39 +0200, Rudi Ahlers escribi??: > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > > Where / how do I use that? > -- Pablo Iranzo G?mez (Pablo.Iranzo at redhat.com) RHCE/Global Profesional Services Consultant Spain Phone: +34 645 01 01 49 (CET/CEST) GnuPG KeyID: 0xFAD3CF0D --- Direcci?n Comercial: C/Jose Bardasano Baos, 9, Edif. Gorbea 3, planta 3?D, 28016 Madrid, Spain Direcci?n Registrada: Red Hat S.L., C/ Velazquez 63, Madrid 28001, Spain Inscrita en el Reg. VolGroupntil de Madrid ? C.I.F. B82657941 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Esta parte del mensaje est? firmada digitalmente URL: From accarlson at gmail.com Fri Feb 29 12:27:05 2008 From: accarlson at gmail.com (Augusto Castelan Carlson) Date: Fri, 29 Feb 2008 09:27:05 -0300 Subject: install xen guest - getting kickstart fail Message-ID: Hello! I'm using Fedora 8 x86_64 re-spin. I'm trying to install a guest using kickstart, but it stay DHCP'ing. The host is the ftp server. The command I'm using is this one: virt-install -p -n guest -r 512 -vnc -l ftp://172.16.4.6/ -x ip=172.16.4.7 -x ks=ftp://172.16.4.6/guest.cfg -f /dev/vg/lv1 I hit ALT+F3 to see the messages: When it start one of the messages is "getting kickstart file" followed by "DHCPDISCOVER". As there is no DHCP server, the installation asks about network configuration, even specifying "ip=" in the kernel argument. I configured it manually (ip address, netmask and gateway) when asked. Once configured, it start "transferring kickstart file", "setting up kickstart file", "setting up language en_US.UTF-8" and appear a message saying that "need setup networking" and start DHCPDISCOVER again. DHCp'ing remains forever now. My kickstart starts with: install url --url ftp://172.16.4.6// lang en_US.UTF-8 keyboard br-abnt2 network --device eth0 --bootproto static --ip 172.16.4.7 --netmask 255.255.255.0 --gateway 172.16.4.1 --nameserver 172.16.4.1 firewall --disabled selinux --disabled .... Any idea of what is wrong? Once I asked here about disable DHCP'ing and I was advised to put kernel arguments, and it appeared to be working. Thanks in advance. Regards, -- Augusto