We have written a series of scripts for SuSE SLES 8that will setup the machine for autoyast installation and reboot the machine for installation. All works well on Intel and opteron platforms, but when we try to make a PPC machine install we run into a few issues. 1. It looks like the machine installs the OS, until the first reboot. At that time the machine refuses to boot up. If we put the installation CD in the cdrom drive, boot with that, we can select 'Boot installed system". 2. After selecting this, the machine boots up and finishes up the installation just like normal. Here is what I see when the machine does it's first reboot during install. /dev/sda invalid device Then I go into rescue CD mode from the installation CD and look at /etc/yaboot.conf which does not exist. So I chroot the mounted and look at lilo.conf which has the following in it. This does not look correct from what I have seen when the system is installed macnually. default=linux timeout=100 boot= activate image = /boot/vmlinux label = linux root = /dev/sda3 initrd = /boot/initrd append = "" If I try to run lilo at this point, it errors out on boot= which should be /dev/sda1. Activate is an extra parameter I have not seen before. Image pointed to /boot/vmlinux.initrd initrd = was not there append = "" was not there Any suggestions, questions, or pointers would be helpful. -- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
Hello Paul,
can you check if the kernel is installed? What happens if you want to boot
from disk? Has it already problems loading yaboot? Or has yaboot problems
finding the kernel?
And do you boot from network or CD? I have noticed that if you boot from
network and have the wrong sequence in the instorder file, it happens that
the new kernel of the sp3 was not installed.
Best regards, Joachim Jordan
Linux IT Architect, EMEA Linux Center of Competence
IBM Laboratories, Dept.3300, Schoenaicherstr.220, D 71032 Boeblingen,
Germany
Sametime/e-mail: jordanj@de.ibm.com, Phone: +49-7031-16-4407; Fax:
+49-7031-16-3232
Paul Edgar
Joachim, Yes the kernel is installed on the system. Once I 'boot the installed system', I can edit lilo.conf and get lilo to work. After this the system boots because yaboot.chrp and yaboot.conf are created. There are other issues that I have to address but cannot address those until the system can reboot on its own. Joachim Jordan wrote:
Hello Paul,
can you check if the kernel is installed? What happens if you want to boot from disk? Has it already problems loading yaboot? Or has yaboot problems finding the kernel?
And do you boot from network or CD? I have noticed that if you boot from network and have the wrong sequence in the instorder file, it happens that the new kernel of the sp3 was not installed.
Best regards, Joachim Jordan
Linux IT Architect, EMEA Linux Center of Competence IBM Laboratories, Dept.3300, Schoenaicherstr.220, D 71032 Boeblingen, Germany Sametime/e-mail: jordanj@de.ibm.com, Phone: +49-7031-16-4407; Fax: +49-7031-16-3232
Paul Edgar
To suse-autoinstall@suse.com 01.03.2004 20:07 cc Subject [suse-autoinstall] Autoyast on pSeries machines.
We have written a series of scripts for SuSE SLES 8that will setup the machine for autoyast installation and reboot the machine for installation.
All works well on Intel and opteron platforms, but when we try to make a PPC machine install we run into a few issues.
1. It looks like the machine installs the OS, until the first reboot. At that time the machine refuses to boot up. If we put the installation CD in the cdrom drive, boot with that, we can select 'Boot installed system".
2. After selecting this, the machine boots up and finishes up the installation just like normal.
Here is what I see when the machine does it's first reboot during install.
/dev/sda invalid device
Then I go into rescue CD mode from the installation CD and look at /etc/yaboot.conf which does not exist. So I chroot the mounted and look at lilo.conf which has the following in it.
This does not look correct from what I have seen when the system is installed macnually.
default=linux timeout=100 boot= activate
image = /boot/vmlinux label = linux root = /dev/sda3 initrd = /boot/initrd append = ""
If I try to run lilo at this point, it errors out on boot= which should be /dev/sda1. Activate is an extra parameter I have not seen before. Image pointed to /boot/vmlinux.initrd initrd = was not there append = "" was not there
Any suggestions, questions, or pointers would be helpful.
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com
No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
Hi Paul,
so what is the exact error message? I understand this means that in the
network install the yaboot installation fails?
Best regards, Joachim Jordan
Linux IT Architect, EMEA Linux Center of Competence
IBM Laboratories, Dept.3300, Schoenaicherstr.220, D 71032 Boeblingen,
Germany
Sametime/e-mail: jordanj@de.ibm.com, Phone: +49-7031-16-4407; Fax:
+49-7031-16-3232
Paul Edgar
Hello Paul,
can you check if the kernel is installed? What happens if you want to
from disk? Has it already problems loading yaboot? Or has yaboot problems finding the kernel?
And do you boot from network or CD? I have noticed that if you boot from network and have the wrong sequence in the instorder file, it happens
boot that
the new kernel of the sp3 was not installed.
Best regards, Joachim Jordan
Linux IT Architect, EMEA Linux Center of Competence IBM Laboratories, Dept.3300, Schoenaicherstr.220, D 71032 Boeblingen, Germany Sametime/e-mail: jordanj@de.ibm.com, Phone: +49-7031-16-4407; Fax: +49-7031-16-3232
Paul Edgar
com>
To
suse-autoinstall@suse.com
01.03.2004 20:07
cc
Subject
[suse-autoinstall] Autoyast on
pSeries machines.
We have written a series of scripts for SuSE SLES 8that will setup the machine for autoyast installation and reboot the machine for
installation.
All works well on Intel and opteron platforms, but when we try to make a PPC machine install we run into a few issues.
1. It looks like the machine installs the OS, until the first reboot. At that time the machine refuses to boot up. If we put the installation CD in the cdrom drive, boot with that, we can select 'Boot installed system".
2. After selecting this, the machine boots up and finishes up the installation just like normal.
Here is what I see when the machine does it's first reboot during
install.
/dev/sda invalid device
Then I go into rescue CD mode from the installation CD and look at /etc/yaboot.conf which does not exist. So I chroot the mounted and look at lilo.conf which has the following in it.
This does not look correct from what I have seen when the system is installed macnually.
default=linux timeout=100 boot= activate
image = /boot/vmlinux label = linux root = /dev/sda3 initrd = /boot/initrd append = ""
If I try to run lilo at this point, it errors out on boot= which should be /dev/sda1. Activate is an extra parameter I have not seen before. Image pointed to /boot/vmlinux.initrd initrd = was not there append = "" was not there
Any suggestions, questions, or pointers would be helpful.
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com
No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced. -- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
I have two identical five-disk SCSI systems, the first installed manually, the
second I am installing via autoyast with a reference profile (including
partitions) generated from the first machine.
The hard drives are partitioned identically, three partitions, first 512MB,
second 8GB, third ~59GB. I use the first as swap on all 5, the second partition
on the first two drives RAID1 under md0, with the third partition of all drives
as RAID5. This leaves the second partition on drives 3,4,5 unused.
When the autoinstall tries to partition the drives, it reverses the order of the
2nd and 3rd partitions on drives 3,4,5 (sdc,d,e{2,3}), giving the large third
partition to the "unused" partition ( mounted by yast2 as /data{1,2,3} ) and
confusing the RAID5 set.
The software raid table is below from system #1, as is the partitioning section
from the ref profile.
Any pointers would be appreciated.
Thanks,
Vernon
# autogenerated /etc/raidtab by YaST2
raiddev /dev/md0
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
persistent-superblock 1
chunk-size 4
device /dev/sda2
raid-disk 0
device /dev/sdb2
raid-disk 1
raiddev /dev/md1
raid-level 5
nr-raid-disks 5
nr-spare-disks 0
persistent-superblock 1
parity-algorithm left-symmetric
chunk-size 128
device /dev/sda3
raid-disk 0
device /dev/sdb3
raid-disk 1
device /dev/sdc3
raid-disk 2
device /dev/sdd3
raid-disk 3
device /dev/sde3
raid-disk 4
------------------------------------------------------
<partitioning xmlns="http://www.suse.com/1.0/yast2ns" config:type="list"
xmlns:config="http://www.suse.com/1.0/configns" >
<drive xmlns="http://www.suse.com/1.0/yast2ns">
<device xmlns="http://www.suse.com/1.0/yast2ns">/dev/md</device>
<partitions xmlns="http://www.suse.com/1.0/yast2ns" config:type="list"
xmlns:config="http://www.suse.com/1.0/configns" >
<partition xmlns="http://www.suse.com/1.0/yast2ns">
<filesystem xmlns="http://www.suse.com/1.0/yast2ns" config:type="symbol"
xmlns:config="http://www.suse.com/1.0/configns" >ext3</filesystem>
<mount xmlns="http://www.suse.com/1.0/yast2ns">/</mount>
Hi, Raid is not supported when you use this feature. How did you manage to get all those namespaces in the control file? What editor did you use? Anas H. Vernon Jones wrote:
I have two identical five-disk SCSI systems, the first installed manually, the second I am installing via autoyast with a reference profile (including partitions) generated from the first machine. The hard drives are partitioned identically, three partitions, first 512MB, second 8GB, third ~59GB. I use the first as swap on all 5, the second partition on the first two drives RAID1 under md0, with the third partition of all drives as RAID5. This leaves the second partition on drives 3,4,5 unused.
When the autoinstall tries to partition the drives, it reverses the order of the 2nd and 3rd partitions on drives 3,4,5 (sdc,d,e{2,3}), giving the large third partition to the "unused" partition ( mounted by yast2 as /data{1,2,3} ) and confusing the RAID5 set.
The software raid table is below from system #1, as is the partitioning section from the ref profile.
Any pointers would be appreciated.
Thanks,
Vernon
# autogenerated /etc/raidtab by YaST2
raiddev /dev/md0 raid-level 1 nr-raid-disks 2 nr-spare-disks 0 persistent-superblock 1 chunk-size 4 device /dev/sda2 raid-disk 0 device /dev/sdb2 raid-disk 1
raiddev /dev/md1 raid-level 5 nr-raid-disks 5 nr-spare-disks 0 persistent-superblock 1 parity-algorithm left-symmetric chunk-size 128 device /dev/sda3 raid-disk 0 device /dev/sdb3 raid-disk 1 device /dev/sdc3 raid-disk 2 device /dev/sdd3 raid-disk 3 device /dev/sde3 raid-disk 4
Thank you for the confirmation about raid and profiles. I used the installed /opt/kde3/bin/kxmleditor. I find copy/pasting large blocks from vi into a mailer awkward. I did find it a bit "busy", but I cannot find the namespace configuration preference in kxmleditor, so I had to write it off as free advertising. :) Vernon Anas Nashif wrote:
Hi, Raid is not supported when you use this feature. How did you manage to get all those namespaces in the control file? What editor did you use?
Anas
H. Vernon Jones wrote:
I have two identical five-disk SCSI systems, the first installed manually, the second I am installing via autoyast with a reference profile (including partitions) generated from the first machine. The hard drives are partitioned identically, three partitions, first 512MB, second 8GB, third ~59GB. I use the first as swap on all 5, the second partition on the first two drives RAID1 under md0, with the third partition of all drives as RAID5. This leaves the second partition on drives 3,4,5 unused.
When the autoinstall tries to partition the drives, it reverses the order of the 2nd and 3rd partitions on drives 3,4,5 (sdc,d,e{2,3}), giving the large third partition to the "unused" partition ( mounted by yast2 as /data{1,2,3} ) and confusing the RAID5 set.
The software raid table is below from system #1, as is the partitioning section from the ref profile.
Any pointers would be appreciated.
Thanks,
Vernon
# autogenerated /etc/raidtab by YaST2
raiddev /dev/md0 raid-level 1 nr-raid-disks 2 nr-spare-disks 0 persistent-superblock 1 chunk-size 4 device /dev/sda2 raid-disk 0 device /dev/sdb2 raid-disk 1
raiddev /dev/md1 raid-level 5 nr-raid-disks 5 nr-spare-disks 0 persistent-superblock 1 parity-algorithm left-symmetric chunk-size 128 device /dev/sda3 raid-disk 0 device /dev/sdb3 raid-disk 1 device /dev/sdc3 raid-disk 2 device /dev/sdd3 raid-disk 3 device /dev/sde3 raid-disk 4
Joachim, Let me see if I can explain this better. We have a set of scripts that builds and then incorporates autoinstall.xml into a platforms initrd. That will the installation kernel from the CD is moved to /boot, the machine is rebooted and an unattended installation goes forward. On the p machines there are a couple of changes that are made to setup the system similar to the way it is installed. We make sure that a vmlinuz.initrd is created. Now that that is said, once the p machine reboots the installation proceeds normally, until the time where the machine reboots. As the p machine tries to come back up, I get an error message that /dev/sda3 is corrupt. So to attempt to correct the situation, I insert the install CD and go thru the manual process and select rescue CD. Once booted I chroot the mounted /dev/sda3 after mount /dev/sda1(/boot) to /boot. In looking at the machine, I noticed that lilo.conf was not filled in correctly, having to add /dev/sda1 to the boot= line. Once added, I can then run lilo to have yaboot.conf, and /boot/yaboot.chrp created. No exiting the rescue cd and rebooting the machine, this installation proceeds sort of. At this time the network card (eth0) is no longer working and I have to skip the last few files(optional for this installation). However, the installation appears to proceed normally, asking for the appropriate passwords for root, userids, and network settings. However, once the machine is finished installing, I get a message that the ethernet card is not valid. I check modules.conf and ifcfg-eth0 and they appear normal. I am sure that if I can get the frist reboot part to work correctly, that the rest of the errors will magically go away since the machine will be able to fully install and configure itself without any issues. So the issue is the first reboot during the installation that I am trying to work through where it looks like lilo is not being executed correctly and the only thing I see is that the boot= parameter is not filled in correctly so lilo errors out. Joachim Jordan wrote:
Hi Paul,
so what is the exact error message? I understand this means that in the network install the yaboot installation fails?
Best regards, Joachim Jordan
Linux IT Architect, EMEA Linux Center of Competence IBM Laboratories, Dept.3300, Schoenaicherstr.220, D 71032 Boeblingen, Germany Sametime/e-mail: jordanj@de.ibm.com, Phone: +49-7031-16-4407; Fax: +49-7031-16-3232
Paul Edgar
To suse-autoinstall@suse.com 02.03.2004 15:55 cc Subject Re: [suse-autoinstall] Autoyast on pSeries machines.
Joachim,
Yes the kernel is installed on the system. Once I 'boot the installed system', I can edit lilo.conf and get lilo to work. After this the system boots because yaboot.chrp and yaboot.conf are created. There are other issues that I have to address but cannot address those until the system can reboot on its own.
Joachim Jordan wrote:
Hello Paul,
can you check if the kernel is installed? What happens if you want to
boot
from disk? Has it already problems loading yaboot? Or has yaboot problems finding the kernel?
And do you boot from network or CD? I have noticed that if you boot from network and have the wrong sequence in the instorder file, it happens
that
the new kernel of the sp3 was not installed.
Best regards, Joachim Jordan
Linux IT Architect, EMEA Linux Center of Competence IBM Laboratories, Dept.3300, Schoenaicherstr.220, D 71032 Boeblingen, Germany Sametime/e-mail: jordanj@de.ibm.com, Phone: +49-7031-16-4407; Fax: +49-7031-16-3232
Paul Edgar
com>
To
suse-autoinstall@suse.com
01.03.2004 20:07
cc
Subject
[suse-autoinstall] Autoyast on
pSeries machines.
We have written a series of scripts for SuSE SLES 8that will setup the machine for autoyast installation and reboot the machine for
installation.
All works well on Intel and opteron platforms, but when we try to make a PPC machine install we run into a few issues.
1. It looks like the machine installs the OS, until the first reboot. At that time the machine refuses to boot up. If we put the installation CD in the cdrom drive, boot with that, we can select 'Boot installed system".
2. After selecting this, the machine boots up and finishes up the installation just like normal.
Here is what I see when the machine does it's first reboot during
install.
/dev/sda invalid device
Then I go into rescue CD mode from the installation CD and look at /etc/yaboot.conf which does not exist. So I chroot the mounted and look at lilo.conf which has the following in it.
This does not look correct from what I have seen when the system is installed macnually.
default=linux timeout=100 boot= activate
image = /boot/vmlinux label = linux root = /dev/sda3 initrd = /boot/initrd append = ""
If I try to run lilo at this point, it errors out on boot= which should be /dev/sda1. Activate is an extra parameter I have not seen before. Image pointed to /boot/vmlinux.initrd initrd = was not there append = "" was not there
Any suggestions, questions, or pointers would be helpful.
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com
No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com
No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
-- Paul Edgar Verification Engineer Linux Technology Center / IBM External 512.838.1493 / TieLine 678-1493 javaoops@us.ibm.com No trees were destroyed in the sending of this message, however, a significant number of electrons were terribly inconvenienced.
participants (4)
-
Anas Nashif
-
H. Vernon Jones
-
Joachim Jordan
-
Paul Edgar