FAI 3.2.10 - Etch.

2008-10-09 Diskussionsfäden Fortier,Vincent [Montreal]
Hi All,

I'm getting this strange behaviour on some of our workstation
instalation (this is new since 3.2.9):
Using config file: /var/lib/fai/config/disk_config/DESKTOP
Executing: parted -s /dev/sda unit TiB print
Executing: parted -s /dev/sda unit B print free
Executing: parted -s /dev/sda unit chs print free
Executing: vgdisplay --units m -s
Command vgdisplay --units m -s had exit code 5
Executing: mdadm --detail --scan --verbose -c partitions
Executing: true
Executing: parted -s /dev/sda mklabel msdos
Executing: parted -s /dev/sda mklabel msdos
Executing: parted -s /dev/sda mkpart primary ext3 32256B 263208959B
Executing: udevsettle --timeout=10
Executing: parted -s /dev/sda mkpart primary linux-swap 263208960B
2410007039B
Executing: udevsettle --timeout=10
Executing: parted -s /dev/sda mkpart primary  2410007040B 160039272959B
Executing: udevsettle --timeout=10
Executing: parted -s /dev/sda set 1 boot on
Executing: mkfs.ext3  /dev/sda1
Executing: mkswap  /dev/sda2
Executing: parted -s /dev/sda set 3 lvm on
Executing: udevsettle --timeout=10
Executing: pvremove -ff -y /dev/sda3
Command pvremove -ff -y /dev/sda3 had exit code 5
Executing: pvcreate /dev/sda3
Command pvcreate /dev/sda3 had exit code 5
Executing: vgcreate vg00 /dev/sda3
Command vgcreate vg00 /dev/sda3 had exit code 5
Executing: vgchange -a y vg00
Command vgchange -a y vg00 had exit code 5
Executing: lvcreate -n var -L 8192 vg00
Command lvcreate -n var -L 8192 vg00 had exit code 5
Executing: udevsettle --timeout=10
Executing: mkfs.xfs  -f /dev/vg00/var
Command mkfs.xfs  -f /dev/vg00/var had exit code 1
Executing: lvcreate -n opt -L 5120 vg00
Command lvcreate -n opt -L 5120 vg00 had exit code 5
Executing: udevsettle --timeout=10
Executing: mkfs.xfs  -f /dev/vg00/opt
Command mkfs.xfs  -f /dev/vg00/opt had exit code 1
Executing: lvcreate -n root -L 1 vg00
Command lvcreate -n root -L 1 vg00 had exit code 5
Executing: udevsettle --timeout=10
Executing: mkfs.xfs  -f /dev/vg00/root
Command mkfs.xfs  -f /dev/vg00/root had exit code 1
Executing: lvcreate -n tmp -L 2048 vg00
Command lvcreate -n tmp -L 2048 vg00 had exit code 5
Executing: udevsettle --timeout=10
Executing: mkfs.xfs  -f /dev/vg00/tmp
Command mkfs.xfs  -f /dev/vg00/tmp had exit code 1
Executing: /lib/udev/vol_id -u /dev/sda1
Executing: /lib/udev/vol_id -l /dev/sda1
Command /lib/udev/vol_id -l /dev/sda1 had exit code 3
Executing: readlink -f /dev/vg00/var
Command readlink -f /dev/vg00/var had exit code 1
Use of uninitialized value in scalar chomp at
/usr/share/fai/setup-storage//Fstab.pm line 207.
Use of uninitialized value in -b at
/usr/share/fai/setup-storage//Fstab.pm line 210.
Failed to resolve /dev/vg00/var

While investigating I've found this:
# pvremove -ff -y /dev/sda3
/etc/lvm/.cache: open failed: Permission denied
/var/lock/lvm: mkdir failed: Permission denied
Locking type 1 initialisation failed.

# echo $?
5

This seems to be only happening on a specific workstation that I had
previously installed using 3.2.9 version of FAI (although I might had
been using static disks instead of LVM at the time).  This config using
3.2.10 runs smoothly everywhere else... So I just don't get it.

Help appreciated!

NOTE: Could similar bug be fixed in 3.2.11 ? Or 3.2.12 SVN ?

Vincent Fortier
Informatique
Environnement Canada
Place Bonaventure
Bureau: 514.283.1143
Cellulaire: 514.919.2014





Regular kernel?

2008-10-09 Diskussionsfäden Fortier,Vincent [Montreal]

I've tried using regular etchnhalf kernels to boot PXE instead of the
usual 2.6.18 + unionfs kernel that FAI uses.  Using any kernel the
nfsroot simply does not mount and I presume it's due to unionsfs module
which is not available or loaded (if available by default with
2.6.24-etchnhalf).

Any clues on how to boot with regular or even backport kernels ?

Thnx!

Vincent Fortier
Informatique






RE: Regular kernel?

2008-10-09 Diskussionsfäden Thomas Lange
 On Thu, 9 Oct 2008 14:27:18 -0400, Fortier,Vincent [Montreal] [EMAIL 
 PROTECTED] said:

 Although there is a:
 /lib/modules/2.6.24-etchnhalf.1-amd64/kernel/fs/ufs/ufs.ko

 Could it just be a matter of forcing the load of the module at boot time?
ufs is not aufs.

-- 
regards Thomas