FAI Stable release 3.4.4 available

[Michael Tautschnig]

Hi Darsha,

Thanks a lot for your feedback!

[...]

> 
> [1] http://thread.gmane.org/gmane.linux.debian.fai/8024
> 
>  * still problems with RAID1 + grub-pc
> 
> Altough I actually /have/ found a working setup for myself I thought I
> should point out this problem again:
> 
> When creating a RAID1 and using the class GRUB_PC, grub only get's
> installed into "(hd0)" (see 10-setup).
> After creating such a setup, boot into the system. Wait until the
> disks are in sync (watch /proc/mdstat). Now shutdown and remove sda.
> You'll notice a non-booting system.
> 
> Altough Michael [Tautschnig] merged some of my patches into the (his?)
> experimental-branch[2], these changes never hit trunk. (Sorry,
> Michael: I really had no time to test your experimental package and in
> the end I forgot about it...)
> 

Would you mind doing that *now*? Well, probably it suffices if you just replace
10-setup with the attached version. The code in stable (and trunk) is known to
work only in very simple cases, but proper testing of our experimental versions
will help in getting the patches merged. BTW: No, this is not "my" branch, it's
open to everyone, it just so happens that until now most of the stuff has been
committed by myself. But it should be very easy to grab the experimental builds,
as described here:

http://wiki.fai-project.org/index.php/Main_Page#getting_FAI

> So thread [2] is still valid. You might consider the (second)
> 10-setup-patch.
> 
> [2] http://thread.gmane.org/gmane.linux.debian.fai/7572/focus=7575
> 

I think I even had merged your work, at least the main point of looping over the
devices is implemented in the attached version, it seems. It should just take
care of a few more corner cases.

[...]

> 
>  * reuse LVM-on-RAID disks (it may be an LVM-only problem but I did not encouter
> this problem without RAID) ...
> 
> Steps to reproduce:
> 
> Use a maschine with two hdds (sda/sdb).  Use a (simple) setup or use a disk
> layout like my RAIDLVM above and deploy the machine. And then do it
> again. (That should wipe away all data...!)
> 
> But: FAI fails with:
> 
> [snip]
> Executing: yes | mdadm --create  /dev/md0 --level=raid1 --force --run --raid-devices=2 /dev/sda1 /dev/sdb1
> Executing: mkfs.ext3  /dev/md0
> Executing: yes | mdadm --create  /dev/md1 --level=raid1 --force --run --raid-devices=2 /dev/sdb2 /dev/sda2
> Executing: pvcreate  /dev/md1
> Command had non-zero exit code
> Error in task partition. Traceback: task_error task_partition task task_install task task_action task main
> FATAL ERROR. Installation stopped.
> 
> FAI then dropped me to the shell. I then did the following to check-out the problem:
> 
> % pvcreate /dev/md1
>   Can't initialize physical volume "/dev/md1" of volume group "vgroup" without -ff
> 
> I manually did a "pvremove -ff /dev/md1" and re-ran FAI. That worked.
> 
> I also checked the fai.log and noticed that setup-storage seems to
> scan for volume groups after partitioning the disks. Maybe it should
> do that after creating the RAID-setup too...
> 

[...]

I think there's fair chance that this has been fixed meanwhile by some of the
patches included in trunk and/or experimental. I would be truly grateful if you
could give the experimental packages a try! In case it still fails, could you
please provide some logs (private mail is fine, of course).

Best regards,
Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
Url : http://lists.uni-koeln.de/pipermail/linux-fai/attachments/20101018/b9bc74d9/attachment.bin