Craig:
> The lockup is a little different now. Whereas previously the system would 
> hang before any of the snapshots were removed, now some of the snapshots are 
> successfully umount'd and lvremove'd before the hang. This could be the 
> result of changes made to the script that does the umount and lvremove, 
> reordering the sequence of the snapshots, which consist of home, root, and 
> var partitions.

That is a red herring. I had done the umounting/removing alternating
until this issue appeared to me. I first assumed that lvremove caused
the hangs but after dividing the script into a split phase operation
(unmount all first, lvremove all as second) the issue appeared always
during the unmount step.


But maybe i have found a workaround: Instead of running the script in context 
of bacula-fd, i use "at" to decouple the unmount/lvremove step from bacula-fd 
and delay it for about two minutes. I had freezes at maybe two out of three 
backup jobs before and no one during the last three backup jobs after 
implementing this change.

I guess the issue lies somewhere in the way kernel, bacula-fd, bash and
unmount interact together. Breaking up this chain by running the
unmount/lvremove part outside bacula-fd and after the backup job itself
has been finished circumvents the issue to appear.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1115753

Title:
  Umount of Multiple LVM Snapshots Causes  'soft lockup CPU#0 stuck for'

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1115753/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to