ZFS doesn't start after dist-upgrade

LXD
1

Hi,

I had a setup of LXC with bridges and ZFS storages running on 16.04. Today I did a dist-upgrade and after reboot I wanted to look if all LXC containers run and ran lxc list and lxc asked me to perform an initial setup. All network bridges are also gone and zfs list is not showing any device.

What’s the way to get the system running again?

Any help is appreciated!

Andy

2

It looks like, that the issue that lxc is not running is that ZFS is not able to start. Because I am using ZFS as file system for LXC.

syslog shows
zfs-import-cache.service: Main process exited, code=killed, status=6/ABRT
Failed to start Import ZFS pools by cache file.
zfs-import-cache.service: Unit entered failed state.
zfs-import-cache.service: Failed with result ‘signal’.

There are several services that are unable to start after the dist-upgrade and reboot. E.g. redis and spreed.

Because of this I posted this issue here https://ubuntuforums.org/showthread.php?t=2382240

If anybody can help, I would be very happy!

Andy

3

What do you see with those:

  • zpool status -v
  • journalctl -u zfs-import-cache
4

zpool status -v says
no pools available

and journalctl -u zfs-import-cache
-- Logs begin at Thu 2018-01-11 04:39:38 CET, end at Thu 2018-01-11 04:40:15 CET. -- Jan 11 04:39:41 server1 systemd[1]: Starting Import ZFS pools by cache file... Jan 11 04:39:42 server1 zpool[972]: internal error: Value too large for defined data type Jan 11 04:39:42 server1 systemd[1]: zfs-import-cache.service: Main process exited, code=killed, status=6/ABRT Jan 11 04:39:42 server1 systemd[1]: Failed to start Import ZFS pools by cache file. Jan 11 04:39:42 server1 systemd[1]: zfs-import-cache.service: Unit entered failed state. Jan 11 04:39:42 server1 systemd[1]: zfs-import-cache.service: Failed with result 'signal'.

I have 3 .img-files that should have been mounted as loop device on /dev/loop2 /dev/loop3 and /dev/loop4.

As it has been a while I setup the server I do not remember how I ran losetup to get them up on boot before I did the dist-upgrade. It looks like the original script to set them up has disappeared while the dist-upgrade. Now I tried to get the loop devices run with upstart with the following script, but it does not work

start on mounted MOUNTPOINT=/data
task

exec losetup /dev/loop2 /data/lxd/device01.img
exec losetup /dev/loop3 /data/lxd/device02.img
exec losetup /dev/loop4 /data/lxd/device03.img

The loopback devices are not setup. If I perform it manually then I can perform zpool import lxd.

5

I closed this thread as the problem is not directly related to LXC