Latest updates break LVM detection on boot
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
lvm2 (Ubuntu) |
New
|
Undecided
|
Unassigned |
Bug Description
Description: Ubuntu 12.04.1 LTS
Release: 12.04
Last night I installed the following updates using the "software updater" (/var/log/
Start-Date: 2012-11-03 02:23:04
Commandline: aptdaemon role='role-
Upgrade: base-files:i386 (6.5ubuntu6.2, 6.5ubuntu6.4), gnome-power-
End-Date: 2012-11-03 02:23:45
Since then, I am unable to boot my system because LVM volumes are not detected. Since the /home folder is on an LVM device, Ubuntu fails to mount it and prompts me to either skip (press S) or try manually (press M).
I am forced to press M, type "vgscan" followed by "mount -a" and then logout to resume the boot normally.
I have attached kern.log for the past few days. It seems to me that the only difference is that after the update my mdraid volumes acquire different device names:
Boot immediately before the update:
Nov 3 02:20:02 elmo kernel: [ 3.984469] md/raid1:md127: active with 2 out of 2 mirrors
Nov 3 02:20:02 elmo kernel: [ 3.984469] md127: detected capacity change from 0 to 8584626176
Nov 3 02:20:02 elmo kernel: [ 3.994084] md/raid1:md126: active with 2 out of 2 mirrors
Nov 3 02:20:02 elmo kernel: [ 3.994120] md126: detected capacity change from 0 to 8584626176
UPDATE: Start-Date: 2012-11-03 02:23:04 (see history.log above)
Boot immediately after the update:
Nov 3 07:16:49 elmo kernel: [ 3.871347] md/raid1:md0: active with 2 out of 2 mirrors
Nov 3 07:16:49 elmo kernel: [ 3.871371] md0: detected capacity change from 0 to 8584626176
Nov 3 07:16:49 elmo kernel: [ 6.837048] md/raid1:md127: active with 2 out of 2 mirrors
Nov 3 07:16:49 elmo kernel: [ 6.837068] md127: detected capacity change from 0 to 8584626176
There is a significant delay in detecting the second md volume. Also, a lot of things run now before the second volume is detected, so I suspect that the LVM scan takes place prematurely (both md raids are part of the same LVM volume).
Rebooted with previous kernel but problem persists.
Ran update-initramfs to recreate initrd and problem persists.
This must be related to the way new udev works?