[DRBD-user] LVM crash maybe due to a drbd issue

Maxence DUNNEWIND maxence at dunnewind.net
Mon Feb 1 21:01:05 CET 2010

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


> You tell us how to reproduce in the current drbd version,
> and we see what we can do to get it fixed.

We have a setup like : 

2 nodes.
a vg on each node.
many drbd devices (3 or 4 on these nodes) using 2 lvs (1 for data, 1 for meta).

so, we have, on each node : hard drive -> pv -> vg -> 2 lv per vm -> 1 drbd device.
On this drbd device, we have a vm running. This setup is pretty complex, but
works well until drbd fails in Timeout.

After some time, because each node is on a different physical location, 1
device fails and go to Timeout. (it's a 500G device, with pretty few i/o). As
soon as it is on Timeout state, all lvm operations (a vgs for example) will go
in endless loop, and the drbdX_receiver process will be in uninterruptible
sleep. It seems vgs can't read the lv informations .

The lvm operations will use lot of cpu, and finally node will crash. The only
way to kill these processes is to physically reboot thez node.

Cheers,

Maxence
-- 
Maxence DUNNEWIND
Contact : maxence at dunnewind.net
Site : http://www.dunnewind.net
06 32 39 39 93
GPG : 18AE 61E4 D0B0 1C7C AAC9  E40D 4D39 68DB 0D2E B533
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20100201/0a50e13c/attachment.pgp>


More information about the drbd-user mailing list