[Xen-users] Xen/drbd - vbd devices locked after domain reboot

[Simon Cozens <simon@xxxxxxxxxxxxxxxx>]

We've had some problems recently with Xen domains not restarting after
they are shut down. I've narrowed it down to locking problems with
the drbd backend.

We're using Xen 3.2-1 and drbd 8.3.2 both from Debian. The symptoms
are, when this happens, that a domain won't start:

pox:~# xm create /etc/xen/shared/bird-flu
Using config file "/etc/xen/shared/bird-flu".
Error: Device 769 (vbd) could not be connected.
/etc/xen/scripts/block failed; error detected.

Attempting to manipulate the underlying drbd device with drbdadm
tells me that someone has the device held open, and drbd-overview
would say something like

25:bird-flu        Connected    Primary/Secondary   UpToDate/UpToDate C
    r----          xen-vbd: _bird-flu
 26:bird-flu-swap   Connected    Primary/Secondary   UpToDate/UpToDate C
    r----          xen-vbd: _bird-flu

Yet xm list confirms that the domain (a managed domain) is not running:

bird-flu                                        1024     1
   0.0

So something somewhere is holding a lock open and not shutting it
properly. I can't find out exactly what's holding it or how to manually
clear it, so I end up prodding wildly and eventually it rights itself.
Sometimes rebooting the Xen server doesn't even clear the lock.

Any clues on what this might be or how to fix it? Or should I just bite
the bullet and go to Remus?

Thanks,
Simon

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users