Re: [Xen-users] Xen shutdown fails to release DRBD device

[Valentin Vidic <Valentin.Vidic@xxxxxxxxx>]

On Wed, Aug 22, 2018 at 03:34:20PM +0200, Roger Pau Monné wrote:
> On Wed, Aug 22, 2018 at 03:24:24PM +0200, Valentin Vidic wrote:
> > Hi,
> > 
> > Any idea why would Xen occasionally fail to release the block
> > device on shutdown?
> > 
> > Aug 21 22:56:00 node1 root: /etc/xen/scripts/block-drbd: remove 
> > XENBUS_PATH=backend/vbd/6/51712
> > Aug 21 22:56:00 node1 kernel: [ 2222.278235] block drbd6: State change 
> > failed: Device is held open by someone
> > Aug 21 22:56:00 node1 kernel: [ 2222.278304] block drbd6:   state = { 
> > cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate r----- }
> > Aug 21 22:56:00 node1 kernel: [ 2222.278340] block drbd6:  wanted = { 
> > cs:Connected ro:Secondary/Secondary ds:UpToDate/UpToDate r----- }
> > Aug 21 22:56:00 node1 root: /etc/xen/scripts/block-drbd: Writing 
> > backend/vbd/6/51712/hotplug-error /etc/xen/scripts/block-drbd failed; error 
> > detected. backend/vbd/6/51712/hotplug-status error to xenstore.
> > Aug 21 22:56:00 node1 root: /etc/xen/scripts/block-drbd: 
> > /etc/xen/scripts/block-drbd failed; error detected.
> > 
> > It seems like /etc/xen/scripts/block-drbd is called a bit too early just
> > sometimes. This is Xen 4.8.4 with DRBD 8.9.10 from Debian stretch.
> 
> The hotplug scripts are called when the backend releases the devices
> (ie: switches to state 6).
> 
> This seems to be an issue with the drbd state itself, and some kind of
> primary/secondary change? (according to the log above).
> 
> The drdb hotplug script is not maintained by the Xen project, so I
> would suggest you to contact the drbd project [0] in order to try to
> figure out what's wrong.

Right, block-drbd tries to move the device to Secondary state on
shutdown but this fails because something in Xen is still holding the
device open. Not sure if anything can be done here on the DRBD side
except retrying. This only happens when shutting down several VMs at
once, so it seems like some kind of timing problem in the backend
release?

-- 
Valentin

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-users