Hi,
We found the problem, and the solution as well.
The problem was that there was inconsistency in the gluster storage volume
which was underneath the libvirt storage pool. One file in multiple
locations with different sizes. After this fixing those inconsistencies
pool-start worked correctly.
But the error message from virsh was way off.
Best regards
Raggi Skula
On Tue, Sep 4, 2012 at 9:45 AM, Ragnar Skúlason <raggi(a)greenqloud.com>wrote:
Hi everyone,
I'm not sure if this belongs here, I sent a question to the users list
yesterday without getting a response, so I'm trying here.
I have storage-pool that for some reason became inactive (possibly due to
an error with pool-refresh in an old version of a Java API binding).
I want to restart the pool but am unable to, due to the volume being
mounted:
# virsh pool-start [uuid]
error: Failed to start pool [uuid]
error: Requested operation is not valid: Target '/mnt/[uuid]' is already
mounted
We have tested this with versions 0.9.12 and 0.9.13.
This did work in versions 0.8.x
Is it possible to bypass this mount check and start the storage pool
(forcefully)?
Many thanks,
Ragnar Skúlason