[Gluster-users] VMs paused - unknown storage error - Stale file handle - distribute 2 - replica 3 volume with sharding

[Marco Lorenzo Crociani]

Hi,
is there a way to recover file from "Stale file handle" errors?

Here some of the tests we have done:

- compared the extended attributes of all of the three replicas of the 
involved shard. Found identical attributes.

- compared SHA512 message digest of all of the three replicas of the 
involved shard. Found identical digests.

- tried to delete the shard from a replica set, one at a time, along 
with its hard link. Shard is always rebuilt correctly but error from 
client persists.

Regards,

-- 
Marco Crociani

Il 22/11/18 13:19, Marco Lorenzo Crociani ha scritto:
> Hi,
> I opened a bug on gluster because I have reading errors on files on a 
> gluster volume:
> https://bugzilla.redhat.com/show_bug.cgi?id=1652548
> 
> The files are many of the VMs images of the oVirt DATA storage domain. 
> oVirt pause the vms because unknown storage errors.
> It's impossibile to copy/clone, manage some snapshots of these vms. The 
> errors on the low level are "stale file handle".
> Volume is distribute 2 replicate 3 with sharding.
> 
> Should I open a bug also on oVirt?
> 
> Gluster 3.12.15-1.el7
> oVirt 4.2.6.4-1.el7
> 
> Regards,
>