[Gluster-users] Gluster 3.5 problems with libgfapi/qemu

[Vijay Bellur]

On 06/12/2014 07:47 PM, Ivano Talamo wrote:
> On 6/12/14 3:44 PM, Vijay Bellur wrote:
>> On 06/11/2014 11:25 PM, Ivano Talamo wrote:
>>> Hello,
>>> I recently update 2 servers (Scientific Linux 6) with a replicate volume
>>> from gluster 3.4 to 3.5.0-2.
>>> The volume was previously used to host qemu/kvm VM images accessed via a
>>> fuse-mounted mount-point.
>>> Now I would like to use the libgfapi but I'm seeing this error:
>>>
>>> [root at cmsrm-service02 ~]# qemu-img info
>>> gluster://cmsrm-service02/vol1/vms/disks/cmsrm-ui01.raw2
>>> [2014-06-11 17:47:22.084842] E [afr-common.c:3959:afr_notify]
>>> 0-vol1-replicate-0: All subvolumes are down. Going offline until atleast
>>> one of them comes back up.
>>> image: gluster://cmsrm-service03/vol1/vms/disks/cmsrm-ui01.raw2
>>> file format: raw
>>> virtual size: 20G (21474836480 bytes)
>>> disk size: 4.7G
>>> [2014-06-11 17:47:22.318034] E [afr-common.c:3959:afr_notify]
>>> 0-vol1-replicate-0: All subvolumes are down. Going offline until atleast
>>> one of them comes back up.
>>>
>>
>> This is a benign error message. qemu-img initializes a glusterfs graph
>> through libgfapi, performs the operation and then cleans up the graph.
>> afr translator in glusterfs displays this log message as part of a
>> graph cleanup operation. IIRC, qemu displays all log messages on
>> stderr by default and hence this message is seen.
>>
> I thought this indeed, since the volume was fine after that and I could
> access it.
>
>>> The error message does not appear if I access the file via the
>>> mount-point.
>>>
>>
>> There should be no functional problem even if this message is seen.
>>
>
> Btw, I have problems starting the VM. the "virsh start <vm-name>"
> command remains waiting forever on a futex call.
> The xml relevant section is:
>      <disk type='network' device='disk'>
>        <driver name='qemu' type='qcow2' cache='none'/>
>        <source protocol='gluster' name='vol1/vms/disks/cmsrm-ui01.qcow2'>
>          <host name='141.108.36.19' port='24007'/>
>        </source>
>        <target dev='vda' bus='virtio'/>
>        <address type='pci' domain='0x0000' bus='0x00' slot='0x05'
> function='0x0'/>
>      </disk>
>

Do you happen to notice any errors in the glusterd and glusterfsd logs?

-Vijay