[Users] 3.1 and centos 6.3 problem

[Fernando Ariel Martinez]

I used pre-allocate just for testing. I'll use thin provision.

To work around the problem (allocate space for the virtual disk), I
installed ovirt according to "Integrating Gluster storage into oVirt" -
by Robert Middleswarth.
I used a data center with storage type "POSIX compliant FS", but then I
found another problem
(https://bugzilla.redhat.com/show_bug.cgi?id=835949) which is mentioned
by Middleswarth. I used the patch http://gerrit.ovirt.org/#/c/6103/ to
fix it.

Up to now, it's working.


Fernando M.

El 20/02/2013 05:54 a.m., Itamar Heim escribió:
> On 19/02/2013 18:43, Fernando Ariel Martinez wrote:
>> I could see that the volume (vol1) was mounted with nfs type.
>> I dismounted the volume and mounted it again in both nodes but using
>> glusterfs type.
>>
>> This time, I could create a 10GB preallocated virtual disk.
>> The events tab showed:
>> 2013-Feb-19, 09:12:39 Add-Disk operation of 'vm1_Disk1' was initiated on
>> VM vm1 by admin at internal.
>> 2013-Feb-19, 09:30:35 The disk 'vm1_Disk1' was successfully added to
>> VM vm1.
>>
>> Too slow, isn't? but worked
>
> why pre-allocate on file based storage? iirc, thin should still be
> "raw" on file based domains?
>
>>
>>
>> El 19/02/2013 08:21 a.m., Fernando Ariel Martinez escribió:
>>> I have installed ovirt 3.1 according to the instructions in
>>> "Integrating
>>> Gluster storage into oVirt" - by Robert Middleswarth, linked in
>>> http://wiki.ovirt.org/Documentation.
>>> I was able to complete the instructions successfully but I have
>>> problems
>>> when a need to create a VM. Especifically when I need to allocate space
>>> for the virtual disk.
>>>
>>> I have:
>>> 1 ovirt engine server, (admin)
>>> 2 nodes, IBM System x 3650M3 (nodo1, nodo2)
>>> 1 switch cisco SG 300-20 (to connect them)
>>>
>>> I created a VM (vm1).
>>> When I add a 10GB preallocated disk to vm1, after a few seconds the
>>> ovirt engine server loses the connection with the node SPM.
>>> I can see in the events tab "Invalid status on Data Center Default.
>>> Setting status to Non-Responsive."
>>> Then, the SPM changes to the other node. At last, in the events tab I
>>> can see "Operation Add-Disk failed to complete on VM vm1."
>>>
>>> I have set SELinux in permissive mode. I have also disabled iptables.
>>> In the ovirt engine server there are an ISO domain, called iso, (NFS
>>> Export Path: admin.cluster:/nfs/iso) and an Export domain, called exp,
>>> (NFS Export Path: admin.cluster:/nfs/export).
>>> In each node with gluster, the /data directory is used to create the
>>> volume vol1 (type: Replicate)
>>> The Data domain is called vms (NFS Export Path: localhost:/vol1).
>>>
>>> Some help wil be appreciated
>>>
>>> Fernando M.
>>
>> _______________________________________________
>> Users mailing list
>> Users at ovirt.org
>> http://lists.ovirt.org/mailman/listinfo/users
>>
>
>