[Users] unable to start vm in 3.3 and f19 with gluster

[Gianluca Cecchi]

On Thu, Oct 3, 2013 at 12:21 AM, Gianluca Cecchi wrote:
> On Wed, Oct 2, 2013 at 9:16 PM, Itamar Heim wrote:
>> On 10/02/2013 12:57 AM, Gianluca Cecchi wrote:
>>>
>>> Today I was able to work again on this matter and it seems related to
>>> spice
>>> Every time I start the VM (that is defined with spice) it goes in
>>
>>
>> and this doesn't happen if the VM is defined with vnc?
>
> No, reproduced both from oVirt and through virsh.
> with spice defined in boot options or in xml (for virsh) the vm
> remains in paused state and after a few minutes it seems the node
> hangs...
> with vnc the VM goes in runnign state
> I'm going to put same config on 2 physical nodes with only local
> storage and see what happens and report...
>
> Gianluca

So I was able to configure 2xHP blades BL685c G1 (Opteron G2 in oVirt)
with only one internal 72Gb disk each.

It seems here GlusterFS works quite well form the initial tests done.
I used same config as the problematic nested one.
Created a CentOS 6.4 VM, configured as server in oVirt but installed
as default Desktop in anaconda.
The 2 servers are on a Gigabit network
The install phase of the 1089 packages had an elapsed of about 17
minutes before completion.
During install I see (with command "bwm-ng -t 1000") throughput goes
at 55MB/s so it seems quite ok.
Also at the moment the distributed-replicated volume is on ovirtmgmt
itself and I don't see any disconnection from hosts in webadmin.....
I'm going to test more on this infra, and check/compare with the
nested environment that was configured quite the same...
unfortunately not much disk space to use and stress... ;-(

Gianluca