[one-users] migrate causes VM to reboot

[Shi Jin]

Actually, after some digging around, it turned out to be the libvirt
problem running on my RHEl-5.4 host.
No memory data was written to the checkpoint file (only 1.2KB in size)
 and thus resulting in a reboot at restore.
I simply did an yum update and rebooted the server, they now seem to work fine.
Thanks.

Shi

On Thu, Apr 29, 2010 at 7:00 AM, Ruben S. Montero <rubensm at dacya.ucm.es> wrote:
> Hi
>
> Actually OpenNebula do a restore from the checkpoint file when the VM
> is "cold" migrated. It should not be restarted from the plain
> deployment file. This looks like a bug.
>
> Please Shi could you send us the vm.log file for this case (i.e. doing
> a migrate between two hosts). Thanks in advance!
>
> Cheers
>
> Ruben
>
> On Tue, Apr 27, 2010 at 3:49 PM, Shi Jin <jinzishuai at gmail.com> wrote:
>> Neil.
>> Thanks for the reply.
>> Yes, livemigration does not reboot.
>> However, it seems quite easy to do a cold migrate without the need to
>> reboot, why not do it?
>> In the current way, the VM will be shutdown improperly. For the case
>> of Windows, it might cause lengthy disk check upon boot.
>> In the worst case, it might generate disk failure (I haven't had it
>> yet. Maybe the checkpoint flushes the cache thus is better than a
>> power lose).
>>
>> Just my thoughts.
>> Shi
>>
>> On Tue, Apr 27, 2010 at 1:22 AM, Neil Mooney <neil.mooney at sara.nl> wrote:
>>>
>>> Hi Shi
>>>
>>> For the machine to transfer without reboot, you probably need to use
>>> livemigrate instead.
>>> Do you have shared storage already ?
>>>
>>> The checkpoint file is probably redundant in a non live migration.(not
>>> sure).
>>>
>>> cheers
>>>
>>> Neil
>>>
>>> --
>>> * livemigrate (Migrates a running VM to another host without downtime)
>>>    onevm livemigrate <vm_id> <host_id>
>>>
>>> Shi Jin wrote:
>>>> Hi there,
>>>>
>>>> I am able to call "onevm migrate" to move a VM from one host to another.
>>>> However, it seems that after each migration, the VM was started from
>>>> the XML deployment file again.
>>>>
>>>> What is interesting is that a checkpoint has already been made.
>>>> However, it does not seem to be used?
>>>>
>>>> I think this is a bug, isn't it?
>>>> Thanks a lot.
>>>>
>>>>
>>>
>>>
>>
>>
>>
>> --
>> Shi Jin, Ph.D.
>> _______________________________________________
>> Users mailing list
>> Users at lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>
>
>
>
> --
> Dr. Ruben Santiago Montero
> Associate Professor (Profesor Titular), Complutense University of Madrid
>
> URL: http://dsa-research.org/doku.php?id=people:ruben
> Weblog: http://blog.dsa-research.org/?author=7
>



-- 
Shi Jin, Ph.D.