live migration problems.

live migration problems.

[Riccardo Veraldi]

Hello,
I have problems with live migration.
I have several VMs in a CentOS 6.2 cluster environment.

When I migrate one virtual machine from nodeA to nodeB the migration 
goes smoothly but
in the same time the VM is migrated an external ping cannot reach the VM.
I ping the VM before migration and after the migration is done ping 
stops to work.
This happens for some VM and does not happen for other VM.
Apparently the VM are configured the same way so I can't understand why 
this is happening.
If I migrate the VM back to the original physical node the ping start 
working again like if the switch did not
realize the VM mac address changed switch port.
To make things work always I Should make a ping from the VM console to 
whatever host and
in this case the arp table is updated on the switch.
The strange thing as I said is that some VM seems to always work after 
migration while other
seems never to work untl they are migrated back to original cluster node.
Any hints or suggestions ?

thank you very much

Rick

Re: live migration problems.

[Andrew Cathrow]

----- Original Message -----
> From: "Riccardo Veraldi" <Riccardo.Veraldi@cnaf.infn.it>
> To: kvm@vger.kernel.org
> Sent: Friday, March 30, 2012 5:45:47 PM
> Subject: live migration problems.
> 
> Hello,
> I have problems with live migration.
> I have several VMs in a CentOS 6.2 cluster environment.
> 
> When I migrate one virtual machine from nodeA to nodeB the migration
> goes smoothly but
> in the same time the VM is migrated an external ping cannot reach the
> VM.
> I ping the VM before migration and after the migration is done ping
> stops to work.
> This happens for some VM and does not happen for other VM.
> Apparently the VM are configured the same way so I can't understand
> why
> this is happening.
> If I migrate the VM back to the original physical node the ping start
> working again like if the switch did not
> realize the VM mac address changed switch port.
> To make things work always I Should make a ping from the VM console
> to
> whatever host and
> in this case the arp table is updated on the switch.
> The strange thing as I said is that some VM seems to always work
> after
> migration while other
> seems never to work untl they are migrated back to original cluster
> node.
> Any hints or suggestions ?

Make sure you have DELAY=0 set in the ifcfg file for the bridge on both hosts.


> 
> thank you very much
> 
> Rick
> 
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: live migration problems.

[Riccardo Veraldi]

ifdown and ifup of the bridge is sufficient to enable the new delay config ?

thank you

Rick


On 3/30/12 11:58 PM, Andrew Cathrow wrote:
>
> ----- Original Message -----
>> From: "Riccardo Veraldi"<Riccardo.Veraldi@cnaf.infn.it>
>> To: kvm@vger.kernel.org
>> Sent: Friday, March 30, 2012 5:45:47 PM
>> Subject: live migration problems.
>>
>> Hello,
>> I have problems with live migration.
>> I have several VMs in a CentOS 6.2 cluster environment.
>>
>> When I migrate one virtual machine from nodeA to nodeB the migration
>> goes smoothly but
>> in the same time the VM is migrated an external ping cannot reach the
>> VM.
>> I ping the VM before migration and after the migration is done ping
>> stops to work.
>> This happens for some VM and does not happen for other VM.
>> Apparently the VM are configured the same way so I can't understand
>> why
>> this is happening.
>> If I migrate the VM back to the original physical node the ping start
>> working again like if the switch did not
>> realize the VM mac address changed switch port.
>> To make things work always I Should make a ping from the VM console
>> to
>> whatever host and
>> in this case the arp table is updated on the switch.
>> The strange thing as I said is that some VM seems to always work
>> after
>> migration while other
>> seems never to work untl they are migrated back to original cluster
>> node.
>> Any hints or suggestions ?
> Make sure you have DELAY=0 set in the ifcfg file for the bridge on both hosts.
>
>
>> thank you very much
>>
>> Rick
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe kvm" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: live migration problems.

[Riccardo Veraldi]

Thank you very much, this solved my problem, not even 1 ping is lost!
I did not find this hint in the documentation though.
Really thanks again

Riccardo


On 3/30/12 11:58 PM, Andrew Cathrow wrote:
>
> ----- Original Message -----
>> From: "Riccardo Veraldi"<Riccardo.Veraldi@cnaf.infn.it>
>> To: kvm@vger.kernel.org
>> Sent: Friday, March 30, 2012 5:45:47 PM
>> Subject: live migration problems.
>>
>> Hello,
>> I have problems with live migration.
>> I have several VMs in a CentOS 6.2 cluster environment.
>>
>> When I migrate one virtual machine from nodeA to nodeB the migration
>> goes smoothly but
>> in the same time the VM is migrated an external ping cannot reach the
>> VM.
>> I ping the VM before migration and after the migration is done ping
>> stops to work.
>> This happens for some VM and does not happen for other VM.
>> Apparently the VM are configured the same way so I can't understand
>> why
>> this is happening.
>> If I migrate the VM back to the original physical node the ping start
>> working again like if the switch did not
>> realize the VM mac address changed switch port.
>> To make things work always I Should make a ping from the VM console
>> to
>> whatever host and
>> in this case the arp table is updated on the switch.
>> The strange thing as I said is that some VM seems to always work
>> after
>> migration while other
>> seems never to work untl they are migrated back to original cluster
>> node.
>> Any hints or suggestions ?
> Make sure you have DELAY=0 set in the ifcfg file for the bridge on both hosts.
>
>
>> thank you very much
>>
>> Rick
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe kvm" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html