[Pacemaker] Virtual IP migration fails to cleanup IP from failed interface

[Arjun Pandey]

No there's a separate interface for cluster communication. I don't
have fencing configured in the setup.


On Thu, Aug 21, 2014 at 5:47 AM, Andrew Beekhof <andrew at beekhof.net> wrote:
>
> On 20 Aug 2014, at 11:45 pm, Arjun Pandey <apandepublic at gmail.com> wrote:
>
>> Hi
>>
>> I have a 2 Node(Active-Passive) mode cluster on CentOS 6.4
>> After setting up the cluster things work fine.
>>
>> There are 4 resources (2 virtual IP's and my application (master-slave ) mode).
>> Additionally i have a constraint that virtual IP's are colocated with
>> the master.
>>
>> If i bring the interface down using ifconfig eth1 down , migration of
>> virtual IP happens and the slave instance gets promoted.
>
> Is this the same interface that the cluster communicates on?
> Have you got fencing configured?
>
>> However the
>> address with the previous is not cleaned up.   If the interface is
>> again brought up, i now see the virtual IP associated with interfaces
>> on both the nodes. Should IP be flushed from interface  via migration
>> ?
>>
>> Regards
>> Arjun
>>
>> _______________________________________________
>> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>> Project Home: http://www.clusterlabs.org
>> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>> Bugs: http://bugs.clusterlabs.org
>
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>