Hi Anil,

Can you please confirm whether the patch for avoiding the additional 
notifications on failover in the Openflowplugin He clustering is available in 
the beryllium branch

Regards,
K.Kalaiselvi

From: Anil Vishnoi [mailto:[email protected]]
Sent: Wednesday, January 27, 2016 2:37 PM
To: K, Kalaiselvi
Cc: [email protected]
Subject: Re: Additional node removal/updated notifications on ovs switch role 
transition to Master in Openflowplugin

Nope that should not happen, i am working on the patch that will fix this 
issue. Planning to push this patch sometime tomorrow, will update you once 
patch is ready for the test.

Thanks
Anil

On Wed, Jan 27, 2016 at 12:46 AM, 
<[email protected]<mailto:[email protected]>> wrote:
Hi,

With the Openflowplugin HE design clustering, when an ovs-switch is connected 
to a cluster of controllers
and one of the openflowplugin instances in the cluster transitions its role 
from Slave to Master as previous master is brought down,
additional node removed and node updated data change notifications are sent out 
even though the ovs-switch is still connected to one of the controllers in the 
cluster.

This causes the applications dependent on the openflowplugin to consider this 
transitioning ovs-switch as new node and not as already learnt node.
This behavior was not observed with the openflowplug Li clustering.
Is there any specific reasons why these additional notifications sent out with 
the He clustering design.

Regards,
K.Kalaiselvi




--
Thanks
Anil
_______________________________________________
openflowplugin-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev

Reply via email to