Replace master VCS peer in cluster (hardware fault)

[David Anstee]

We need to replace a faulty Cisco VCS in a cluster. The peer was the master peer before failing. 

I I understand from the cluster documentation that we will need to remove the VCs from the cluster and add it back in.

My question is will this incur any live service outage on the cluster itself? Do either the existing live VCs or replacement need to be rebooted etc?

the cluster is running X7.1

thanks

david

[Wayne DeNardi · Accepted solution]

Hi David,

No, the cluster will still be operational, just with one less VCS in its membership.

Obviously, any calls that are active on the VCS that you are removing will be affected, but the overall operation of the VCS cluster will not (the other however many others you have in the cluster devices will continute without any reboot required).

Check the VCS Cluster Creation and Maintenance guide for full instructions.

Wayne
--
Please remember to rate responses and to mark your question as answered if appropriate.

[Patrick Sparkman]

In addition to what Wayne stated above about the operation of the VCS cluster, once you remove the master from the cluster, any call licenses that particular VCS has installed on it, will remain shared within the cluster for up to two weeks.  After two weeks, whatever licenses that VCS had, will become unavailable to the cluster.  So be sure to time the replacement wisely and keep in mind the usage of the cluster during that time.  This is mentioned on pg 139 of the Cisco VCS Administrator Guide.

Also, since the master VCS was the one that went down, it would be best to use one of the VCS slaves to become the new master.  This is because it should contain all the configurations from the master that failed, and when the new VCS is added to the cluster, that data will replicate to the new VCS as a slave.  Other wise, if you tried to setup the replacement VCS as the master, you must make sure you have it configured exactly correct with whatever configuration the old master was running, because when you add it to the cluster it would override anything else the slaves are configured with, potentially breaking something.

[David Anstee]

Great, I thought that was the case, but just wanted to make sure.

Thank-you both for your comments.