Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

Rebooting UCCX HA cluster

I have a question regarding rebooting servers in the HA cluster.

From what I have read, you should always reboot the Pub server first, then once Pub is completely up, reboot the Sub.

Does it matter if the Pub is not the Master?

If the Sub is the Master and Pub is the Slave, does the reboot sequence change?

Secondly, what about rebooting just the Sub (aside from the above question), is there any problem with just rebooting the Sub in the cluster and not rebooting Pub?

For example, we usually have our HA cluster set with the Sub being Master and the Pub is the Slave. We had a problem with CAD agents not able to log in, so the CCX engine was restarted on the Master, pushing the Master to the other server (Pub) agent were then able to log in. The Sub was then rebooted and the CCX engine was restarted on the Master again, pushing the Master back to the Sub.

This worked ok, but I don't know if this is an acceptable practice to reboot the Sub only.

2 REPLIES
Silver

Re: Rebooting UCCX HA cluster

If you want your Publisher to be master, you should always reboot that first, then your subscriber.

It is OK to just reboot one server if needs be, I've done this before in production without any issues. In a situation where you want your subscriber to be master, you would reboot the subscriber first then the publisher.

I always prefer to have the publisher as the master, and I only allow the subscriber to be master for short periods of time. If your agent desktops are having difficulties connecting to the publisher I suggest opening a tac case.

Brian

Super Bronze

Re: Rebooting UCCX HA cluster

I have to say I'm not a huge fan of this kind of setup. I find that typically customers who prefer a particular server as master do so because their poorly written wallboard solution can't cope with HA and blindly reads the stats from the 'stale' server.

Customers with no wallboards generally don't care which server is 'master' as both work equally well.

Aaron Harrison

Principal Engineer at Logicalis UK

Please rate helpful posts...

Aaron Please remember to rate helpful posts to identify useful responses, and mark 'Answered' if appropriate!
903
Views
0
Helpful
2
Replies
CreatePlease login to create content