01-29-2018 05:20 PM - edited 03-17-2019 12:04 PM
I have a unity connection cluster running on 10.5.2. Over the weekend i saw some alerts regarding cluster manager state and we kicked off the week with users complaining about MWI. When we checked the status of the cluster it showed both the servers were in secondary. We tried to make publisher the primary but it refused. So with TAC on the call we decided to give subscriber a reboot and then the connection messager service on subscriber craps out. Then the TAC suggests running the recovery on sub which we did. After the recovery we attempted a start of the conversation manager and it came back up. However both the nodes are in below state with MWI still being an issue.
Server Name Member ID Server State Internal State Reason
----------- --------- ------------ -------------- ---------------------
rih-ucon-p 0 Secondary Pri Failover Critical Service Down
cor-ucon-s 1 Initializing Sec Connecting Unknown
What was bogging me down all along is that how is just the MWI an issue in scenario like this. I expect all the services to work including MWI. But apparently that wasn't the case. Also, TAC suggested a reboot of publisher which i was kind of expecting and will be doing so in a couple of hours. But i am wondering if anyone else has encountered something like this before.
Solved! Go to Solution.
01-29-2018 09:07 PM
Just in case if anybody else runs into similar kind of issue, Reboot of the publisher fixed this. MWI would only work on the Primary server through the connection notifier unlike the regular voice messaging or call processing that doesn't care for a primary server and as long as conversation manager services are running. Rebooting the pub forced the primary role and that kicked off the MWI. After this we reset the DB replication to force the sub to be in sync. I also learnt during this process that stopping the calls on primary would result in MWI failure as well.
01-29-2018 09:07 PM
Just in case if anybody else runs into similar kind of issue, Reboot of the publisher fixed this. MWI would only work on the Primary server through the connection notifier unlike the regular voice messaging or call processing that doesn't care for a primary server and as long as conversation manager services are running. Rebooting the pub forced the primary role and that kicked off the MWI. After this we reset the DB replication to force the sub to be in sync. I also learnt during this process that stopping the calls on primary would result in MWI failure as well.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide