Unity Connection 8.x Cluster "not reachable" issue

[rajesh.kumar]

Hi

We have Unity 8.0.3.20000-2 cluster.

Publisher and secondary is configured.

There is issue in cluster, secondary is always "not reachable"

Here I have attached the screenshot and cluster status.

Pls. suggest

Rgds

Rajesh Kumar

[Aman Soi]

Hi Rajesh,

Split Brain Recovery occurs when both servers assume the role of Primary.Even though I have never implemented  but do u find Server Redundancy Manager Service in cisco unity serviceability servers , so that it can be restarted .

I was seeing your earlier POST also where similar kind of problem is there.

regds,

aman

[rajesh.kumar]

Dear Aman

We have same problem with multiple customer. What I observed is DBRplication is in good condition, but only cluster is giving trouble. Previous similar case is similar issue but version is CUC 9.1.

I have restared the entire Unity Connection cluster. But still the problem is same.

If you can pls. suggest me on my previuos posted issue on CUC 9.1 cluster issue.

Rgds

Rajesh Kumar

[Aman Soi]

Hi Rajesh,

I have been trying to search for the problem which u have been facing in both Unity Clusters but could not find anything related to it.

I could find one bug but still that is also not helpful.Anybody who has faced this problem could guide or u can open the TAC case.

CSCtc52702 - Early process restart during split brain resolution results in CP fail

Symptom:

During recovery from split brain condition on the Secondary side, there is a race condition between DB Copy and restart of the group 0 processes. In one such condition, if group 0 processes are restarted before DB copy is completed then SGA which is a group 0 process maps the uncompleted shared memory tables which results in SGA crashing when it gets data to be updated in shm.

Conditions:

Recovery from Split Brain Condition.

Workaround:

Bring the secondary platform down, clear the shm and then bring the secondary platform up.

Further Problem Description:

Group 0 Processes restarted before DB Copy is completed.

May be cisco has come up with patch upgradation to resolve the issue.

regds,

aman

[rajesh.kumar]

Dear Aman

did you get change to look into similar issue on CUC 9.1 ?

bug issue is with CUC 9.1 as well ? During this issue, DB relication seems to be alright.

https://supportforums.cisco.com/message/3956379#3956379

Rgds

Rajesh Kumar

[yahsiel2004]

Are you using DNS? If so, verify that the CUC servers can reach the DNS server. If the CUC servers are on different subnets, verify that they can reach each other. I would power down both servers, then power on the CUC Pub and wait for it to be the primary and that the conversation manger is running, then power on the CUC Sub and verify that it becomes the secondary.

Regards,

Yosh

[rajesh.kumar]

Dear Yosh

DNS is enabled and both CUC can access each other using DNS name.

Would you pls. look into anoter post, where I am facing similar issue with CUC 9.1

https://supportforums.cisco.com/message/3956379#3956379

Rgds

Rajesh Kumar

[Trying tobeone]

Hi Rajesh

You mean to say the reachability and routing part is fine.

So in that case can you make all the services down on pub and sub and try activating all services again.

Regards

TT

[yahsiel2004]

Did you power down both servers, then power on the CUC Pub and wait for it to  be the primary and that the conversation manger is running, then power  on the CUC Sub and verify that it becomes the secondary?

Please rate all helpful replies.

HTH

Regards,

Yosh