cancel
Showing results for 
Search instead for 
Did you mean: 

Enqueue replication server does not terminate after failover

ashish_mishra2
Contributor
0 Kudos

Hi,

We are trying to setup high availability of enqueue server where in we have running enqueue server on node-A and ERS on node-B all the time.

Whenever enqueue is stopped on node-A, it automatically failovers on node-B but after replication of lock table, enqueue does not terminate the ERS running on node-B and as a result our enqueue and ERS both keeps running on the same host (Failover node-B) which should not be the case.

We havenu2019t configured polling in that scenario SAP note-1018968 depicts the same however this is applicable only for version-640 and 700.

Ideally when enqueue server switches to node-B, it should terminate the ERS on the same node after replication and then HA software would take care of its restart on node -A.

We have ERS running of version 701; could anyone please let me know if the same behaviour is common for 701 version as well?

Or there is any additional configuration to be done to make it working.

Thanks in advance.

Cheers !!!

Ashish

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

You need add the scenario in your HA Cluster setup where if your enqueue server failovers to node-B, after starting ENQ shutdown ERS.

Hope this helps.

Thanks,

Naveed

ashish_mishra2
Contributor
0 Kudos

Hi Naveed,

Stopping ERS is suppose to be taken care by SAP only and not the HA software.

Once ERS stops on node -B there would be a fault reported and as a result HA software will restart the ERS on node A.

Please refer to a section of SAP Note 1018968 - 'Enqueue replication server does not terminate after failover'

"Therefore, the cluster software must only organise the restart of the replication server and does not need to do anything for the shutdown."

Another blog about the same:

http://www.symantec.com/connect/blogs/veritas-cluster-server-sap

- After the successful lock table takeover, the Enqueue Replication Server will fault on this node (initiated by SAP). Veritas Cluster Server recognizes this failure and initiates a failover to a remaining node to create SAP Enqueue redundancy again. The Enqueue Replication Server will receive the complete Enqueue table from the Enqueue Server (SCS) and later Enqueue lock updates in a synchronous fashion.

So it is nothing about HA software, it is the SAP which should control ERS on node-B.

Cheers !!!

Ashish

ashish_mishra2
Contributor

Hi,

Solution to this issue was provided by SAP and ERS is functioning properly in way it should be.

Actually the ERS was getting restarted and not terminated, because of Restart command entry in

the profile START_ERS<nr.>_<hostname>

SAP suggested to replace Restart_Program_00* command with Start_Program_00* in Start profile of ERS instance.

After making above changes everything worked fine in the way it should be.

Thanks to all who provided their valuable feedbacks.

Regards,

Ashish

Answers (0)