Connector java process takes long time to shut down in cluster failed over

Document ID : KB000111168
Last Modified Date : 15/08/2018
Show Technical Document Details
Issue:
We have recently uninstalled the old UIM connector 3.6.0 and installed 3.9.0 in a the cluster environment. The installation was successful but when we perform fail over from one node to another the connector java process is not getting stopped automatically while the node is failing over, it takes 5-10 minute for the java process to exit. We have to stop the process manually to fail over the node.
Resolution:
- In CA\Catalyst\CatalystConnector\jsw\conf\Catalyst-Container.conf we set the following
  wrapper.shutdown.timeout-60
  wrapper.jvm_exit.timeout=60
- Offlined Catalyst Container service from cluster manager
- Ended the java.exe process from task manager
- Deleted all files from \CA\Catalyst\CatalystConnector\container\registry-cache and from CA\Catalyst\CatalystConnector\container\nls- store
- Onlined the Catalyst Container service from cluster manager.
- Waited until the UIM connector shows online
- Offlined Catalyst Container service from cluster manager again and now the java process exited within a minute.
- We failed over and did the same and java process took around a minute to exit.