I have a timing issue. Sometimes I am unable to read the cache on failover. In the code that runs just after failover that gets triggered when a node gets made the new cluster loader, I read the cache but I get null back. I look in the cache afterwards and all the data is there.
To be specific, I have a node and its peer running concurently. I have a piece of restart code running as a jboss service and the method to invoke on restart is restoreRuntimeState ( see fragment of jboss-service.xml below)
When the method called "restoreRuntimeState" is invoked, the cache read returns no data. I look through the mbean for the cache and the data is there so I am wondering why this happening. This is happening intermittently i.e. it works sometimes and at other times it does not.
Can anybody help. Thanks
Are you saying you read the cache content right after your own node is restarted? Can you try to sleep it say 200 ms to see if that makes a difference. That will help troubleshooting.