Given that I’m developer and knowledge of app servers is scoped to the basics such as deploying, declaring / mapping resources.
My first step is to investigate the app and server logs (sysout and syserr) to find any clues and see the last thing the server instance did before the shutdown. For example with a locally deployed Tomcat, if it suddenly shutsdown perhaps due to memory issue, you would see a stack trace related to PermGen.
My question is, if there isn’t any indication that something went wrong in the logs before the shutdown, where do I go from there? What other places should I look into, and what would be the possible causes of this sudden shutdown?
In my case, this is a WAS 6.1 where there are several instances, where only 1 instance died. This instance that died has several apps, not more than 5.
I’d recommend downloading Visual VM 1.3.2, installing all the available plug-ins, and attaching to the JVM process after restarting your Java EE server. It’ll let you see all the threads, all the generations of the heap, memory usage by class, etc. If you can reproduce the scenario that caused the shutdown, Visual VM 1.3.2 will give you a lot more visibility and insight.