Repeated NPEs using SFSBs

mrhightech May 29, 2006 3:17 PM

I have done some tests on clustering stateful beans and came accross the following exception, repeated over and over, when I tried removing the tutorial.jar file from the farm directory.

Incidently, I had beans online when I did this, and a message appared on the other clustered server saying it was unable to delete the jar from the farm directory.

so first, the exception and then a question

5:09:09,015 ERROR [STDERR] java.lang.NullPointerException
5:09:09,093 ERROR [STDERR] at org.jboss.ejb3.cache.tree.StatefulEvictionPolicy.evict(StatefulEvictionPolicy.java:48)
5:09:09,187 ERROR [STDERR] at org.jboss.cache.eviction.LRUAlgorithm.evictCacheNode(LRUAlgorithm.java:197)
5:09:09,265 ERROR [STDERR] at org.jboss.cache.eviction.LRUAlgorithm.prune(LRUAlgorithm.java:303)
5:09:09,343 ERROR [STDERR] at org.jboss.cache.eviction.LRUAlgorithm.process(LRUAlgorithm.java:55)
5:09:09,421 ERROR [STDERR] at org.jboss.cache.eviction.EvictionTimerTask.run(EvictionTimerTask.java:37)
5:09:09,515 ERROR [STDERR] at java.util.TimerThread.mainLoop(Timer.java:512)
5:09:09,593 ERROR [STDERR] at java.util.TimerThread.run(Timer.java:462)

The question is: What is the proper way to undeploy a jar accross a cluster using the farm directory.

Thanks!

1. Re: Repeated NPEs using SFSBs

brian.stansberry May 29, 2006 4:56 PM (in response to mrhightech)

The exception appears to be due to http://jira.jboss.com/jira/browse/EJBTHREE-581 . It keeps happening over and over because the eviction algorithm keep retrying the eviction.

However, I don't see how this would prevent undeployment of the farmed EJB. The way to remove the farmed deployment is to remove the deployment unit (e.g. tutorial.jar) from the farm directory on a running node. Sounds like that's what you did.

What exactly did it say on the other server?
Actions