Recently, we fixed a bug in our JMS message producer that was not closing the connection to the Queue it was interacting with. Due to this, the ActiveThreadCount kept growing. At that point, we did not take the thread dump to verify count of UIL threads. Now, after the fix, we see the ActiveThreadCount growing but not so fast. However, when we take the thread dump from the web console, we see a numerous occurences of SelectorThread. Is it related to JMS? We googled for this snippet but did not find any precedents. The thread description in the dump is as follows:
Thread: SelectorThread : priority:5, demon:true, threadId:328, threadState:RUNNABLE, threadLockName:null
The JBoss installation in our system interfaces with Weblogic in two ways:
1. A message producer in Weblogic (8.1 sp6) posts a message to a Queue configured in JBoss (4.0.5). An MDB in JBoss listens for messages in this Queue and consumes them. We fixed the bug in the message producer where it was not closing QueueConnections to JBoss Queue.
2. A session bean is exposed as a webservice using Apache Axis. This session bean is deployed in JBoss. This session bean receives web service requests from our partners and synchronously calls an EJB in Weblogic.
Is it possible that the SelectorThread is being spawned by the second scenario where our application code, residing in JBoss, is making calls into Weblogic?
Thanks in advance,
This is Sun (IIOP) code and has nothing to do with JMS.
If you were using JacORB for IIOP then you might be able to get an answer in that forum.