We're running into a problem where something is breaking in our sessions. The exact situation is that our server is running with two agent-nodes on the same machine; one node-agent behaves as expected, and the other constantly produces the stack trace shown below.
The question is: has anyone encountered this type of problem before? There are remarkable few google hits for "close session with active objects", so I'm guessing this is a rare pheomenon.
Does this look like an SSL problem? A Sun-specific bug? A configuration problem with the high-availability database that stores these sessions between the two instances?
We are running (on some fancy-pants Sun hardware): Sun Java System Application Server Enterprise Edition 8.1_02 (build b06-fcs). As mentioned, there are two instances on the same machine. There is an http-load balancer in front of these two instances.
Any help or further clues -- even just "me-too" posts -- would be helpful. Thanks!