This last weekend 3-28 to 3-30, we upgraded our Hardware to all be virtual weblogic servers. Each of our environments has a single Admin server and a clustered A and B server. We also, upgraded our weblogic versions from 9.1 to 9.2.3 and we are using JRockit version jrockit_150_12. When we open our application and have multiple users in our environment one of the managed servers freezes and produces no log output. When we connect with jconsole we just stop getting output back, we do not even get a stack trace. Any ideas what is going on?
"Post Today 10:52:10 AM Subject: We upgraded from Weblogic 9.1 to 9.2.3 and JRockit hangs and freezes
What is the difference between the affected environment and the others?
Have you opened a ticket with Weblogic support?"
Yes we did that this morning and they are looking at it. We managed to get a binary dump from windbg and they have that now. Its about 500 mbs so I cant really post it. The difference in the environments is we have the admin, managed a and managed b on sep virtual servers, the version of JRockit is different since the weblogic version is different. Before our upgrade, we had the admin and managed a on the same physical server and managed b on its own physical server.
JRockit version: jrockit90_150_04
Weblogic version: 9.1
New Virtual Environment:
JRockit version: jrockit_150_12
Weblogic version: 9.2.3
Jared Sheehan wrote: the version of JRockit is different since the weblogic version is different.
I've had problems using WLST if my path had one version of Sun Java and Weblogic was running a different version. RMI would not work between the two, though in my case it was the client that failed. I'd try to get all the servers running the same version.
Joined: Mar 26, 2008
All the new application servers are running the same version of JRockit, I was referring to the old environments. Basically what happens when the issue occurs is that either Managed A or B will just freeze and be in an infinite loop and be in a "Failed not restartable state". We have to resume then stop then start again in the weblogic console to get the server back up. Oh one other thing we noticed, when i rdp into the affected managed server, and open the taskmgr, the java process is eating up 99 - 100% of our cpu. That obviously is very bad but what does it mean the server is doing in its endless loop where we get NO logs...