Just a note (it'll be the last) to tell you that some "nagiosed" task
did call the 8093 port of the server (which is in fact our UIL2's port).
That clearly explains the stacks.
Thanks again for the good pointer !
Regards,
Arnaud
View the original post :
http://www.jboss.org/index.html?module
Thanks for the answers Adrian !
Regards,
Arnaud
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3829911#3829911
Reply to the post :
http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3829911
-
No idea about the 5 minute thing, it sounds like you have some client
connecting the UIL2 port, but it is not a messaging client.
This is how we discovered this bug in the first place.
Regards,
Adrian
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3829831#38298
Please forget last question , since we will wait for the 3.2.4 release.
Having been through JBoss's Thread dump, it appears that this morning, before start of
production, there is about 300 WriteTask threads left for only 5 or 6 ReadTask.
| Thread: UIL2.SocketManager.WriteTask#1314 : priority
Thanks a lot Adrian, I wasn't aware of it !
So , should I install 3.2.4 RC1 or is it a way for me to upgrade 3.2.3 release with
this modified class ?
Best regards,
Arnaud
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3829605#3829605
Reply to the post
This is a bug which has already been fixed:
http://cvs.sourceforge.net/viewcvs.py/jboss/jbossmq/src/main/org/jboss/mq/il/uil2/SocketManager.java?r1=1.1.4.8&r2=1.1.4.9
UIL2 does use more threads than OIL making it more concurrent/scalable
when there are multiple sessions on the same connection.
It
