AwareIM locking up (but not an AwarIM issue)

rocketman

Last couple of weeks I've been having some issues on my Virtual Private Server in that AwareIM will suddenly lock up after the login notification is fired.

On the most recent occasions this coincided with me doing some work on the server using remote desktop (windows server 2003) and I noticed that the server had become unresponsive to mouse clicks and was generally very sluggish. I opened up Task Manager and saw that the dual 2.1Ghz Quad core processors were thrashing around at 90% - but none of my processes were using anywhere near that much.

I resisted the temptation to shut down AwareIM and do a restart - as I've done in the past. After about 5 minutes the processors throttled back to below 50% and I was able to log in again.

So my question to any hardware gurus out there in forum land is. Assuming the above processors on a single dedicated machine dedicated to AwareIM and MySQL - and with say max 30 users at any one time, what should I expect the CPU usage to be?

Also, if anyone else is running a VPS - do you ever get up to 90% loads with your service provider? is this normal. At the moment I'm seriously considering castrating my SP with a blunt teaspoon - but I'll hold back to see what others think.

rocketman

This may be an unrelated issue - I just don't know.

There have been a number of occasions recently when members who are already logged in continue to work normally whilst other trying to log in just get the spinning ball.

It seems something (AwareIM, Tomcat, JBoss, MySQL - I know not what) refuses to accept any more connections.

In all such cases, the loginNotification fires and an entry is made in the user login log (A BO in my app) but the visual perspective never arrives on screen

Memory allocation is AwareIM = 512Mb, tomcat and jboss 128Mb

Any pointers appreciated

aware_support

Check Tomcat output when this happens - you may need to increase the number of concurrent connections for Tomcat

rocketman

I made some tweaks to MySql and tomcat and the system stayed up for 7 days until yesterday - Monday. Then the usual lock-up. Users who were in stayed working. Others couldn't log in.

This from the jBoss Log

02:11:03,385 WARN [TransactionImpl] Transaction TransactionImpl:XidImpl [FormatId=257, GlobalId=S15247674//63588, BranchQual=] timed out. status=STATUS_ACTIVE

02:11:32,075 WARN [TransactionImpl] Transaction TransactionImpl:XidImpl [FormatId=257, GlobalId=xxremovedxx//63589, BranchQual=] timed out. status=STATUS_ACTIVE

02:06:41,448 WARN [TransactionImpl] Transaction TransactionImpl:XidImpl [FormatId=257, GlobalId=xxremovedxx///156157, BranchQual=] timed out. status=STATUS_ACTIVE

02:06:48,917 WARN [TransactionImpl] Transaction TransactionImpl:XidImpl [FormatId=257, GlobalId=xxremovedxx///156158, BranchQual=] timed out. status=STATUS_ACTIVE

02:07:16,028 WARN [TransactionImpl] Lock contention, tx=TransactionImpl:XidImpl [FormatId=257, GlobalId=xxremovedxx///156157, BranchQual=]

Note [xxremovedxx] is simply me replacing my server identity

Found this post - but not being a jBoss man - need some explanation please

http://community.jboss.org/message/111660#111660