It's all about the answers!

Ask a question

CCM app crashing, McAfee to blame?


Francis Kemp (7212944) | asked Nov 06 '13, 9:18 p.m.
Our environment is CLM V403, WAS, DB2 V9.7, Enterprise Topology

Our CCM application is on its own VM with 8GB of RAM allocated for the JVM.  Over the last several months, we have had to reboot the CCM server once per week because it would become unresponsive. This week, we have needed to reboot 5 times already due to similar circumstances.

We originally thought that the JVM was running out of memory due to a memory leak but that doesn't appear to be the case this week.  We have noticed a McAfee process (we assume on demand scan) using a bunch of disk I/O at about the time the server fails.  We suspect changes may have been made this week to McAfee settings by the System Admin team. Unfortunately, we can't change these settings ourselves to test other configurations.

Anyone else seen anything like this?  Any suggestions for McAfee settings so that it plays nicely with CCM?


Thanks in advance,
Fran






Comments
Ralph Schoon commented Nov 07 '13, 3:08 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER

Have you looked into Configuring and Tuning? See https://jazz.net/wiki/bin/view/Deployment/DeploymentConfiguringAndTuning , especially https://jazz.net/wiki/bin/view/Deployment/TuningWAS and related topics? I can imagine that a virus scanner on the server, especially if it scans index and other folders can impact performance heavily, but maybe there are other effects contributing too?


Sumant Renukarya commented Nov 07 '13, 4:29 a.m.

Fran


--> Have you isolated and identified that the CCM application crash is due to McAfee? 
If yes, you can continue to troubleshoot based on the comment above w.r.t tuning and related topics

--> If this is a production server that you are dealing with here (I assume not), you should be opening a PMR with IBM Rational Support if you haven't yet. 


Francis Kemp commented Nov 14 '13, 10:47 p.m.

Ralph, Sumant,

Thanks for responding.  We were able to make the change last Thursday to stop the Jazz installation directory from being scanned by McAfee. It ran fine until today when it needed to be rebooted again. Clearly, turning off McAfee scanning in the Jazz installation directory helped, but the problem is still occurring.

We have Theming enabled to display Top and Bottom banners on pages for CCM (and the other Web Apps).  When we start getting errors that require the CCM server to be rebooted, we get Theming Errors in the ccm.log file.  (I forgot to print the actual error). 

We're not certain if the Theming errors are causing the server problem or are a symptom of the problem. Have you heard of any issues where Theming causes situations like this?

I administer a similar Enterprise CLM installation at another site. The main difference is that Theming is not used at the other site. That CCM server runs perfectly and never requires a reboot, so our suspicion is that Theming might be causing the problem somehow, 

Any thoughts? I did open a PMR with IBM last week for this problem.

Fran


Ralph Schoon commented Nov 15 '13, 5:49 a.m.
FORUM ADMINISTRATOR / FORUM MODERATOR / JAZZ DEVELOPER

Fran, that is out of the envelope of my experience. I hope support can help.


Kevin Ramer commented Nov 15 '13, 2:12 p.m.

Is your DB2 also on a VM ?  Our experience is that sooner or later performance of DB2 will degrade.   My group tried to deply Build Forge totally in a VM environment.  Ran into this kind of issue until database was moved to a physical machine. 

Maybe gather the active services of the jazz application from info you can find here:

https://jazz.net/forum/questions/124420/how-to-get-alert-if-a-service-has-been-running-for-more-than-x-hours-on-jts-or-rtc


One answer



permanent link
Stephanie Bagot (2.1k1513) | answered Nov 23 '13, 9:01 a.m.
FORUM MODERATOR / JAZZ DEVELOPER
Websphere has a hang mustgather - http://www-01.ibm.com/support/docview.wss?uid=swg21052641
I would recommend gathering the information mentioned and open up a PMR with the WAS IBM Support team to investigate the root cause of the outage.
Javacores and heapdumps will be helpful as well.

Your answer


Register or to post your answer.


Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.