Home › Forums › Luminis Platform / System

Webserver on IV becomes randomly unresponsive?

Wed, 2008-08-20 19:17 — jkstilwell

We have seen a situation where the luminis webserver becomes unresponsive randomly. This has happened twice now in the last 24 hours. There is nothing out of the ordinary in the cp.log files. All the other system components work just fine when this happens. So, I'm scratching my head over this one. We are running linux currently (RHEL 3 AS) on a VMWare virtual machine.

Has anyone else experienced anything similar to this with the base version of luminis IV?

0
No votes yet
‹ CAS SSO for INB, Self_Serv and Luminis Luminis IV migration problem ›

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Fri, 2008-08-22 13:37 — Jason

webserver unresponsive

If you look at the catalina.out and catalina.err logs, do they tell you anything? ($cp_root/products/tomcat/tomcat-cp/logs/)

I would suspect that you'll see your problem in there, which could be anything from needing a larger "maxThreads" value in your tomcat connector to needing increased java heap size.

Tue, 2008-08-26 13:35 — jkstilwell

Tomcat server issues on Luminis IV

We've been back and forth with Sungard support now for about 17 hours with a critical support case. It seems after applying some tuning recommendations to the tomcat webserver in $CP_WEB_ROOT/conf/server.xml and $CP_ROOT/bin/tomcat-cp-conf our portal is unable to handle logins. We're seeing an "java.lang.OutOfMemoryError: unable to create new native thread" error. This is odd considering the OS reports about 14Gb of free space available with zero swap space used.

We are running on VMWare currently. Has anyone else had issues with running Luminis IV on VMWare?

Jared Stilwell

Harding University
Application Development and Enhancement

Tue, 2008-08-26 18:47 — danderson

Login Issues

Jared,

Which log are you seeing the java.lang.OutOfMemoryError messages in?

We just went live on Monday with Luminis IV.1.0.23 and have had log in and performance issues. We've increased our java heap size and "maxThreads" value, applied all the tuning recommendations, and rebuilt our ldap on the recommendation of support. We are not running on VMWare though. Nor do we have parallel deployment.

One of the processes that we are seeing spike shortly before the system becomes unresponsive is "ns-slapd" (on Solaris). We've given support a thread dump, along with access and error logs for tomcat to help us identify the cause of the spikes.

Our first two days on Luminis IV have been rough.

Denise Anderson
Portal Administrator
Wright State University

Mon, 2008-10-06 18:52 — zigafoose

Jared, How did you fix this?

Jared,

How did you fix this?

Mon, 2008-10-06 19:17 — jkstilwell

No end in sight ;)

We haven't fixed it yet. We had a couple of steady weeks, but then last Friday things got extremely bogged down. The memory usage seemed extremely high over the weekend with 90-95% heap usage consistently, until we restarted (stopcp -a) and it seems to return to pretty decent responsiveness. We still don't know for sure where the bottle neck is or even how to effectively diagnose the issue other than do our own load testing which seems a bit out of reach at this point.

Jared Stilwell

Harding University
Application Development and Enhancement

Mon, 2008-10-06 23:24 — zigafoose

We're having the same issues.

Jared,

You mind contacting me to talk about this. We are having the same issues you are seeing in the same setup. Wanted to see what you tried so far and I can let you know what we've done.

eapelin(AT)hpu.edu

Thanks,
E. Apelin

Fri, 2008-08-29 14:16 — Jason

web server unresponsive

What are the system loads like when the web server becomes unresponsive?

Specifically, what do your database loads look like while the web server hangs? We had to go through several upgrades of both hardware and software, and tuning/indexes, before our database worked well enough to allow logins at a good rate.