Hi All.

At UNISA uir.unisa.ac.za, we have upgraded from DSpace version 1.4.2 to
version 1.6.2. All looked well when we tested the system, but we have
been picking up a few issues since we have gone live.

The first issue was that there was not enough perm-gen space for the
JVM. I rectified that by adding 

-Xmx4096M -Xms2048M -Dfile.encoding=UTF-8 -XX:+UseConcMarkSweepGC -XX:
+CMSPermGenSweepingEnabled -XX:+CMSClassUnloadingEnabled
-XX:MaxPermSize=512M

to my startup script and using lambda probe, monitored the machine for
16 days in which there were no problems.

Then we started getting
java.util.NoSuchElementException: Timeout waiting for idle object

Research showed that this was due to insufficient database connections.
I increased the db.maxconnections = 100 from the default 30 in
dspace.cfg. I used 100 as postgres default is also 100. I also increased
shared_buffers = 50000 and effective_cache_size = 100000 in
postgresql.conf as well as increasing postgres' maximum connections to
200.

The system seems to run pretty stable for a day, but then starts giving
these errors again.

This specific error is not logged in catalina.out, but I get a lot of 
Feb 16, 2011 9:42:44 AM org.apache.catalina.loader.WebappClassLoader
clearThreadLocalMap
SEVERE: The web application [] created a ThreadLocal with key of type
[java.lang.ThreadLocal] (value [java.lang.ThreadLocal@7d5f26a2]) and a
value of type [org.apache.cocoon.environment.internal.EnvironmentStack]
(value [[]]) but failed to remove it when the web application was
stopped. This is very likely to create a memory leak.

These are continuously being logged.

So my question is:
Are there any best practises for tuning postgres for DSpace, and how can
I get rid of the above error message?

Many thanks.

Shaun Donovan.




------------------------------------------------------------------------------
The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
Pinpoint memory and threading errors before they happen.
Find and fix more than 250 security defects in the development cycle.
Locate bottlenecks in serial and parallel code that limit performance.
http://p.sf.net/sfu/intel-dev2devfeb
_______________________________________________
DSpace-tech mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-tech

Reply via email to