03-03-2013 06:15 PM - edited 03-03-2013 06:27 PM
We are seeing some weirdness in our legacy HP QC 9.2 installation, so wanted to see if anyone had seen this before.
A few weeks back, after a long period of time and when 50+ users started connecting within 10 minutes. we started seeing a ton of database sessions spawning and sitting around in Oracle that had been created by JDBC (we're using Websphere connecting to Oracle 9i) - once we hit the max configured sessions, QC totally hangs (debugs showing "GetProjectCustomizationData" running for 10+ minutes and users can't get in - eventually, the JVM spawns a ton of threads, maxes out the memory on the host and starts core-dumping/heap-dumping. Said core-dumps contain a lot of "OutOfMemoryError" and "Pipes Broken" errors at the time of the issue.
When we faced it this morning, we had the DBA kill the sessions from Oracle - once they did that, QC started responding like greased lightning - absolutely smoking fast.
We are currently on QC 9.2 Patch 17 and it has been running without many problems for years (we are in the process of moving toward migration to QC 11, but need to keep this one running a little while longer). We haven't changed any setups on our systems in 2+ years. We're planning on installing Patch 35 as per guidance from HP Support, but wanted to see if anyone else out there had dealt with this issue.
Thanks in advance for the help - this community has been very helpful during our investigation/troubleshooting.