What could be inferred from this snippet of heap dump analysis?
RTC 4.0.3 crashed with java.lang.OutOfMemoryException. Analyzed heap dump using Heap Analyzer. It reports this:
2,099,867,416 bytes (58.6 %) of Java heap is used by 488 instances of com/ibm/team/workitem/common/internal/rcp/dto/impl/SaveResultDTOCustomImpl Contains 4 instances of the following leak suspects:I also see mentions of com/ibm/team/process/internal/common/model/settings/OperationBehaviorConfiguration objects a few nodes beneath the top leak suspect. So this leads me to believe that work item save somehow began to clog the pipes. Earlier this week configuration for SPoRT (won't mean much except to IBMr's) was added to a project. SPoRT config adds a number of items on Save Work Item behavior Suggestions? |
One answer
Hi Kevin.
I wouldn't read too much into the specifics of the Heap Analyzer results. While it's possible that your server actually has a memory leak, it's more likely that it just ran out of Java heap due to the increased demands of this "SPoRT" feature, or just a heavier workload in general.
It looks like your server is running with 4 GB of Java heap - is that right? Many customers find they need to bump that up to 6 or 8 GB or more as the number of users & projects increases. Try that and see if it takes care of the out-of-memory exceptions. See these documents for more info:
https://jazz.net/blog/index.php/2012/09/10/keeping-my-jazz-server-happy-and-my-users-overjoyed/ ("Pay Attention to Your JVM Settings)
-Matt
|
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.