Garbage Collection/Indexing/Session Issue - Possible Memory Leak?

[a_lex_]

We have been having an issue on our author instance that we are currently trying to find a solution to.

After starting our author instance, our lucene indexing initiates, runs for a while, and then comes to a halt. At this moment, the GC goes crazy on the box and the entire heap is allocated to old generation with full-GC's occuring every second or so - basically killing the usability of the instance. We are trying to figure out what is causing this. We are currently running AEM6.0 SP2 and Oak 1.0.19

Things we have tried:

• Fully regenerated all indexes
• Upgraded Oak to 1.0.19, trying 1.0.23 soon
• Adjusted JVM settings to the following:

        

We are also seeing the following in the logs:

This is a very similar issue to what we are experiencing: http://stackoverflow.com/questions/30407105/aem-session-object-not-garbaged-collected

This is the memory pool behavior we are seeing:

If anyone has ANY ideas, they would be much appreciated!

[leeasling · Accepted solution]

We experienced similar issues with the same version of AEM after doing a migration and decided not to continue.  We then did the same with AEM 6.1 and the latest hot fixes and the issue no longer exists.

I don't know if upgrading to 6.1 is an option, but i think it would be worth looking at.

[Amit_Kumar]

This will be resolved if you can use:

• Oak version>> 1.2.2
• Jackrabbit API version 2.10
• CRX API version 2.5.30.

You can install SP3 and Oak 1.0.23 stand-alone hot fix. This will resolve the problem.

[a_lex_]

Updating to 6.1 is not an option at the moment unfortunately, so we applied SP3 and the Oak 1.0.23 hotfix in our test environment. I will update this in a few days with results for anyone else that is running into the same issue.