Yep, it is pretty crazy with strings in this dump. Consider the duplicate strings report:
(string, waste memory, # of duplicates) "RequestHandlerBase" 36803312 304888 "type" 33161040 152525 "name" 33146672 152474 "scope" 33144584 152444 "requestTimes" 8456864 38112 "numRequests" 8231896 38112 "numTimeouts" 8231896 38112 "numErrors" 8223976 38112 So the first string is repeated 304888 times, leading to a waste of 36mb. You get the picture. And yes, they're all referenced from com.yammer* (that problematic ConcurrentHashMap). Dawid On Wed, Dec 26, 2012 at 4:58 PM, Uwe Schindler <[email protected]> wrote: > It is there! Just scroll up in the thread and download it... I made the > jenkins run sticky: > http://jenkins.sd-datasolutions.de/job/Lucene-Solr-trunk-Linux/3421/artifact/heapdumps/java_pid13141.hprof > > ----- > Uwe Schindler > H.-H.-Meier-Allee 63, D-28213 Bremen > http://www.thetaphi.de > eMail: [email protected] > > >> -----Original Message----- >> From: [email protected] [mailto:[email protected]] On Behalf >> Of Dawid Weiss >> Sent: Wednesday, December 26, 2012 4:53 PM >> To: [email protected] >> Subject: Re: [JENKINS] Lucene-Solr-trunk-Linux (32bit/jdk1.6.0_37) - Build # >> 3421 - Failure! >> >> > I am still working on getting the root of the permgen issue. I have no >> hammer to tell visualvm to get a list of "interned" strings. How to >> "identify" >> an interned string? Having fun with OQL (Object Query Language)... >> >> Can you make that dump available? I'll check with yourkit. They have licenses >> for open source developers too and it's a great piece of software. >> >> > Dawid's note about Java 7 no longer puts interned strings into permgen >> seems "hardened" by the fact that only Java 6 builds have the permgen >> issue. >> >> Yeah, this is a known gotcha. In Java8 they've removed the permgen entirely >> (or are planning to; I didn't track progress of it recently) so it'll get >> simpler. >> >> Dawid >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] For additional >> commands, e-mail: [email protected] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
