accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From GitBox <...@apache.org>
Subject [GitHub] [accumulo] andrewglowacki opened a new issue #1011: Slow memory leak
Date Tue, 05 Mar 2019 16:39:32 GMT
andrewglowacki opened a new issue #1011: Slow memory leak
URL: https://github.com/apache/accumulo/issues/1011
 
 
   I've observed this in the past, I've started to worry more recently.
   
   I've configured my tservers to use 24g (Xmx) of java memory, and 2g for the native map
(tserver.memory.maps.max). With this, one would expect the resident memory size (RES) to maybe
be around 28g-30g (including native thread memory/etc), but I don't think it should get above
30g.
   
   My tservers on fresh restart sit at approximately 24g of RES and slowly grow well above
30g to as high as 40g at times. What's causing this slow leak? Java memory usage is low, 4g-12g.
   
   After quite a bit of digging, I tentatively believe it's an issue with NativeMap creation/cleanup
or something not closing it's NativeMap. Doing a grep | wc -l for 'Allocated native map' and
'Deallocated native map' in a server debug log, the count differential is about 100K. Looking
at the code, I suspect the issue may be the tablet isn't closing it's in memory map when it
gets unloaded. Dumping the servers memory and analyzing with MAT, I can confirm there are
about 100K live NativeMap references.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

Mime
View raw message