[ https://issues.apache.org/jira/browse/HBASE-19320?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16261796#comment-16261796 ]
Duo Zhang commented on HBASE-19320: ----------------------------------- I observed this problem years ago in java6, so until now the Java team still haven’t fixed it? The thread local cache is just a nightmare for an rpc server with hundreds of handler threads. I used to maintain a DBB pool with fixed size and if there is a large message I will write it chunk by chunk. After that I start to use netty and the problem is gone. Netty has its own buffer pool which works like a jemalloc. So, let’s start using netty by default? Thanks. > document the mysterious direct memory leak in hbase > ---------------------------------------------------- > > Key: HBASE-19320 > URL: https://issues.apache.org/jira/browse/HBASE-19320 > Project: HBase > Issue Type: Improvement > Affects Versions: 2.0.0, 1.2.6 > Reporter: huaxiang sun > Assignee: huaxiang sun > Attachments: Screen Shot 2017-11-21 at 4.43.36 PM.png, Screen Shot > 2017-11-21 at 4.44.22 PM.png > > > Recently we run into a direct memory leak case, which takes some time to > trace and debug. Internally discussed with our [~saint....@gmail.com], we > thought we had some findings and want to share with the community. > Basically, it is the issue described in > http://www.evanjones.ca/java-bytebuffer-leak.html and it happened to one of > our hbase clusters. > Create the jira first and will fill in more details later. -- This message was sent by Atlassian JIRA (v6.4.14#64029)