[ 
https://issues.apache.org/jira/browse/HBASE-4733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13144289#comment-13144289
 ] 

Andrew Purtell commented on HBASE-4733:
---------------------------------------

bq. FYI this patch applies well to 0.90.4, those folks that use it in 
production should be able to fix their REST issue right away. 

Yes
                
> Rest client does not close zookeeper sessions (leaking sessions for each GET 
> or PUT)
> ------------------------------------------------------------------------------------
>
>                 Key: HBASE-4733
>                 URL: https://issues.apache.org/jira/browse/HBASE-4733
>             Project: HBase
>          Issue Type: Bug
>          Components: rest
>    Affects Versions: 0.90.4
>         Environment: Fedora 13.
>            Reporter: jack levin
>              Labels: newbie
>             Fix For: 0.90.4
>
>         Attachments: my.patch
>
>
> This will appear in the log once the zookeeper connection/session leaking 
> will grow to 2000, zookeeper won't be able to accept any more connections 
> causing REST RPC calls to fail, here is the log when the problem is in 
> progress:
> 2011-10-26 01:35:49,270 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.3 - max is 2000
> 2011-10-26 01:35:49,300 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.4 - max is 2000
> 2011-10-26 01:35:49,317 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.4 - max is 2000
> 2011-10-26 01:35:49,321 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.3 - max is 2000
> 2011-10-26 01:35:49,323 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.101.255.3 - max is 2000
> 2011-10-26 01:35:49,367 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.101.255.2 - max is 2000
> 2011-10-26 01:35:49,375 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.4 - max is 2000
> 2011-10-26 01:35:49,382 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.2 - max is 2000
> 2011-10-26 01:35:49,404 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.101.255.2 - max is 2000
> 2011-10-26 01:35:49,429 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.2 - max is 2000
> 2011-10-26 01:35:49,439 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.3 - max is 2000
> 2011-10-26 01:35:49,469 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.3 - max is 2000
> 2011-10-26 01:35:49,489 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.101.255.3 - max is 2000
> 2011-10-26 01:35:49,501 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.101.255.2 - max is 2000
> 2011-10-26 01:35:49,584 WARN org.apache.zookeeper.server.NIOServerCnxn: Too 
> many connections from /10.102.255.2 - max is 2000
> After the fix, the log looks much better, and we can observe zookeeper 
> connections closing after every RPC call:
> 2011-11-02 15:50:14,339 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Accepted socket connection from /10.101.255.2:37225
> 2011-11-02 15:50:14,339 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Client attempting to establish new session at /10.101.255.2:37225
> 2011-11-02 15:50:14,340 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Established session 0x3352857cb1f19f with negotiated timeout 180000 for 
> client /10.101.255.2:37225
> 2011-11-02 15:50:14,363 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Closed socket connection for client /10.101.255.2:37225 which had sessionid 
> 0x3352857cb1f19f
> 2011-11-02 15:50:14,723 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Accepted socket connection from /10.101.255.2:38089
> 2011-11-02 15:50:14,723 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Client attempting to establish new session at /10.101.255.2:38089
> 2011-11-02 15:50:14,725 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Established session 0x3352857cb1f1a0 with negotiated timeout 180000 for 
> client /10.101.255.2:38089
> 2011-11-02 15:50:14,771 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Closed socket connection for client /10.101.255.2:38089 which had sessionid 
> 0x3352857cb1f1a0
> 2011-11-02 15:50:16,326 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Accepted socket connection from /10.101.255.3:34085
> 2011-11-02 15:50:16,326 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Client attempting to establish new session at /10.101.255.3:34085
> 2011-11-02 15:50:16,328 INFO org.apache.zookeeper.server.NIOServerCnxn: 
> Established session 0x3352857cb1f1a1 with negotiated timeout 180000 for 
> client /10.101.255.3:34085

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to