[jira] [Updated] (HBASE-7944) Replication leaks file reader resource & not reset currentNbOperations

["Jeffrey Zhong (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-7944?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jeffrey Zhong updated HBASE-7944:
---------------------------------

    Attachment: hbase-7944-addendum-fix-0.patch

In the trunk, we found even test case pass while it throws NullPointerException.

The issue is triggered by the last check in where we set internal reader to null after we close the reader inside function repLogReader.closeReader(). The root cause is an old issue that we don't reset this.reader when we close repLogReader.

More details can be seen at hbase-7982.

Thanks,
-Jeffrey

                
> Replication leaks file reader resource & not reset currentNbOperations
> ----------------------------------------------------------------------
>
>                 Key: HBASE-7944
>                 URL: https://issues.apache.org/jira/browse/HBASE-7944
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Jeffrey Zhong
>            Assignee: Jeffrey Zhong
>             Fix For: 0.95.0, 0.98.0, 0.94.6
>
>         Attachments: 0.94.patch, hbase-7944-addendum-fix-0.patch, hbase-7944.patch
>
>
> When I trace test case TestReplicationQueueFailover, I found current replication logic has two following issues:
> 1) leak file reader when finishing read a file in function finishCurrentFile
> 2) never reset currentNbOperations(this can be seen in the attached error log) such as "currentNbOperations:12655 and seenEntries:0 and size: 0" inconsistant info. 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira