[
https://issues.apache.org/jira/browse/HBASE-17035?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15643987#comment-15643987
]
ramkrishna.s.vasudevan commented on HBASE-17035:
------------------------------------------------
Let me explain what I see. I saw your latest comment and patch.
But am not sure if it is going to solve this problem.
Since requestLogRoll is having a map if the request log comes for the same log
we roll it only once.
But I can see that syncCompleted is called at the same time and there I feel
that there is a chance that the writer reference got updated and so we have two
immediate rolls happening one for the old file and the other for the new file
that just got rolled since the writer instance got updated.
BTW after applying your patch (the new one in HBASE-17021) the time of
execution of PE remains the same.
So log roll is not the only culprit here in my opinion. There is something more
to it. Let us check with WALPE.
> Check why we roll a wal writer at 10MB when the configured roll size is 120M+
> with AsyncFSWAL
> ---------------------------------------------------------------------------------------------
>
> Key: HBASE-17035
> URL: https://issues.apache.org/jira/browse/HBASE-17035
> Project: HBase
> Issue Type: Sub-task
> Components: wal
> Affects Versions: 2.0.0
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Fix For: 2.0.0
>
>
> Found this when addressing HBASE-16890. It is one of the possible reason that
> why AsyncFSWAL performs worse than FSHLog when running PE tool.
> https://issues.apache.org/jira/browse/HBASE-16890?focusedCommentId=15636688&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15636688
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
