[
https://issues.apache.org/jira/browse/HBASE-3504?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12990704#comment-12990704
]
Jean-Daniel Cryans commented on HBASE-3504:
-------------------------------------------
I did some work previously to do as much as we can outside the updateLock, and
looking at the code there's only one thing remaining that slows us down inside
rollWriter once it gets the lock which is cleanupCurrentWriter that calls close
on the file. This is a bit involved and takes a long time. I think we could
switch the writing to the new log first and then close the file after we're
done with the lock (but we have to be careful).
> HLog performance improvement
> ----------------------------
>
> Key: HBASE-3504
> URL: https://issues.apache.org/jira/browse/HBASE-3504
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
>
> The HLog.updateLock protects the rolling of logs with concurrent writes to
> the HDFS log file. This is a scalability bottleneck for a workload that
> comprises mostly of counter-increments.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
