[
https://issues.apache.org/jira/browse/HBASE-4528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13121745#comment-13121745
]
dhruba borthakur commented on HBASE-4528:
-----------------------------------------
Hi Ted, if the HLog.sync() throws an exception, then it is not clear whether
the transaction made it to the disk or not, but the changes to memstore are
already made. So, it won't help to catch the exception from Hlog.sync() and
continue; instead it is better to fail completely, do you agree?
Kannan: thanks for pointing out the problem related to memstore flushes. I had
a solution for this deployed internally, will upload a new patch that will
include the fix to the problem. Essentially, the flush will wait for all
currently running transactions to quiesce before committing the flush.
> The put operation can release the rowlock before sync-ing the Hlog
> ------------------------------------------------------------------
>
> Key: HBASE-4528
> URL: https://issues.apache.org/jira/browse/HBASE-4528
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
> Attachments: appendNoSyncPut1.txt, appendNoSyncPut2.txt
>
>
> This allows for better throughput when there are hot rows. A single row
> update improves from 100 puts/sec/server to 5000 puts/sec/server.
--
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
