[
https://issues.apache.org/jira/browse/HBASE-1892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12851943#action_12851943
]
Jonathan Gray commented on HBASE-1892:
--------------------------------------
Related to stack's comment, we should be doing what they do in bigtable. Notes
about this in HBASE-2375 as well.
Rather than: Decide to split, block updates, perform flush.
We should: Decide to split, perform flush, block updates, perform flush.
In the former, we have to block while writing what could be a very large
memstore. In the latter, we flush and then block. The final flush would only
have to flush data that was written during the first flush which should be
significantly less data in almost all cases.
> [performance] make hbase splits run faster
> ------------------------------------------
>
> Key: HBASE-1892
> URL: https://issues.apache.org/jira/browse/HBASE-1892
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: stack
> Fix For: 0.21.0
>
> Attachments: HBASE-1892.patch
>
>
> hbase-1506 tried and failed making splits faster in 0.20 context. This issue
> is about doing it in 0.21 where we'll have to tools to do.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
