[
https://issues.apache.org/jira/browse/HBASE-16013?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ramkrishna.s.vasudevan updated HBASE-16013:
-------------------------------------------
Attachment: HBASE-16013.patch
In the default case a simple memstore iterator makes things much faster. I
could see the improvement in the cluster. But this is a thing to be discussed
and see how can this be implemented. In a default case may be we don't even
need a compaction, just directly flush from the pipelines?
> Compaction process can be improved for a default case
> ------------------------------------------------------
>
> Key: HBASE-16013
> URL: https://issues.apache.org/jira/browse/HBASE-16013
> Project: HBase
> Issue Type: Sub-task
> Affects Versions: 2.0.0
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Fix For: 2.0.0
>
> Attachments: HBASE-16013.patch
>
>
> There is an ongoing discussion on how to handle the default case (no
> duplicates/deletes) and how to use the new CompactingMemstore in such a case.
> One improvement is to avoid the StoreScanner and go with a simple
> MemstoreScanner. This avoids lot of comparisons that the StoreScanner does.
> Just raising this sub-task as part of some analysis going on.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
