[
https://issues.apache.org/jira/browse/HBASE-15381?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15208548#comment-15208548
]
Ted Yu commented on HBASE-15381:
--------------------------------
For step 2:
bq. record the mapping <regionServer, regions> in Zookeeper.
Have you considered other persistence medium ? We should reduce reliance on
zookeeper.
For step 8:
bq. Check if all found regions in this region server that are recorded in
Zookeeper before the procedure is started are still online,
What if some regions are transitioned offline during the compaction process ?
bq. adding a new item to metadata is a more reasonable choice
And this is amenable to compatibility.
What if duration of MOB compaction is longer than the interval which
ScheduledChore runs ?
> Implement a distributed MOB compaction by procedure
> ---------------------------------------------------
>
> Key: HBASE-15381
> URL: https://issues.apache.org/jira/browse/HBASE-15381
> Project: HBase
> Issue Type: Improvement
> Components: mob
> Reporter: Jingcheng Du
> Assignee: Jingcheng Du
> Attachments: mob distributed compaction design.pdf
>
>
> In MOB, there is a periodical compaction which runs in HMaster (It can be
> disabled by configuration), some small mob files are merged into bigger ones.
> Now the compaction only runs in HMaster which is not efficient and might
> impact the running of HMaster. In this JIRA, a distributed MOB compaction is
> introduced, it is triggered by HMaster, but all the compaction jobs are
> distributed to HRegionServers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
