[
https://issues.apache.org/jira/browse/HBASE-11861?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14294214#comment-14294214
]
Jonathan Hsieh commented on HBASE-11861:
----------------------------------------
I think for a first cut having the master do the mob compactions is reasonable.
Distributing the work will make this siginficatly more complicated but would
be a good follow on. (this will follow the life cycling similar to log
splitting -- initially used MR, then on master, and then later distributed).
I'm also thinking that with some of the notification bus ideas, it would be
better to us that instead of piggy backing off the distributed log splitting
mechanism or writing infrastructure to do distributed mob compaction.
> Native MOB Compaction mechanisms.
> ---------------------------------
>
> Key: HBASE-11861
> URL: https://issues.apache.org/jira/browse/HBASE-11861
> Project: HBase
> Issue Type: Sub-task
> Components: regionserver, Scanners
> Affects Versions: 2.0.0
> Reporter: Jonathan Hsieh
> Assignee: Jingcheng Du
> Attachments: 141030-mob-compaction.pdf, HBASE-11861-V1.diff,
> HBASE-11861-V2.diff, HBASE-11861.diff, mob compaction-out-of-region.pdf, mob
> compaction.pdf
>
>
> Currently, the first cut of mob will have external processes to age off old
> mob data (the ttl cleaner), and to compact away deleted or over written data
> (the sweep tool).
> From an operational point of view, having two external tools, especially one
> that relies on MapReduce is undesirable. In this issue we'll tackle
> integrating these into hbase without requiring external processes.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
