[
https://issues.apache.org/jira/browse/HBASE-7763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13572984#comment-13572984
]
stack commented on HBASE-7763:
------------------------------
bq. Seq nums are just breaking a tie on ordering for key values that have the
same timestamp. We've always said that it's non-deterministic which kv you will
get back if you have two writes at the same timestamp.
Yeah, it'd break the tie in favor of the mostly recently written which would be
less surprising than if it were the reverse. I suppose too, we would need to
always select a contiguous set of files -- contiguous in the order in which
they were written -- if we wanted to be sure to return the last written (if we
could select any set, we might not include the file that had the key w/ most
recent update). I believe it used to work this way (thats why the note in the
refguide that Sergey quotes).
This is great stuff you fellas are doing. It can make all the difference in
the world... any improvement found herein.
> Compactions not sorting based on size anymore.
> ----------------------------------------------
>
> Key: HBASE-7763
> URL: https://issues.apache.org/jira/browse/HBASE-7763
> Project: HBase
> Issue Type: Bug
> Components: Compaction
> Affects Versions: 0.96.0, 0.94.4
> Reporter: Elliott Clark
> Assignee: Elliott Clark
> Priority: Critical
> Fix For: 0.96.0, 0.94.6
>
> Attachments: HBASE-7763-trunk-TESTING.patch,
> HBASE-7763-trunk-TESTING.patch, HBASE-7763-trunk-TESTING.patch
>
>
> Currently compaction selection is not sorting based on size. This causes
> selection to choose larger files to re-write than are needed when bulk loads
> are involved.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
