[
https://issues.apache.org/jira/browse/HBASE-2457?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12857905#action_12857905
]
stack commented on HBASE-2457:
------------------------------
Thanks Todd. I like your cut at a better algorithm. As per Jon, it may be the
case that the contiguous-only no longer applies after yesterdays commit that
leaves deletes in files until major compaction. Also, anything that improves
our i/o profile is ++ultra super important.
So, I suggest that we move the compaction review out of this issue into a new
blocker against 0.20.5. This new issue would be about trying your suggestion
and breaking out compaction so it standalone/testable. You're interested in
this one Jon?
For this issue, for 0.20.4, I suggest we apply the band-aid. For the super
pathological case you turned up above, the band-aid will mean we only do two
files at a time compressing -- way sub-optimal but at least we wouldn't be
stuck.
> RS gets stuck compacting region ad infinitum
> --------------------------------------------
>
> Key: HBASE-2457
> URL: https://issues.apache.org/jira/browse/HBASE-2457
> Project: Hadoop HBase
> Issue Type: Bug
> Affects Versions: 0.20.4
> Reporter: Todd Lipcon
> Priority: Critical
> Attachments: log.gz, stack
>
>
> Testing 0.20_pre_durabil...@934643, I ended up in a state where one region
> server got stuck compacting a single region over and over again forever. This
> was with a special config with very low flush threshold in order to stress
> test flush/compact code.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
