[
https://issues.apache.org/jira/browse/CASSANDRA-1876?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis updated CASSANDRA-1876:
--------------------------------------
Attachment: 1876-reformatted.txt
Thanks for the patch!
Attaching patch w/ CompactionManager changes reformatted to illustrate
http://wiki.apache.org/cassandra/CodeStyle.
There are two main things I would like to fix here:
- it would be cleaner to make the existing executor multithreaded, rather than
adding a second with somewhat cumbersome coordination
- this basically sub-bucketizes the buckets we've decided to compact, into
smaller groups, which means we're going to have to re-compact the sub-buckets
again very soon. this is a very high penalty to pay. (if that's what we
_want_, then a user can do that w/in the existing framework by lowering
MaximumCompactionThreshold, if we modify submitMinorIfNeeded to submit multiple
buckets' worth. either way we should not re-group a second time outside of
submitMinorIfNeeded.)
ideally we would parallelize within a single sstable (breaking out the
deserialize / merge / write stages) but this is Hard.
> Allow minor Parallel Compaction
> -------------------------------
>
> Key: CASSANDRA-1876
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1876
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Germán Kondolf
> Priority: Minor
> Attachments: 1876-reformatted.txt, compactionPatch-V2.txt
>
>
> Hi,
> According to the dev's list discussion (1) I've patched the CompactionManager
> to allow parallel compaction.
> Mainly it splits the sstables to compact in the desired buckets, configured
> by a new parameter: compaction_parallelism with the current default of "1".
> Then, it just submits the units of work to a new executor and waits for the
> finalization.
> The patch was created in the trunk, so I don't know the exact affected
> version, I assume that is 0.8.
> I'll try to apply this patch to 0.6.X also for my current production
> installation, and then reattach it.
> (1) http://markmail.org/thread/cldnqfh3s3nufnke
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
