[
https://issues.apache.org/jira/browse/CASSANDRA-6851?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14003149#comment-14003149
]
Marcus Eriksson commented on CASSANDRA-6851:
--------------------------------------------
The idea is that we can do anticompaction at the same time over several
SSTables, if a key X is in all included files, it has been repaired in all
those files, and can go into the same repaired output file (re-reading my
description I see that it was not very clear, sorry about that)
I think what is needed is to pass in more sstables when creating the compaction
scanners instead of doing Arrays.asList(sstable);
Perhaps a first step could be to always give it two sstables so that we atleast
don't increase the number of sstables by doing anticompaction? I guess we could
experiment a bit here and find a heuristic that is good for most cases.
> Improve anticompaction after incremental repair
> -----------------------------------------------
>
> Key: CASSANDRA-6851
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6851
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Marcus Eriksson
> Assignee: Russell Alexander Spitzer
> Priority: Minor
> Labels: compaction, lhf
> Fix For: 2.1.1
>
>
> After an incremental repair we iterate over all sstables and split them in
> two parts, one containing the repaired data and one the unrepaired. We could
> in theory double the number of sstables on a node.
> To avoid this we could make anticompaction also do a compaction, for example,
> if we are to anticompact 10 sstables, we could anticompact those to 2.
> Note that we need to avoid creating too big sstables though, if we
> anticompact all sstables on a node it would essentially be a major compaction.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
