cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yuki Morishita (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-5078) save compaction merge counts in a system table
Date Thu, 03 Oct 2013 17:11:44 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-5078?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13785358#comment-13785358
] 

Yuki Morishita commented on CASSANDRA-5078:
-------------------------------------------

Size-tiered/leveled compaction choose files that aren't compacting when determining candidates,
and try marking acquired candidates as compacting. If marking fails, then compaction would
be skipped(and go on to find next candidates).
So two or more threads can grab different set of SSTables inside the same ColumnFamily.

See, for example https://github.com/apache/cassandra/blob/cassandra-2.0/src/java/org/apache/cassandra/db/compaction/SizeTieredCompactionStrategy.java#L57
and https://github.com/apache/cassandra/blob/cassandra-2.0/src/java/org/apache/cassandra/db/DataTracker.java#L188.

> save compaction merge counts in a system table
> ----------------------------------------------
>
>                 Key: CASSANDRA-5078
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-5078
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Matthew F. Dennis
>            Assignee: lantao yan
>            Priority: Minor
>              Labels: lhf
>         Attachments: 5078-v3.txt, 5078-v4.txt, patch1.patch
>
>
> we should save the compaction merge stats from CASSANDRA-4894 in the system table and
probably expose them via JMX (and nodetool)



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message