cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yuki Morishita (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-5151) Implement better way of eliminating compaction left overs.
Date Fri, 18 Jan 2013 01:54:12 GMT


Yuki Morishita commented on CASSANDRA-5151:

It is. Except schema_* are flushed at the end of migration, other system cf used before commit
log recovery like local also flushes every time after its data updated. If we go using system
cf for this, we need to do the same, and doing so requires 2 flushes during one compaction,
one for inserting log and one for removing at the end.

I think plain file under cf data dir is easy to implement both for writing log and recovery.
> Implement better way of eliminating compaction left overs.
> ----------------------------------------------------------
>                 Key: CASSANDRA-5151
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>    Affects Versions: 1.1.3
>            Reporter: Yuki Morishita
>            Assignee: Yuki Morishita
>             Fix For: 1.2.1
>         Attachments: 5151-1.2.txt
> This is from discussion in CASSANDRA-5137. Currently we skip loading SSTables that are
left over from incomplete compaction to not over-count counter, but the way we track compaction
completion is not secure.
> One possible solution is to create system CF like:
> {code}
> create table compaction_log (
>   id uuid primary key,
>   inputs set<int>,
>   outputs set<int>
> );
> {code}
> to track incomplete compaction.

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:

View raw message