[
https://issues.apache.org/jira/browse/CASSANDRA-5685?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis resolved CASSANDRA-5685.
---------------------------------------
Resolution: Duplicate
Proposed w/ more details in CASSANDRA-6602.
> time based compaction strategy
> ------------------------------
>
> Key: CASSANDRA-5685
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5685
> Project: Cassandra
> Issue Type: New Feature
> Reporter: Edward Capriolo
>
> We have a column family with all sites ttl. The idea is that the odd data
> files would roll off after they are passed retention. What ends happening is
> they get compacted with new file and don't roll off. Leveled might help but
> still not perfect.
> Suggesting a compaction strategy that will not compact sstables together if
> the oldest and youngesr column are too far apart. This wilk 3nsure old data
> does not compact with new xata and help run closer to capacity.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
