[jira] [Commented] (CASSANDRA-10742) Real world DateTieredCompaction tests

["Jeff Jirsa (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-10742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15018287#comment-15018287 ] 

Jeff Jirsa commented on CASSANDRA-10742:
----------------------------------------

Our primary production cluster that was on DTCS was 95% write, 5% read, with writes and reads at QUORUM. Writes come strictly in order, reads read random pieces of the past ~3-7 days. 

In order to hit the 'constant-recompaction' behavior observed in DTCS, you may want to occasionally restart nodes as well (or simulate other outages/pauses), because the foreground read repair is likely necessary for that behavior (writing with downgrading consistency may also be sufficient if your insertion rate is close enough to the box's throughput limit).



> Real world DateTieredCompaction tests
> -------------------------------------
>
>                 Key: CASSANDRA-10742
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-10742
>             Project: Cassandra
>          Issue Type: Test
>            Reporter: Marcus Eriksson
>
> So, to be able to actually evaluate DTCS (or TWCS) we need stress profiles that are similar to something that could be found in real production systems.
> We should then run these profiles for _weeks_, and do regular operational tasks on the cluster - like bootstrap, decom, repair etc.
> [~jjirsa] [~jshook] (or anyone): could you describe any write/read patterns you have seen people use with DTCS in production?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)