[jira] [Commented] (CASSANDRA-6074) Read regression in 2.0.1

["Ryan McGuire (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-6074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13773744#comment-13773744 ] 

Ryan McGuire commented on CASSANDRA-6074:
-----------------------------------------

It appears the commit that introduced this is 0fc15e5ebd211a0921aee805b4f883ae6e2f728d. The immediately prior commit 830507b0312fad5f418deb5f34bd50c8636ebca6 is good.

!6074 Bisect.png!

[see data here|http://ryanmcguire.info/ds/graph/graph.html?stats=stats.6074.bisect.1.json&metric=interval_op_rate&operation=stress-read&smoothing=1]
                
> Read regression in 2.0.1
> ------------------------
>
>                 Key: CASSANDRA-6074
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6074
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Ryan McGuire
>         Attachments: 6074 bisect.png, stress 2.0.0 vs 2.0.01.png
>
>
> I'm seeing a regression in stress reads introduced in 2.0.1. Note, that this is a separate regression from CASSANDRA-5933.
> Setup
>  * 4 node cluster using default settings
>  * Write: -F 30000000 -n 30000000 -i 5 -l 2
>  * Read: -n 30000000 -o read -i 5
> I tried this three times on 2.0.0 and on 2.0.1(tentative; eb96db6c)
> 2.0.0 is fine, but 2.0.1 timed out of it's read 2 out of the 3 times, and on the one that did succeed, it struggled quite a bit at the start:
> !stress 2.0.0 vs 2.0.01.png!
> [Data can be viewed here|http://ryanmcguire.info/ds/graph/graph.html?stats=stats.2.0.0v2.0.1.json&metric=interval_op_rate&operation=stress-read&smoothing=4]
> I will bisect and update this when I find the commit that introduced this.

--
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: http://www.atlassian.com/software/jira