You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@cassandra.apache.org by "Stefania (JIRA)" <ji...@apache.org> on 2016/09/19 02:29:20 UTC

[jira] [Commented] (CASSANDRA-12664) GCCompactionTest is flaky

    [ https://issues.apache.org/jira/browse/CASSANDRA-12664?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15502058#comment-15502058 ] 

Stefania commented on CASSANDRA-12664:
--------------------------------------

cc [~blambov] and [~krummas].

> GCCompactionTest is flaky
> -------------------------
>
>                 Key: CASSANDRA-12664
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12664
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Local Write-Read Paths
>            Reporter: Stefania
>            Priority: Minor
>             Fix For: 3.x
>
>
> {{GCCompactionTest}} was introduced by CASSANDRA-7019 and appears to be flaky, see for example [here|https://cassci.datastax.com/view/Dev/view/sbtourist/job/sbtourist-CASSANDRA-9318-trunk-testall/lastCompletedBuild/testReport/org.apache.cassandra.cql3/GcCompactionTest/testGcCompactionStatic/]. 
> I think it's the same root cause as CASSANDRA-12282: the tables in the test keyspace are dropped asynchronously after each test, and this might cause additional flush operations for all dirty tables in the keyspace. See the [callstack|https://issues.apache.org/jira/browse/CASSANDRA-12282?focusedCommentId=15399098&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15399098] in 12282. 
> A possible solution is to use KEYSPACE_PER_TEST, which is instead dropped synchronously.



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