[jira] [Reopened] (CASSANDRA-10879) HSHA dtest for closing connections almost always fails on CASSCI developer branches

["Branimir Lambov (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/CASSANDRA-10879?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Branimir Lambov reopened CASSANDRA-10879:
-----------------------------------------

No, this is a different and IMO more serious issue. The test does not flap or fail at all for 2.1 or 2.2, unless you are testing a developer branch, where it usually fails.

This looks like a problem of the infrastructure. A fix for 10863 _could_ fix this, but most probably won't.

> HSHA dtest for closing connections almost always fails on CASSCI developer branches
> -----------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-10879
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-10879
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Testing
>            Reporter: Branimir Lambov
>            Priority: Critical
>
> The test consistently succeeds on 2.1/2.2 branches, but has been failing almost every time on developer ones for quite a while, for example [9258 dtest|http://cassci.datastax.com/view/Dev/view/blambov/job/blambov-dikang85-9258-dtest/lastCompletedBuild/testReport/], [10817 dtest|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-10817-dtest/], [10059 (committed) dtest|http://cassci.datastax.com/view/Dev/view/blambov/job/blambov-10059-dtest/lastCompletedBuild/testReport/], [unchanged 2.2 dtest|http://cassci.datastax.com/view/Dev/view/blambov/job/blambov-cassandra-2.2-dtest/lastCompletedBuild/testReport/] and many others. The failures cannot be genuine if the same code committed to the main branch no longer fails the test.
> This is making it hard to identify genuine failures and blocking CASSANDRA-9669 in particular.



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