You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Russ Hatch (JIRA)" <ji...@apache.org> on 2014/05/06 00:40:15 UTC
[jira] [Commented] (CASSANDRA-7163) dtests should reuse existing
clusters where possible
[ https://issues.apache.org/jira/browse/CASSANDRA-7163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13990059#comment-13990059 ]
Russ Hatch commented on CASSANDRA-7163:
---------------------------------------
One possibility here might be some simple helpers in the Tester class to track a keyspace name on a per-test basis. A uuid would be great for this, and would prevent table name collisions. Then the tests can just ask for a keyspace when needed (and not really care what the actual value is).
> dtests should reuse existing clusters where possible
> ----------------------------------------------------
>
> Key: CASSANDRA-7163
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7163
> Project: Cassandra
> Issue Type: Test
> Reporter: Ryan McGuire
> Assignee: Ryan McGuire
> Priority: Minor
>
> Many dtests don't require special setup, specifically the cql tests. We can reuse the clusters we setup across multiple tests to save time.
> Suggestion: only share clusters across a single test suite, and don't share the cluster by default, turn it on per-class. If we share them more broadly than that we may get into weird states because some tests definitely do mess with the cluster setup.
--
This message was sent by Atlassian JIRA
(v6.2#6252)