[
https://issues.apache.org/jira/browse/FLINK-5268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15821889#comment-15821889
]
Jark Wu edited comment on FLINK-5268 at 1/13/17 3:03 PM:
---------------------------------------------------------
Hi [~fhueske] , do we have the case that an IT case should run both on
collection environment and cluster environment ?
And when should we use cluster environment ? To cover the combine phase is the
only case I know. I think we should add some comments to remind developers when
to use {{TableProgramsClusterTestBase}} .
was (Author: jark):
Hi [~fhueske] , do we have the case that an IT case should run both on
collection environment and cluster environment ?
And when should we use cluster environment ? I think we should add some
comments to remind developers when to use {{TableProgramsClusterTestBase}} .
> Split TableProgramsTestBase into TableProgramsCollectionTestBase and
> TableProgramsClusterTestBase
> -------------------------------------------------------------------------------------------------
>
> Key: FLINK-5268
> URL: https://issues.apache.org/jira/browse/FLINK-5268
> Project: Flink
> Issue Type: Sub-task
> Components: Table API & SQL
> Affects Versions: 1.2.0
> Reporter: Fabian Hueske
> Priority: Minor
>
> Currently the {{TableProgramsTestBase}} allows to run tests on a collection
> environment and a MiniCluster by setting a testing parameter. This was done
> to cover different execution path. However, testing on a MiniCluster is quite
> expensive and should only be done in rare cases.
> I propose to split the {{TableProgramsTestBase}} into
> * {{TableProgramsCollectionTestBase}} and
> * {{TableProgramsClusterTestBase}}
> to have the separation of both execution backends more clear.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
