[
https://issues.apache.org/jira/browse/SPARK-31845?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dongjoon Hyun updated SPARK-31845:
----------------------------------
Component/s: Tests
> DAGSchedulerSuite: Reuse completeNextStageWithFetchFailure and support
> checkAnswer
> ----------------------------------------------------------------------------------
>
> Key: SPARK-31845
> URL: https://issues.apache.org/jira/browse/SPARK-31845
> Project: Spark
> Issue Type: Sub-task
> Components: Spark Core, Tests
> Affects Versions: 3.1.0
> Reporter: jiaan.geng
> Priority: Major
>
> *First*
> DAGSchedulerSuite provides completeNextStageWithFetchFailure to make next
> stage occurs fetch failure.
> But many test case uses complete directly as follows:
> {code:java}
> complete(taskSets(2), Seq(
> (FetchFailed(BlockManagerId("hostA-exec2", "hostA", 12345),
> firstShuffleId, 0L, 0, 0, "ignored"),
> null)
> ))
> {code}
> We need to improve completeNextStageWithFetchFailure and reuse it.
> *Second*
> DAGSchedulerSuite also check the results show below:
> {code:java}
> complete(taskSets(0), Seq((Success, 42)))
> assert(results === Map(0 -> 42))
> {code}
> We can extract it as a generic method of `checkAnswer`.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
