[
https://issues.apache.org/jira/browse/YARN-10040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17006070#comment-17006070
]
Abhishek Modi commented on YARN-10040:
--------------------------------------
I looked into it and Opportunistic Scheduling is working fine. I checked
running distributed shell jobs locally with opportunistic containers and they
are also getting completed. The issue is
"yarn.nodemanager.opportunistic-containers-max-queue-length" is not being set
while bringing up nodemanager in the test and thus no nodemanager is accepting
the Opportunistic container.
Prior to YARN-9697, NM would accept Opportunistic containers even if max
opportunistic containers allowed on them is set to 0. That's why these tests
were passing before. I will provide a patch for fixing these tests.
> DistributedShell test failure on X86 and ARM
> --------------------------------------------
>
> Key: YARN-10040
> URL: https://issues.apache.org/jira/browse/YARN-10040
> Project: Hadoop YARN
> Issue Type: Bug
> Components: applications/distributed-shell
> Environment: X86/ARM
> OS: ubuntu1804
> Java 8
> Reporter: zhao bo
> Assignee: Abhishek Modi
> Priority: Major
>
> *
> org.apache.hadoop.yarn.applications.distributedshell.TestDistributedShell.testDSShellWithOpportunisticContainers
> *
> org.apache.hadoop.yarn.applications.distributedshell.TestDistributedShell.testDSShellWithEnforceExecutionType
> Please see the Apache Jenkins Test result:
> [https://builds.apache.org/job/hadoop-multibranch/job/PR-1767/1/testReport/]
>
> These 2 tests are failed on both X86 and ARM platform.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
