[
https://issues.apache.org/jira/browse/FLINK-24727?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17444385#comment-17444385
]
Yangze Guo commented on FLINK-24727:
------------------------------------
The job's parallelism is set to 3 while the `taskmanager.numberOfTaskSlots` is
set to 1. So, we need at least 3 TMs. It's a bit wired that the cluster
resource is `<memory:5000, vCores:2>`. So, we can only start the JM with one TM
at all. Does anyone know the context of this e2e case or what is the expected
behavior of this case?
> Running Kerberized YARN per-job on Docker test (no fat-jar) fails on Azure
> --------------------------------------------------------------------------
>
> Key: FLINK-24727
> URL: https://issues.apache.org/jira/browse/FLINK-24727
> Project: Flink
> Issue Type: Bug
> Components: Deployment / YARN, Tests
> Affects Versions: 1.14.0
> Reporter: Till Rohrmann
> Priority: Critical
> Labels: test-stability
> Fix For: 1.15.0, 1.14.1
>
>
> The e2e {{Running Kerberized YARN per-job on Docker test (no fat-jar)}} fails
> on Azure.
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=25772&view=logs&j=c88eea3b-64a0-564d-0031-9fdcd7b8abee&t=070ff179-953e-5bda-71fa-d6599415701c
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
