[
https://issues.apache.org/jira/browse/FLINK-27550?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Junfan Zhang updated FLINK-27550:
---------------------------------
Description:
When invoking the method of {{checkYarnQueues}} in YarnClusterDescriptor,
it will check the specified yarnQueue whether exists in the queues gotten by
the YarnClient.QueueInfo.
However when using the capacity-scheduler, the yarn queues path should be
retrieved by the api of {{QueueInfo.getQueuePath}}
instead of {{getQueueName}}.
Due to this, it will always print out the yarn all queues log, but it also can
be submitted to Yarn successfully.
The api of getQueuePath is introduced in the latest hadoop
version(https://issues.apache.org/jira/browse/YARN-10658), so it's hard to
solve this problem in the older hadoop cluster.
According to the above description, the process of checking is unnecessary.
was:
When invoking the method of {{checkYarnQueues}} in YarnClusterDescriptor,
it will check the specified yarnQueue whether exists in the queues gotten by
the YarnClient.QueueInfo.
However when using the capacity-scheduler, the yarn queues path should be
retrieved by the api of {{QueueInfo.getQueuePath}}
instead of {{getQueueName}}.
Due to this, it will always print out the yarn all queues log, but it also can
be submitted to Yarn successfully.
According to the above description, the process of checking is unnecessary.
> Remove checking yarn queues before submitting job to Yarn
> ---------------------------------------------------------
>
> Key: FLINK-27550
> URL: https://issues.apache.org/jira/browse/FLINK-27550
> Project: Flink
> Issue Type: Improvement
> Components: Deployment / YARN
> Reporter: Junfan Zhang
> Priority: Major
>
> When invoking the method of {{checkYarnQueues}} in YarnClusterDescriptor,
> it will check the specified yarnQueue whether exists in the queues gotten by
> the YarnClient.QueueInfo.
> However when using the capacity-scheduler, the yarn queues path should be
> retrieved by the api of {{QueueInfo.getQueuePath}}
> instead of {{getQueueName}}.
> Due to this, it will always print out the yarn all queues log, but it also
> can be submitted to Yarn successfully.
> The api of getQueuePath is introduced in the latest hadoop
> version(https://issues.apache.org/jira/browse/YARN-10658), so it's hard to
> solve this problem in the older hadoop cluster.
> According to the above description, the process of checking is unnecessary.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
