[
https://issues.apache.org/jira/browse/LIVY-642?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16916511#comment-16916511
]
Zhefeng Wang edited comment on LIVY-642 at 8/27/19 8:22 AM:
------------------------------------------------------------
Hi [~runzhiwang] Sorry, this case happened in some of our batchsession. and it
caused by yarn AM has not sync state and still RUNNING, but final status has
changed to SUCCEEDED. so i cannot reproduce this state. this state is illegal
in livy and mapped to SparkApp.State.FAILED
was (Author: jassy):
Hi [~runzhiwang] Sorry, this case happened in some of our batchsession. and it
caused by yarn AM has not sync state and still RUNNING, but final status has
changed to SUCCEEDED. this state is illegal in livy and mapped to
SparkApp.State.FAILED
> A rare status happened in yarn cause SparkApp change into error state
> ---------------------------------------------------------------------
>
> Key: LIVY-642
> URL: https://issues.apache.org/jira/browse/LIVY-642
> Project: Livy
> Issue Type: Bug
> Components: Server
> Affects Versions: 0.6.0
> Reporter: Zhefeng Wang
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Some batch session execute successfully but return error state. in livy
> log，we find:
> {quote}{{2019-08-08 20:11:37,678 ERROR [Logging.scala:56] - Unknown YARN
> state RUNNING for app application_1559632632227_39801506 with final status
> SUCCEEDED.}}
> {quote}
> and this situation with yarn state RUNNING and final status SUCCEEDED is a
> *correct* state in yarn, so this should not mapped to SparkApp.State.FAILED
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
