[
https://issues.apache.org/jira/browse/FLINK-9788?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16646450#comment-16646450
]
ASF GitHub Bot commented on FLINK-9788:
---------------------------------------
StefanRRichter commented on issue #6810: [FLINK-9788] Fix ExecutionGraph
inconsistency for global failures when restarting
URL: https://github.com/apache/flink/pull/6810#issuecomment-428964847
LGTM 👍
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> ExecutionGraph Inconsistency prevents Job from recovering
> ---------------------------------------------------------
>
> Key: FLINK-9788
> URL: https://issues.apache.org/jira/browse/FLINK-9788
> Project: Flink
> Issue Type: Bug
> Components: Core
> Affects Versions: 1.6.0
> Environment: Rev: 4a06160
> Hadoop 2.8.3
> Reporter: Gary Yao
> Assignee: Till Rohrmann
> Priority: Blocker
> Labels: pull-request-available
> Fix For: 1.7.0, 1.6.2, 1.5.5
>
> Attachments: jobmanager_5000.log
>
>
> Deployment mode: YARN job mode with HA
> After killing many TaskManagers in succession, the state of the
> ExecutionGraph ran into an inconsistent state, which prevented job recovery.
> The following stacktrace was logged in the JobManager log several hundred
> times per second:
> {noformat}
> -08 16:47:18,855 INFO org.apache.flink.runtime.executiongraph.ExecutionGraph
> - Job General purpose test job (37a794195840700b98feb23e99f7ea24)
> switched from state RESTARTING to RESTARTING.
> 2018-07-08 16:47:18,856 INFO
> org.apache.flink.runtime.executiongraph.ExecutionGraph - Restarting
> the job General purpose test job (37a794195840700b98feb23e99f7ea24).
> 2018-07-08 16:47:18,857 DEBUG
> org.apache.flink.runtime.executiongraph.ExecutionGraph - Resetting
> execution vertex Source: Custom Source -> Timestamps/Watermarks (1/10) for
> new execution.
> 2018-07-08 16:47:18,857 WARN
> org.apache.flink.runtime.executiongraph.ExecutionGraph - Failed to
> restart the job.
> java.lang.IllegalStateException: Cannot reset a vertex that is in
> non-terminal state CREATED
> at
> org.apache.flink.runtime.executiongraph.ExecutionVertex.resetForNewExecution(ExecutionVertex.java:610)
> at
> org.apache.flink.runtime.executiongraph.ExecutionJobVertex.resetForNewExecution(ExecutionJobVertex.java:573)
> at
> org.apache.flink.runtime.executiongraph.ExecutionGraph.restart(ExecutionGraph.java:1251)
> at
> org.apache.flink.runtime.executiongraph.restart.ExecutionGraphRestartCallback.triggerFullRecovery(ExecutionGraphRestartCallback.java:59)
> at
> org.apache.flink.runtime.executiongraph.restart.FixedDelayRestartStrategy$1.run(FixedDelayRestartStrategy.java:68)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> {noformat}
> The resulting jobmanager log file was 4.7 GB in size. Find attached the first
> 5000 lines of the log file.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
