Robert Metzger created FLINK-6643:
-------------------------------------
Summary: Flink restarts job in HA even if NoRestartStrategy is set
Key: FLINK-6643
URL: https://issues.apache.org/jira/browse/FLINK-6643
Project: Flink
Issue Type: Bug
Components: JobManager
Affects Versions: 1.3.0
Reporter: Robert Metzger
Priority: Critical
While testing Flink 1.3 RC1, I found that the JobManager is trying to recover a
job that had the {NoRestartStrategy} set.
{code}
2017-05-19 15:09:04,038 INFO org.apache.flink.yarn.YarnJobManager
- Attempting to recover all jobs.
2017-05-19 15:09:04,039 DEBUG
org.apache.flink.runtime.jobmanager.ZooKeeperSubmittedJobGraphStore -
Retrieving all stored job ids from ZooKeeper under
flink/application_1494870922226_0064/jobgraphs.
2017-05-19 15:09:04,041 INFO org.apache.flink.yarn.YarnJobManager
- There are 1 jobs to recover. Starting the job recovery.
2017-05-19 15:09:04,043 INFO org.apache.flink.yarn.YarnJobManager
- Attempting to recover job f94b1f7a0e9e3dbcb160c687e476ca77.
2017-05-19 15:09:04,043 DEBUG
org.apache.flink.runtime.jobmanager.ZooKeeperSubmittedJobGraphStore -
Recovering job graph f94b1f7a0e9e3dbcb160c687e476ca77 from
flink/application_1494870922226_0064/jobgraphs/f94b1f7a0e9e3dbcb160c687e476ca77.
2017-05-19 15:09:04,078 WARN org.apache.hadoop.util.NativeCodeLoader
- Unable to load native-hadoop library for your platform... using
builtin-java classes where applicable
2017-05-19 15:09:04,142 INFO
org.apache.flink.runtime.jobmanager.ZooKeeperSubmittedJobGraphStore -
Recovered SubmittedJobGraph(f94b1f7a0e9e3dbcb160c687e476ca77, JobInfo(clients:
Set((Actor[akka.tcp://flink@permanent-qa-cluster-master.c.astral-sorter-757.internal:40391/user/$a#-155566858],EXECUTION_RESULT_AND_STATE_CHANGES)),
start: 1495206476885)).
2017-05-19 15:09:04,142 INFO org.apache.flink.yarn.YarnJobManager
- Submitting recovered job f94b1f7a0e9e3dbcb160c687e476ca77.
2017-05-19 15:09:04,143 INFO org.apache.flink.yarn.YarnJobManager
- Submitting job f94b1f7a0e9e3dbcb160c687e476ca77
(CarTopSpeedWindowingExample) (Recovery).
2017-05-19 15:09:04,151 INFO org.apache.flink.yarn.YarnJobManager
- Using restart strategy NoRestartStrategy for
f94b1f7a0e9e3dbcb160c687e476ca77.
2017-05-19 15:09:04,163 INFO
org.apache.flink.runtime.executiongraph.ExecutionGraph - Job recovers
via failover strategy: full graph restart
{code}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
