[
https://issues.apache.org/jira/browse/NIFI-9293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17481397#comment-17481397
]
ASF subversion and git services commented on NIFI-9293:
-------------------------------------------------------
Commit b5dd35431e2251802bd0c13c5de016477ede49c5 in nifi's branch
refs/heads/main from Mark Payne
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=b5dd354 ]
NIFI-9293: Ensure that we properly set the scheduled flag in the LifecycleState
when stopping processors. Added system test to verify that @OnScheduled,
onTrigger, @OnUnscheduled, @OnStopped are all called and in the expected order
This closes #5706.
Signed-off-by: Peter Turcsanyi <[email protected]>
> Stateless engine does not stop processors properly
> --------------------------------------------------
>
> Key: NIFI-9293
> URL: https://issues.apache.org/jira/browse/NIFI-9293
> Project: Apache NiFi
> Issue Type: Bug
> Components: NiFi Stateless
> Reporter: Peter Turcsanyi
> Assignee: Mark Payne
> Priority: Major
> Fix For: 1.16.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> When the embedding application calls {{StatelessDataflow.shutdown()}} to stop
> the stateless flow and all its processors, then the expected behaviour would
> be (similar to traditional NiFi):
> - call {{@OnUnscheduled}} method(s) of the processor
> - wait for {{onTrigger()}} to finish up
> - call {{@OnStopped}} method(s) of the processor
> Currently {{@OnUnscheduled}} is not called at all and {{@OnStopped}} gets
> called while {{onTrigger()}} is still running.
> These may lead to improper shutdown of the components: e.g. ConsumeJMS stays
> connected to the message broker after is has been stopped (up until the JVM
> exits).
> It is more problematic for long-running embedding applications that can
> start/stop flows dynamically within their JVM uptime (like Kafka Connect and
> NiFi Stateless connectors).
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
