[
https://issues.apache.org/jira/browse/NIFI-9729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17498279#comment-17498279
]
ASF subversion and git services commented on NIFI-9729:
-------------------------------------------------------
Commit 6cea5ea520f48c5d05a30df93fa582cf42d3438e in nifi's branch
refs/heads/main from Mark Payne
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=6cea5ea ]
NIFI-9729: When restarting components in the VersionedFlowSynchronizer, first
filter out any components that are intended to be stopped.
Signed-off-by: Joe Gresock <[email protected]>
This closes #5806.
> If a component is stopped while a node is disconnected from cluster, node may
> not be updated when rejoining
> -----------------------------------------------------------------------------------------------------------
>
> Key: NIFI-9729
> URL: https://issues.apache.org/jira/browse/NIFI-9729
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework
> Reporter: Mark Payne
> Assignee: Mark Payne
> Priority: Blocker
> Fix For: 1.16.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> This issue does not affect any already-released version of NiFi. It affects
> only 1.16.0-SNAPSHOT.
> To replicate:
> * Have a running NiFi cluster
> * Start components on the canvas
> * Disconnect a node
> * Stop one of the components
> * Reconnect the node
> * Once the node is reconnected, the component should be stopped (or
> stopping) but currently will still be running.
> The opposite does not appear to be a concern. I.e., if a component was
> stopped, and changed to running while the node is disconnected, the node will
> properly update the component state and start running the component upon
> rejoining the cluster.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
