[
https://issues.apache.org/jira/browse/NIFI-3900?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16013022#comment-16013022
]
ASF subversion and git services commented on NIFI-3900:
-------------------------------------------------------
Commit 494a0e89288c5e51f9467b719e7ee582cd74c875 in nifi's branch
refs/heads/master from [~markap14]
[ https://git-wip-us.apache.org/repos/asf?p=nifi.git;h=494a0e8 ]
NIFI-3900: Ensure that when we serialize a flow to send to Cluster Coordinator
that we include the Scheduled State for processors as they are intended to be,
not as they are currently because during startup they may not have been started
yet.
NIFI-3900: Rebased against master and updated new unit test to use new method
signature for FlowSerializer.serialize
This closes #1804.
Signed-off-by: Bryan Bende <[email protected]>
> Cluster - Inconsistent schedule state across the cluster
> --------------------------------------------------------
>
> Key: NIFI-3900
> URL: https://issues.apache.org/jira/browse/NIFI-3900
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework
> Reporter: Matt Gilman
> Assignee: Mark Payne
> Priority: Critical
> Fix For: 1.3.0
>
>
> When nodes join a cluster there appears to be a timing issue that causes
> inconsistency with the scheduled state of components across the cluster. Due
> to delayed initialization, component's schedule state may be incorrect when a
> proposed flow is sent to a node.
> It appears that a node will get an inconsistent scheduled states when it's
> not able to connect to the cluster initially and receives a try-again-later
> response from the coordinator.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
