[
https://issues.apache.org/jira/browse/NIFI-4602?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16272862#comment-16272862
]
ASF subversion and git services commented on NIFI-4602:
-------------------------------------------------------
Commit 45df23b1e09f340fb2c7f3c3234462570ddb5f6b in nifi's branch
refs/heads/master from [~markap14]
[ https://git-wip-us.apache.org/repos/asf?p=nifi.git;h=45df23b ]
NIFI-4607, NIFI-3975, NIFI-4602, NIFI-4606: This closes #2272. Fixed bug in
TailFile that caused new Primary Node to not pull current Clustered State when
File Location was set to Remote. Fixed bug that caused TailFile to occasionally
become 'stuck' when the file it is tailing is renamed and a new file is created
with the same name. Removed the 'Rolling Strategy' property from TailFile
because it is not actually used in the processor anymore. Deleted
MonitorMemoryTest because the unit test was testing the behavior of
FlowController more than the behavior of the reporting task itself and in order
to do so had a dependency in the pom.xml on nifi-framework-core, which means
that it no longer compiles when FlowController is modified.
Signed-off-by: joewitt <[email protected]>
> TailFile does not recover cluster-wide state when Primary Node changes
> ----------------------------------------------------------------------
>
> Key: NIFI-4602
> URL: https://issues.apache.org/jira/browse/NIFI-4602
> Project: Apache NiFi
> Issue Type: Bug
> Components: Extensions
> Reporter: Mark Payne
> Assignee: Mark Payne
> Fix For: 1.5.0
>
>
> I created a TailFile processor and set the "File Location" property to
> "Remote." I then added some text to a file and started the TailFile processor
> (Primary Node Only).
> I disconnected the primary node from the cluster. A second node became the
> new Primary Node and continued tailing the data as expected. As new data was
> added, the node continued to emit FlowFiles.
> Then, I added Node 1 back to the cluster. Verified that Node 2 continued to
> poll data. I then disconnected Node 2 so that Node 1 once again became the
> Primary Node. At this point, Node 1 continued tailing from where it last left
> off, not where Node 2 left off.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
