[
https://issues.apache.org/jira/browse/NIFI-5849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16702922#comment-16702922
]
ASF subversion and git services commented on NIFI-5849:
-------------------------------------------------------
Commit 30f2f4205121113c26bb00ac5a8697dffaeb8206 in nifi's branch
refs/heads/master from [~ijokarumawak]
[ https://git-wip-us.apache.org/repos/asf?p=nifi.git;h=30f2f42 ]
NIFI-5849: ListXXX can lose cluster state on processor restart
NIFI-5406 introduced the issue by trying to use the resetState variable for
different purposes. AbstractListProcessor should have had a different variable
to control whether to clear state for tracking entity strategy.
Signed-off-by: Pierre Villard <[email protected]>
This closes #3189.
> ListSFTP, ListFTP processors lose Cluster state on restart.
> -----------------------------------------------------------
>
> Key: NIFI-5849
> URL: https://issues.apache.org/jira/browse/NIFI-5849
> Project: Apache NiFi
> Issue Type: Bug
> Affects Versions: 1.8.0
> Reporter: Truong Duc Kien
> Assignee: Koji Kawamura
> Priority: Critical
>
> In a cluster environment, restarting a processor that extends
> {{AbstractListProcessor}} could lead to the erasure of that processor's
> existing cluster states.
> This bug was introduced in commit
> [https://github.com/apache/nifi/commit/8b9d4461185848fd552a639ac14b7926164d5426]
> .
> Before that commit, the {{resetState}} variable is changed to false after the
> state is clear, inside {{updateState}} method. However, the commit moved this
> line to inside the {{onTrigger}} method. This is problematic, because
> {{onTrigger}} method is only called on the primary node. As a result, the
> {{resetState}} variable is only reset to false on primary node, but not on
> the other nodes. When the user restarts the processor, every nodes will
> re-execute the {{updateState}} method and the Cluster state will be cleared.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
