[
https://issues.apache.org/jira/browse/NIFI-3792?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16654108#comment-16654108
]
ASF GitHub Bot commented on NIFI-3792:
--------------------------------------
Github user JPercivall commented on the issue:
https://github.com/apache/nifi/pull/3090
Hey @patricker, here's the processor I mentioned. It attempts to avert the
livelock scenario by looking at the input queue if it's over a certain
size/count divert FFs but there's still the chance for livelock in some
scenarios.
> A processor to facilitate retrying FlowFiles
> --------------------------------------------
>
> Key: NIFI-3792
> URL: https://issues.apache.org/jira/browse/NIFI-3792
> Project: Apache NiFi
> Issue Type: New Feature
> Reporter: Joseph Percivall
> Assignee: Joseph Percivall
> Priority: Major
>
> When dealing with processor failures in production, potentially related to
> network issues, the current best practice is to retry the flowfile a couple
> times before declaring it a failure[1]. This currently requires multiple
> processors and penalizing the flowfile isn't possible. Also if the flow is
> not fast enough, back-pressure can cause a livelocked state which requires
> manual intervention.
> A new processor should be added to not only encourage best practices but also
> offer configuration options to deal with un-optimal situations.
> [1]
> https://community.hortonworks.com/questions/77336/nifi-best-practices-for-error-handling.html
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
