[
https://issues.apache.org/jira/browse/FLUME-984?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13213156#comment-13213156
]
Juhani Connolly commented on FLUME-984:
---------------------------------------
I'm not sure if just logging all exceptions and leaving the sink runner(and
repeatedly failing for certain failure patterns) is the best choice.
There should definitely be some kind of unified behavior to all sinks regarding
when and how they fail, along with how they should be handled.
Perhaps we should add an exception that sinks can throw to indicate they are
unrecoverable?
> SinkRunner should catch unhanded exceptions and log them like
> PollingSourceRunner
> ---------------------------------------------------------------------------------
>
> Key: FLUME-984
> URL: https://issues.apache.org/jira/browse/FLUME-984
> Project: Flume
> Issue Type: Improvement
> Components: Sinks+Sources
> Affects Versions: v1.0.0
> Reporter: Brock Noland
> Assignee: Brock Noland
> Attachments: FLUME-984-0.patch
>
>
> If a Sink throws any exception but EventDeliveryException the sink will die.
> We should catch the exception, perhaps backoff a bit, and retry like
> PollingSourceRunner.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
