[
https://issues.apache.org/jira/browse/STORM-2544?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Prasanna Ranganathan updated STORM-2544:
----------------------------------------
Description:
Situation: Spout configured to use manual commit with a finite number of
retries.
In the above scenario if and when a tuple fails repeatedly and hits the retry
limit, it will neither be scheduled for an attempt again nor properly accounted
for in the ack() method and in OffsetManager. This will block commits for the
partition that the tuple belongs to.
was:
Situation: Spout configured to use manual commit with a finite number of
retries.
In the above scenario if and when a tuple fails repeatedly and hits the retry
limit, it will neither be schedule for an attempt again nor properly accounted
for in the ack() method and in OffsetManager. This will block commits for the
partition that the tuple belongs to.
> Bugs in the Kafka Spout retry logic when using manual commit
> ------------------------------------------------------------
>
> Key: STORM-2544
> URL: https://issues.apache.org/jira/browse/STORM-2544
> Project: Apache Storm
> Issue Type: Bug
> Components: storm-kafka-client
> Affects Versions: 1.x
> Reporter: Prasanna Ranganathan
>
> Situation: Spout configured to use manual commit with a finite number of
> retries.
> In the above scenario if and when a tuple fails repeatedly and hits the retry
> limit, it will neither be scheduled for an attempt again nor properly
> accounted for in the ack() method and in OffsetManager. This will block
> commits for the partition that the tuple belongs to.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
