[
https://issues.apache.org/jira/browse/SAMZA-62?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Riccomini updated SAMZA-62:
---------------------------------
Attachment: SAMZA-62.0.patch
Attaching a fix.
https://reviews.apache.org/r/14878/
Testing done on a real Kafka cluster, which triggered a replication count > num
brokers exception when trying to create the checkpoint topic. We can now see
such exceptions in the checkpoint manager.
> KafkaCheckpointManager fails when offset request has error code
> ---------------------------------------------------------------
>
> Key: SAMZA-62
> URL: https://issues.apache.org/jira/browse/SAMZA-62
> Project: Samza
> Issue Type: Bug
> Components: kafka
> Affects Versions: 0.6.0
> Reporter: Chris Riccomini
> Assignee: Chris Riccomini
> Fix For: 0.7.0
>
> Attachments: SAMZA-62.0.patch
>
>
> When the KafkaCheckpointManager makes its offset request to the broker for
> each checkpoint topic, it doesn't check the offset response's error code.
> Consequently, the error is swallowed. This makes it very hard to debug when
> something goes wrong.
> We should just do an ErrorMapping.maybeThrow check on the error code, which
> will trigger a reconnect when things go wrong.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
