[
https://issues.apache.org/jira/browse/STORM-1848?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15289186#comment-15289186
]
ASF GitHub Bot commented on STORM-1848:
---------------------------------------
Github user srdo commented on the pull request:
https://github.com/apache/storm/pull/1428#issuecomment-220071724
You're right, I forgot about backwards compatibility, I mainly wanted to
ensure this doesn't pop up in other spouts :) Is there any way for the Kafka
spout to register with Kryo instead?
How about a comment on the SpoutOutputCollector interface about this?
> NotSerializableException when using storm-kafka spout with event logging
> ------------------------------------------------------------------------
>
> Key: STORM-1848
> URL: https://issues.apache.org/jira/browse/STORM-1848
> Project: Apache Storm
> Issue Type: Bug
> Affects Versions: 2.0.0, 1.0.1
> Reporter: Stig Rohde Døssing
> Assignee: Stig Rohde Døssing
> Priority: Minor
>
> We deployed a topology to multiple workers and set
> topology.eventlogger.executors to be identical to the number of workers. When
> the debug button in Storm UI is pressed, the spout will start throwing
> NotSerializableExceptions.
> java.lang.RuntimeException: java.lang.RuntimeException:
> java.io.NotSerializableException:
> org.apache.storm.kafka.PartitionManager$KafkaMessageId at
> org.apache.storm.utils.DisruptorQueue.consumeBatchToCursor(DisruptorQueue.java:452)
> at (snip)
> KafkaMessageId should be made serializable.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
