[ https://issues.apache.org/jira/browse/STORM-1455?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15106930#comment-15106930 ]
ASF GitHub Bot commented on STORM-1455: --------------------------------------- GitHub user abhishekagarwal87 opened a pull request: https://github.com/apache/storm/pull/1026 STORM-1455: Do not reset the emittedOffset for offsetOutOfRangeExceptions There are two changes in this PR - If there is a TopicOffsetOutOfRangeException, it can be due to an old failed tuple. We should not reset the state back to the beginning. - Apart from the last completed offset, also emit the last emitted offset in the metric. latestEmittedOffset now points to the last emitted offset. You can merge this pull request into a Git repository by running: $ git pull https://github.com/abhishekagarwal87/storm kafka-spout Alternatively you can review and apply these changes as the patch at: https://github.com/apache/storm/pull/1026.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #1026 ---- commit c1e8ec576a391c9431ec50eb8eb60c083417b4c7 Author: Abhishek Agarwal <abhishek.agar...@inmobi.com> Date: 2016-01-19T16:13:22Z STORM-1455: Do not reset the emittedOffset for offsetOutOfRangeExceptions ---- > kafka spout should not reset to the beginning of partition when > offsetoutofrange exception occurs > ------------------------------------------------------------------------------------------------- > > Key: STORM-1455 > URL: https://issues.apache.org/jira/browse/STORM-1455 > Project: Apache Storm > Issue Type: Bug > Components: storm-kafka > Affects Versions: 1.0.0 > Reporter: Abhishek Agarwal > Assignee: Abhishek Agarwal > > https://github.com/apache/storm/blob/master/external/storm-kafka/src/jvm/storm/kafka/PartitionManager.java#L190 > {noformat} > try { > msgs = KafkaUtils.fetchMessages(_spoutConfig, _consumer, > _partition, offset); > } catch (TopicOffsetOutOfRangeException e) { > _emittedToOffset = KafkaUtils.getOffset(_consumer, > _partition.topic, _partition.partition, > kafka.api.OffsetRequest.EarliestTime()); > LOG.warn("{} Using new offset: {}", _partition.partition, > _emittedToOffset); > {noformat} > If there was one old offset out of range, partition manager will re-send all > the offsets from EarliestTime to _emittedOffset. -- This message was sent by Atlassian JIRA (v6.3.4#6332)