Github user tpiscitell commented on the pull request:
https://github.com/apache/storm/pull/405#issuecomment-94448772
@miguno I don't think either of those JIRAs fixed this issue. The problem
here is that PartitionManager.fill() will always attempt to fetch any failed
tuples first:
```
// Are there failed tuples? If so, fetch those first.
if (had_failed) {
offset = failed.first();
} else {
offset = _emittedToOffset;
}
```
However, even with the patches for those two JIRAs, the `failed` list is
never pruned. Instead PartitionManager.fill() just update `_emitedToOffset` and
returns:
```
} catch (TopicOffsetOutOfRangeException e) {
_emittedToOffset = KafkaUtils.getOffset(_consumer,
_spoutConfig.topic, _partition.partition, _spoutConfig);
LOG.warn("Using new offset: {}", _emittedToOffset);
// fetch failed, so don't update the metrics
return;
}
```
Just to be called again, and the process repeats:
```
public EmitState next(SpoutOutputCollector collector) {
if (_waitingToEmit.isEmpty()) {
fill();
}
```
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
