Github user srdo commented on the issue:
https://github.com/apache/storm/pull/1919
There are two different settings influencing retry behavior here. Auto
commit on the spout, and number of ackers configured for Storm
(https://storm.apache.org/releases/1.0.0/javadocs/org/apache/storm/Config.html#TOPOLOGY_ACKER_EXECUTORS).
You can configure Storm to require tuple acking, while also using auto commit
for this spout, in which case the spout will retry tuples that fail, but only
on a best-effort basis (i.e. if the spout crashes, tuples will not be
replayed). In that case, the spout actually needs to track tuples even though
auto commit is on. You can also set acker executors to 0, which will cause
Storm to just ack tuples immediately when emitted. In that case it doesn't hurt
anything if the spout adds tuples to emitted, because they'll just get removed
again when ack() is called.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
