Github user hmcl commented on a diff in the pull request:
https://github.com/apache/storm/pull/1433#discussion_r63924300
--- Diff: docs/State-checkpointing.md ---
@@ -94,6 +94,9 @@ is saved and then the checkpoint tuple is forwarded to
the next component. Each
streams before it saves its state so that the state represents a
consistent state across the topology. Once the checkpoint spout receives
ACK from all the bolts, the state commit is complete and the transaction
is recorded as committed by the checkpoint spout.
+The state checkpointing does not currently checkpoint the state of the
spout. Yet, once the states of all bolts are checkpointed and once the
checkpoint tuples is acked, the tuples emitted by the spout are acked.
+It also implies that `topology.state.checkpoint.interval.ms` is lower than
`topology.message.timeout.secs`.
--- End diff --
once the *STATE* of all the bolts are checkpointed, and once the checkpoint
tuples *ARE* acked.... by the spout are *ALSO* acked.
---
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.
---
