mmmmxa commented on code in PR #4219:
URL: https://github.com/apache/iggy/pull/4219#discussion_r4080677227
##########
core/connectors/sinks/delta_sink/src/sink.rs:
##########
@@ -122,9 +158,6 @@ impl Sink for DeltaSink {
return Ok(());
}
- // TODO: all partition consume() calls serialize on this single lock,
holding it
- // through flush_and_commit() I/O. fix: per-partition writers keyed by
partition_id.
- // Ref:
https://github.com/apache/iggy/pull/2889/#discussion_r2936719763
let mut state_guard = self.state.lock().await;
Review Comment:
The TODO's premise was researched and the per-partition-writer fix was
deliberately not pursued — see the writeup on #3839: a single consumer task per
topic already serializes partition data before it reaches `consume()`,
cross-topic contention on one sink is rare, and Delta's own guidance to write
large commits means `flush_and_commit()` runs infrequently regardless. Deleting
the TODO reflects that decision rather than dodging it. Kept `Closes #3839` in
the PR body and added a short pointer comment at the lock site so future
readers land on that reasoning instead of a stale action item.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
