jihoonson commented on a change in pull request #10278:
URL: https://github.com/apache/druid/pull/10278#discussion_r470879995
##########
File path:
indexing-service/src/main/java/org/apache/druid/indexing/common/task/batch/parallel/ParallelIndexSupervisorTask.java
##########
@@ -788,18 +785,7 @@ private static void publishSegments(TaskToolbox toolbox,
Map<String, PushedSegme
if (published) {
LOG.info("Published [%d] segments", newSegments.size());
} else {
- LOG.info("Transaction failure while publishing segments, checking if
someone else beat us to it.");
Review comment:
Oh, I think the logging here was wrong. In parallel indexing, unlike in
streaming ingestion, there is no concept of replicas which could publish the
same set of segments. So, if it failed to publish segments, the task should
just fail without checking if those segments were already published by another
task since it cannot happen.
> "Our segments really do exist, awaiting handoff."
This log is also misleading since batch tasks never waits for segment
handoff.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
