kfaraz commented on code in PR #19091:
URL: https://github.com/apache/druid/pull/19091#discussion_r2895398461
##########
indexing-service/src/main/java/org/apache/druid/indexing/seekablestream/supervisor/SeekableStreamSupervisor.java:
##########
@@ -2669,18 +2677,50 @@ private void verifyAndMergeCheckpoints(
sequenceCheckpoint -> {
killTask(
sequenceCheckpoint.lhs,
- "Killing task[%s], as its checkpoints[%s] are not consistent
with group checkpoints[%s]"
- + " or latest persisted sequences in metadata store[%s].",
- sequenceCheckpoint.lhs,
- sequenceCheckpoint.rhs,
- taskGroup.checkpointSequences,
- latestOffsetsFromDb
+ "Killing task as its checkpoints are not consistent with group
checkpoints"
+ + " or latest persisted sequences in metadata store."
);
taskGroup.removeTask(sequenceCheckpoint.lhs);
}
);
}
+ /**
+ * Checks if there is another {@link TaskGroup} publishing to any of the
partitions
+ * that are being read by the given {@param taskGroup}. If this method
returns
+ * true, it indicates that the current taskGroup would need to wait for the
+ * older taskGroups to finish publishing before it can publish its own
offsets.
+ */
+ private boolean isAnotherTaskGroupPublishingToPartitionsOf(TaskGroup
taskGroup)
+ {
+ final Set<PartitionIdType> partitionsPendingPublishFromOtherGroups =
+ pendingCompletionTaskGroups
+ .values()
+ .stream()
+ .flatMap(Collection::stream)
+ .filter(group -> !group.equals(taskGroup))
Review Comment:
I had to update the equality check since method signature has also changed.
I am now simply checking the set of taskIds in the `TaskGroup` against the
target `taskId`.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
