attilapiros commented on a change in pull request #24283: [SPARK-27355][SS]
Make query execution more sensitive to epoch message late or lost
URL: https://github.com/apache/spark/pull/24283#discussion_r272159472
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/continuous/EpochCoordinator.scala
##########
@@ -224,24 +227,24 @@ private[continuous] class EpochCoordinator(
partitionOffsets.collect { case ((e, _), o) if e == epoch => o }
if (thisEpochOffsets.size == numReaderPartitions) {
logDebug(s"Epoch $epoch has offsets reported from all partitions:
$thisEpochOffsets")
- query.addOffset(epoch, stream, thisEpochOffsets.toSeq)
+ queryExecution.addOffset(epoch, stream, thisEpochOffsets.toSeq)
resolveCommitsAtEpoch(epoch)
}
checkProcessingQueueBoundaries()
}
private def checkProcessingQueueBoundaries() = {
- if (partitionOffsets.size > epochBacklogQueueSize) {
- query.stopInNewThread(new IllegalStateException("Size of the partition
offset queue has " +
- "exceeded its maximum"))
+ if (partitionOffsets.size > epochMsgBacklogQueueSize) {
Review comment:
As I see you are reinterpret the new config
`CONTINUOUS_STREAMING_EPOCH_MSG_BACKLOG_QUEUE_SIZE` in place of the old
`CONTINUOUS_STREAMING_EPOCH_BACKLOG_QUEUE_SIZE `.
I think this is can be done (if good reasons supports it) as the old config
was only introduced into Spark 3.0.0-SNAPSHOT in the PR
https://github.com/apache/spark/pull/23156.
----------------------------------------------------------------
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]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
