efeg commented on a change in pull request #8579:
URL: https://github.com/apache/kafka/pull/8579#discussion_r419833344
##########
File path: core/src/main/scala/kafka/server/AbstractFetcherThread.scala
##########
@@ -382,6 +382,11 @@ abstract class AbstractFetcherThread(name: String,
"that the partition is being moved")
partitionsWithError += topicPartition
+ case Errors.UNKNOWN_TOPIC_OR_PARTITION =>
+ warn(s"Remote broker does not host the partition
$topicPartition, which could indicate " +
+ "that the partition is being created or deleted.")
Review comment:
@ijuma Thanks for your reply!
We both agree that relying on warnings in logs to _detect_ propagation
issues is not practical -- metrics indeed should be used in "detection" of such
issues.
I believe the focus of this discussion isthe _investigation_ of detected
issues.
Metrics typically fail to provide the level of details to examine the
interaction of various events in a distributed environment. Hence, in case we
need to look at the logs for figuring out such cases, I claim that being able
to grep warns and see undesired events is useful -- as opposed to trying to
browse info logs for bad/undesirable events.
Note that in `trunk`, such `UNKNOWN_TOPIC_OR_PARTITION` events are already
printed in a stronger level (i.e. `error`).
----------------------------------------------------------------
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:
us...@infra.apache.org
