squah-confluent commented on code in PR #18046:
URL: https://github.com/apache/kafka/pull/18046#discussion_r1871373379
##########
group-coordinator/src/main/java/org/apache/kafka/coordinator/group/GroupMetadataManager.java:
##########
@@ -1053,6 +1055,12 @@ ConsumerGroup convertToConsumerGroup(ClassicGroup
classicGroup, List<Coordinator
throw new GroupIdNotFoundException("Cannot upgrade the classic
group " + classicGroup.groupId() +
" to consumer group because the embedded consumer protocol is
malformed.");
+ } catch (UnsupportedVersionException e) {
+ log.warn("Cannot upgrade the classic group " +
classicGroup.groupId() +
+ " to consumer group: " + e.getMessage() + ".", e);
+
+ throw new GroupIdNotFoundException("Cannot upgrade the classic
group " + classicGroup.groupId() +
+ " to consumer group because a custom assignor is in use.");
Review Comment:
> Completely unrelated but I wonder if we always provide such good error
message when the upgrade is not possible. Do you mind checking?
I had a look and there are three reasons in
[`validateOnlineUpgrade`](https://github.com/apache/kafka/blob/e99561e1f30a30f9b151669a78d340f286a3ead6/group-coordinator/src/main/java/org/apache/kafka/coordinator/group/GroupMetadataManager.java#L1008-L1023)
that we convert to "Group %s is not a consumer group":
* Cannot upgrade classic group {} to consumer group because the online
upgrade is disabled.
* Cannot upgrade classic group {} to consumer group because the group does
not use the consumer embedded protocol.
* Cannot upgrade classic group {} to consumer group because the group size
exceeds the consumer group maximum size.
Apart from that and SchemaException when subscriptions or metadata are
malformed, I can't see any other error paths directly related to the upgrade.
--
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]
