bowenli86 commented on code in PR #223:
URL:
https://github.com/apache/flink-connector-kafka/pull/223#discussion_r2789766372
##########
flink-connector-kafka/src/main/java/org/apache/flink/streaming/connectors/kafka/table/DynamicKafkaTableFactory.java:
##########
@@ -216,6 +217,20 @@ private static void applyDynamicDiscoveryOptions(
.STREAM_METADATA_DISCOVERY_FAILURE_THRESHOLD
.key(),
Integer.toString(value)));
+ tableOptions
+ .getOptional(IDLE_WATERMARK_DELAY_MS)
Review Comment:
IDLE_WATERMARK_DELAY_MS is out of this ticket scope, can you revert it and
create a new ticket/pr?
##########
docs/content/docs/connectors/datastream/dynamic-kafka.md:
##########
@@ -143,6 +144,33 @@ DynamicKafkaSource<String> source =
{{< /tab >}}
{{< /tabs >}}
+### Split Assignment Mode
+
+Dynamic Kafka Source supports two split-assignment modes:
+
+* `per_cluster` (default): assigns splits within each Kafka cluster
independently.
+* `global`: assigns splits across all discovered clusters using one global
balancing strategy.
Review Comment:
can you add doc on how per_cluster and global mode determines next owner of
new split, e.g. global mode, owner = knownActiveSplitIds.size() % numReaders,
and how to rebalance in global mode if it become skewed?
--
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]
