frankvicky commented on code in PR #20108:
URL: https://github.com/apache/kafka/pull/20108#discussion_r2191333602
##########
test-common/test-common-runtime/src/main/java/org/apache/kafka/common/test/ClusterInstance.java:
##########
@@ -258,7 +257,63 @@ default Set<GroupProtocol> supportedGroupProtocols() {
//---------------------------[wait]---------------------------//
default void waitTopicDeletion(String topic) throws InterruptedException {
- waitForTopic(topic, 0);
+ Collection<KafkaBroker> brokers = aliveBrokers().values();
+ // wait for metadata
+ TestUtils.waitForCondition(
+ () -> brokers.stream().allMatch(
+ broker ->
broker.metadataCache().numPartitions(topic).isEmpty()),
+ 60000L, topic + " metadata not propagated after 60000 ms");
+
+ for (ControllerServer controller : controllers().values()) {
+ long controllerOffset =
controller.raftManager().replicatedLog().endOffset().offset() - 1;
+ TestUtils.waitForCondition(
+ () -> brokers.stream().allMatch(broker -> ((BrokerServer)
broker).sharedServer().loader().lastAppliedOffset() >= controllerOffset),
+ 60000L, "Timeout waiting for controller metadata propagating
to brokers");
+ }
+
+ List<TopicPartition> topicPartitions = List.of(new
TopicPartition(topic, 0));
Review Comment:
Yes, we don't need it. It's odd that the original logic uses intStream to
create a list, but we could clean up in this PR.
--
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]
