showuon commented on code in PR #15305: URL: https://github.com/apache/kafka/pull/15305#discussion_r1570111991
########## connect/runtime/src/test/java/org/apache/kafka/connect/runtime/distributed/WorkerCoordinatorTest.java: ########## @@ -533,6 +537,45 @@ public void testSkippingAssignmentFails() { verify(configStorage).snapshot(); } + @Test + public void testPollTimeoutExpiry() throws InterruptedException { + + when(configStorage.snapshot()).thenReturn(configState1); + + client.prepareResponse(FindCoordinatorResponse.prepareResponse(Errors.NONE, groupId, node)); + coordinator.ensureCoordinatorReady(time.timer(Long.MAX_VALUE)); + + client.prepareResponse(joinGroupFollowerResponse(1, "member", "leader", Errors.NONE)); + client.prepareResponse(syncGroupResponse(ConnectProtocol.Assignment.NO_ERROR, "leader", configState1.offset(), Collections.emptyList(), + Collections.singletonList(taskId1x0), Errors.NONE)); + + client.prepareResponse(FindCoordinatorResponse.prepareResponse(Errors.NONE, groupId, node)); + client.prepareResponse(FindCoordinatorResponse.prepareResponse(Errors.NONE, groupId, node)); + + // prepare 3 heartBeatResponses because we will trigger 3 heartBeat requests until rebalanceTimeout, + // that is (sessionTimeoutMs - 1) * 3 > rebalanceTimeoutMs + client.prepareResponse(new HeartbeatResponse(new HeartbeatResponseData())); + client.prepareResponse(new HeartbeatResponse(new HeartbeatResponseData())); + client.prepareResponse(new HeartbeatResponse(new HeartbeatResponseData())); + + try (LogCaptureAppender logCaptureAppender = LogCaptureAppender.createAndRegister(WorkerCoordinator.class)) { + coordinator.ensureActiveGroup(); + coordinator.poll(0, () -> { + return null; + }); + + // We keep the heartbeat thread running behind the scenes and poll frequently so that eventually + // the time goes past now + rebalanceTimeoutMs which triggers poll timeout expiry. + TestUtils.waitForCondition(() -> { + // sleep until sessionTimeoutMs to trigger a heartBeat request to avoid session timeout. + // Not sure if this will be flaky in CI because the heartbeat thread might not send out the heartBeat request in time. Review Comment: Should we remove this line? ########## connect/runtime/src/test/java/org/apache/kafka/connect/runtime/distributed/WorkerCoordinatorTest.java: ########## @@ -533,6 +537,45 @@ public void testSkippingAssignmentFails() { verify(configStorage).snapshot(); } + @Test + public void testPollTimeoutExpiry() throws InterruptedException { + + when(configStorage.snapshot()).thenReturn(configState1); + + client.prepareResponse(FindCoordinatorResponse.prepareResponse(Errors.NONE, groupId, node)); + coordinator.ensureCoordinatorReady(time.timer(Long.MAX_VALUE)); + + client.prepareResponse(joinGroupFollowerResponse(1, "member", "leader", Errors.NONE)); + client.prepareResponse(syncGroupResponse(ConnectProtocol.Assignment.NO_ERROR, "leader", configState1.offset(), Collections.emptyList(), + Collections.singletonList(taskId1x0), Errors.NONE)); + + client.prepareResponse(FindCoordinatorResponse.prepareResponse(Errors.NONE, groupId, node)); + client.prepareResponse(FindCoordinatorResponse.prepareResponse(Errors.NONE, groupId, node)); + Review Comment: Hmm.. I see. The reason is because we don't make sure the heartbeat is sent after `time.sleep(sessionTimeoutMs - 1);`. Do you think we could make sure the heartbeat request or response is sent before next sleep? -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org