[kafka] branch 2.5 updated: KAFKA-10158: Fix flaky testDescribeUnderReplicatedPartitionsWhenReassignmentIsInProgress (#9022)

[ij...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

ijuma pushed a commit to branch 2.5
in repository https://gitbox.apache.org/repos/asf/kafka.git


The following commit(s) were added to refs/heads/2.5 by this push:
     new e9b58cf7 KAFKA-10158: Fix flaky testDescribeUnderReplicatedPartitionsWhenReassignmentIsInProgress (#9022)
e9b58cf7 is described below

commit e9b58cf7e936be8011ccef6015be7f7399881394
Author: Brian Byrne <bb...@confluent.io>
AuthorDate: Sun Jul 26 08:28:47 2020 -0700

    KAFKA-10158: Fix flaky testDescribeUnderReplicatedPartitionsWhenReassignmentIsInProgress (#9022)
    
    Set `replica.fetch.max.bytes` to `1` and produce multiple record batches to allow
    for throttling to take place. This helps avoid a race condition where the
    reassignment would complete more quickly than expected causing an
    assertion to fail some times.
    
    Reviewers: Lucas Bradstreet <lu...@confluent.io>, Jason Gustafson <ja...@confluent.io>, Chia-Ping Tsai <ch...@gmail.com>, Ismael Juma <is...@juma.me.uk>
---
 .../kafka/admin/TopicCommandWithAdminClientTest.scala | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/core/src/test/scala/unit/kafka/admin/TopicCommandWithAdminClientTest.scala b/core/src/test/scala/unit/kafka/admin/TopicCommandWithAdminClientTest.scala
index 87750f3..4b25af7 100644
--- a/core/src/test/scala/unit/kafka/admin/TopicCommandWithAdminClientTest.scala
+++ b/core/src/test/scala/unit/kafka/admin/TopicCommandWithAdminClientTest.scala
@@ -49,14 +49,20 @@ class TopicCommandWithAdminClientTest extends KafkaServerTestHarness with Loggin
   /**
     * Implementations must override this method to return a set of KafkaConfigs. This method will be invoked for every
     * test and should not reuse previous configurations unless they select their ports randomly when servers are started.
+    *
+    * Note the replica fetch max bytes is set to `1` in order to throttle the rate of replication for test
+    * `testDescribeUnderReplicatedPartitionsWhenReassignmentIsInProgress`.
     */
   override def generateConfigs: Seq[KafkaConfig] = TestUtils.createBrokerConfigs(
     numConfigs = 6,
     zkConnect = zkConnect,
     rackInfo = Map(0 -> "rack1", 1 -> "rack2", 2 -> "rack2", 3 -> "rack1", 4 -> "rack3", 5 -> "rack3"),
     numPartitions = numPartitions,
-    defaultReplicationFactor = defaultReplicationFactor
-  ).map(KafkaConfig.fromProps)
+    defaultReplicationFactor = defaultReplicationFactor,
+  ).map { props =>
+    props.put(KafkaConfig.ReplicaFetchMaxBytesProp, "1")
+    KafkaConfig.fromProps(props)
+  }
 
   private val numPartitions = 1
   private val defaultReplicationFactor = 1.toShort
@@ -655,8 +661,13 @@ class TopicCommandWithAdminClientTest extends KafkaServerTestHarness with Loggin
     adminClient.createTopics(
       Collections.singletonList(new NewTopic(testTopicName, partitions, replicationFactor).configs(configMap))).all().get()
     waitForTopicCreated(testTopicName)
+
+    // Produce multiple batches.
+    TestUtils.generateAndProduceMessages(servers, testTopicName, numMessages = 10, acks = -1)
     TestUtils.generateAndProduceMessages(servers, testTopicName, numMessages = 10, acks = -1)
 
+    // Enable throttling. Note the broker config sets the replica max fetch bytes to `1` upon to minimize replication
+    // throughput so the reassignment doesn't complete quickly.
     val brokerIds = servers.map(_.config.brokerId)
     TestUtils.setReplicationThrottleForPartitions(adminClient, brokerIds, Set(tp), throttleBytes = 1)
 
@@ -686,6 +697,10 @@ class TopicCommandWithAdminClientTest extends KafkaServerTestHarness with Loggin
       topicService.describeTopic(new TopicCommandOptions(Array("--under-replicated-partitions"))))
     assertEquals("--under-replicated-partitions shouldn't return anything", "", underReplicatedOutput)
 
+    // Verify reassignment is still ongoing.
+    val reassignments = adminClient.listPartitionReassignments(Collections.singleton(tp)).reassignments.get().get(tp)
+    assertFalse(Option(reassignments).forall(_.addingReplicas.isEmpty))
+
     TestUtils.removeReplicationThrottleForPartitions(adminClient, brokerIds, Set(tp))
     TestUtils.waitForAllReassignmentsToComplete(adminClient)
   }