hachikuji commented on a change in pull request #9022: URL: https://github.com/apache/kafka/pull/9022#discussion_r455401060
########## File path: core/src/test/scala/unit/kafka/admin/TopicCommandWithAdminClientTest.scala ########## @@ -56,12 +56,17 @@ class TopicCommandWithAdminClientTest extends KafkaServerTestHarness with Loggin zkConnect = zkConnect, rackInfo = Map(0 -> "rack1", 1 -> "rack2", 2 -> "rack2", 3 -> "rack1", 4 -> "rack3", 5 -> "rack3"), numPartitions = numPartitions, - defaultReplicationFactor = defaultReplicationFactor + defaultReplicationFactor = defaultReplicationFactor, + replicaFetchMaxBytes = replicaFetchMaxBytes(), ).map(KafkaConfig.fromProps) private val numPartitions = 1 private val defaultReplicationFactor = 1.toShort + private def replicaFetchMaxBytes() = + if (testName.getMethodName == "testDescribeUnderReplicatedPartitionsWhenReassignmentIsInProgress") Some(1) Review comment: It's always an annoyance with Junit that there is no way for the test case to override initialization in a `@Before`. What we often end up doing is removing the annotation and calling the setup method explicitly in each test case. I slightly prefer that option since it is easier to understand, but not sure it is possible since we're extending `KafkaServerTestHarness`, which has its own initialization logic. For the sake of argument, would it be possible to set max fetch bytes to 1 for all tests? Either that or maybe we should just produce more data in the test case. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org