cychiu8 commented on code in PR #22252: URL: https://github.com/apache/kafka/pull/22252#discussion_r3254362751
########## server/src/test/java/org/apache/kafka/server/requests/FetchRequestTest.java: ########## @@ -0,0 +1,738 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.kafka.server.requests; + +import kafka.server.KafkaBroker; + +import org.apache.kafka.clients.admin.Admin; +import org.apache.kafka.clients.admin.NewPartitionReassignment; +import org.apache.kafka.clients.admin.TopicDescription; +import org.apache.kafka.clients.consumer.Consumer; +import org.apache.kafka.clients.consumer.ConsumerConfig; +import org.apache.kafka.clients.consumer.RangeAssignor; +import org.apache.kafka.clients.producer.Producer; +import org.apache.kafka.clients.producer.ProducerRecord; +import org.apache.kafka.common.Node; +import org.apache.kafka.common.TopicIdPartition; +import org.apache.kafka.common.TopicPartition; +import org.apache.kafka.common.Uuid; +import org.apache.kafka.common.message.FetchRequestData; +import org.apache.kafka.common.message.FetchResponseData; +import org.apache.kafka.common.protocol.ApiKeys; +import org.apache.kafka.common.protocol.Errors; +import org.apache.kafka.common.requests.FetchMetadata; +import org.apache.kafka.common.requests.FetchRequest; +import org.apache.kafka.common.requests.FetchResponse; +import org.apache.kafka.common.test.ClusterInstance; +import org.apache.kafka.common.test.api.ClusterConfigProperty; +import org.apache.kafka.common.test.api.ClusterTest; +import org.apache.kafka.common.test.api.ClusterTestDefaults; +import org.apache.kafka.common.test.api.Type; +import org.apache.kafka.common.utils.Utils; +import org.apache.kafka.common.utils.annotation.ApiKeyVersionsSource; +import org.apache.kafka.server.IntegrationTestUtils; +import org.apache.kafka.server.TestUtils; + +import org.junit.jupiter.api.Test; +import org.junit.jupiter.params.ParameterizedTest; +import org.junit.jupiter.params.provider.Arguments; +import org.junit.jupiter.params.provider.MethodSource; + +import java.net.Socket; +import java.time.Duration; +import java.util.ArrayList; +import java.util.Collection; +import java.util.Collections; +import java.util.HashMap; +import java.util.HashSet; +import java.util.LinkedHashMap; +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.Optional; +import java.util.Set; +import java.util.concurrent.ExecutorService; +import java.util.concurrent.Executors; +import java.util.concurrent.Future; +import java.util.concurrent.TimeUnit; +import java.util.stream.Stream; + +import static org.apache.kafka.coordinator.group.GroupCoordinatorConfig.OFFSETS_TOPIC_REPLICATION_FACTOR_CONFIG; +import static org.apache.kafka.server.TestUtils.consumeRecords; +import static org.apache.kafka.server.config.ReplicationConfigs.REPLICA_SELECTOR_CLASS_CONFIG; +import static org.apache.kafka.server.config.ServerConfigs.BROKER_RACK_CONFIG; +import static org.apache.kafka.server.config.ServerConfigs.CONTROLLED_SHUTDOWN_ENABLE_CONFIG; +import static org.apache.kafka.server.config.ServerLogConfigs.NUM_PARTITIONS_CONFIG; +import static org.apache.kafka.test.TestUtils.DEFAULT_MAX_WAIT_MS; +import static org.apache.kafka.test.TestUtils.waitForCondition; +import static org.junit.jupiter.api.Assertions.assertEquals; +import static org.junit.jupiter.api.Assertions.assertNotEquals; +import static org.junit.jupiter.api.Assertions.assertThrows; + +@ClusterTestDefaults( + types = {Type.KRAFT}, + brokers = FetchRequestTest.BROKER_COUNT, + serverProperties = { + @ClusterConfigProperty(id = 0, key = BROKER_RACK_CONFIG, value = "0"), + @ClusterConfigProperty(id = 1, key = BROKER_RACK_CONFIG, value = "1"), + @ClusterConfigProperty(key = REPLICA_SELECTOR_CLASS_CONFIG, + value = "org.apache.kafka.common.replica.RackAwareReplicaSelector"), + @ClusterConfigProperty(key = CONTROLLED_SHUTDOWN_ENABLE_CONFIG, value = "false"), + @ClusterConfigProperty(key = NUM_PARTITIONS_CONFIG, value = "1"), + @ClusterConfigProperty(key = OFFSETS_TOPIC_REPLICATION_FACTOR_CONFIG, value = "2"), + } +) +public class FetchRequestTest { + + public static final int BROKER_COUNT = 2; + public static final String TOPIC = "test-fetch-from-follower"; + public static final int LEADER_BROKER_ID = 0; + public static final int FOLLOWER_BROKER_ID = 1; + + private ClusterInstance cluster; + + @ClusterTest + public void testFollowerCompleteDelayedFetchesOnReplication(ClusterInstance cluster) throws Exception { + this.cluster = cluster; + cluster.createTopicWithAssignment(TOPIC, Map.of(0, List.of(LEADER_BROKER_ID, FOLLOWER_BROKER_ID))); + int leaderId = waitUntilLeaderIsKnown(cluster.brokers().values(), new TopicPartition(TOPIC, 0)); + + var topicPartition = new TopicPartition(TOPIC, 0); + assertEquals(LEADER_BROKER_ID, leaderId); + + LinkedHashMap<TopicPartition, FetchRequest.PartitionData> fetchData = createPartitionMap(1000, List.of(topicPartition), Map.of(topicPartition, 0L)); + FetchRequest fetchRequest = FetchRequest.Builder.forConsumer(ApiKeys.FETCH.latestVersion(), 20000, 1, fetchData) + .setMaxBytes(1000) + .rackId("") + .build(); + + int followerPort = cluster.brokers().get(FOLLOWER_BROKER_ID).boundPort(cluster.clientListener()); + try (Socket socket = IntegrationTestUtils.connect(followerPort)) { + IntegrationTestUtils.sendRequest(socket, + Utils.toArray(fetchRequest.serializeWithHeader( + IntegrationTestUtils.nextRequestHeader(ApiKeys.FETCH, ApiKeys.FETCH.latestVersion())))); + + try (Producer<byte[], byte[]> producer = cluster.producer(Map.of())) { + producer.send(new ProducerRecord<>(TOPIC, "key".getBytes(), "value".getBytes())).get(); + } + + FetchResponse response = IntegrationTestUtils.receive(socket, ApiKeys.FETCH, ApiKeys.FETCH.latestVersion()); + assertEquals(Errors.NONE, response.error()); + assertEquals(Map.of(Errors.NONE, 2), response.errorCounts()); + } + } + + @ClusterTest + public void testFetchFromLeaderWhilePreferredReadReplicaIsUnavailable(ClusterInstance cluster) throws Exception { + this.cluster = cluster; + cluster.createTopicWithAssignment(TOPIC, Map.of(0, List.of(LEADER_BROKER_ID, FOLLOWER_BROKER_ID))); + waitUntilLeaderIsKnown(cluster.brokers().values(), new TopicPartition(TOPIC, 0)); + + produceMessages(10); + assertEquals(1, getPreferredReplica()); + + cluster.brokers().get(FOLLOWER_BROKER_ID).shutdown(); + TopicPartition topicPartition = new TopicPartition(TOPIC, 0); + + waitForCondition( + () -> { + KafkaBroker leaderBroker = cluster.brokers().get(LEADER_BROKER_ID); + Map<Integer, Node> endpoints = leaderBroker.metadataCache() + .getPartitionReplicaEndpoints(topicPartition, cluster.clientListener()); + return !endpoints.containsKey(FOLLOWER_BROKER_ID); + }, + "follower is still reachable." + ); + + assertEquals(-1, getPreferredReplica()); + } + + @ClusterTest + public void testFetchFromFollowerWithRoll(ClusterInstance cluster) throws Exception { + this.cluster = cluster; + cluster.createTopicWithAssignment(TOPIC, Map.of(0, List.of(LEADER_BROKER_ID, FOLLOWER_BROKER_ID))); + waitUntilLeaderIsKnown(cluster.brokers().values(), new TopicPartition(TOPIC, 0)); + + var followerConsumerProps = new HashMap<String, Object>(); + followerConsumerProps.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, cluster.bootstrapServers()); + followerConsumerProps.put(ConsumerConfig.GROUP_ID_CONFIG, "test-group-follower"); + followerConsumerProps.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest"); + followerConsumerProps.put(ConsumerConfig.CLIENT_RACK_CONFIG, String.valueOf(FOLLOWER_BROKER_ID)); + + var leaderConsumerProps = new HashMap<String, Object>(); + leaderConsumerProps.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, cluster.bootstrapServers()); + leaderConsumerProps.put(ConsumerConfig.GROUP_ID_CONFIG, "test-group-leader"); + leaderConsumerProps.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest"); + + try (Consumer<byte[], byte[]> followerConsumer = cluster.consumer(followerConsumerProps); + Consumer<byte[], byte[]> leaderConsumer = cluster.consumer(leaderConsumerProps)) { + followerConsumer.subscribe(List.of(TOPIC)); + leaderConsumer.subscribe(List.of(TOPIC)); + + // Wait until preferred replica is set to follower. + waitForCondition( + () -> getPreferredReplica() == FOLLOWER_BROKER_ID, + "Preferred replica is not set to follower" + ); + + // Produce and consume from both consumers. + produceMessages(1); + consumeRecords(followerConsumer, 1); + consumeRecords(leaderConsumer, 1); + + // Shutdown follower, produce and consume should work for both consumers. + cluster.shutdownBroker(FOLLOWER_BROKER_ID); + produceMessages(1); + consumeRecords(followerConsumer, 1); + consumeRecords(leaderConsumer, 1); + + // Start the follower and wait until preferred replica is set to follower. + cluster.startBroker(FOLLOWER_BROKER_ID); + waitForCondition( + () -> getPreferredReplica() == FOLLOWER_BROKER_ID, + "Preferred replica is not set to follower after restart" + ); + + // Produce and consume should still work for both consumers. + produceMessages(1); + consumeRecords(followerConsumer, 1); + consumeRecords(leaderConsumer, 1); + } + } + + @ClusterTest + public void testRackAwareRangeAssignor(ClusterInstance cluster) throws Exception { + this.cluster = cluster; + List<Integer> partitionList = cluster.brokers().keySet().stream() + .sorted() + .toList(); + + String topicWithAllPartitionsOnAllRacks = "topicWithAllPartitionsOnAllRacks"; + cluster.createTopic(topicWithAllPartitionsOnAllRacks, + cluster.brokers().size(), + (short) cluster.brokers().size()); + + // Racks are in order of broker ids, assign leaders in reverse order + String topicWithSingleRackPartitions = "topicWithSingleRackPartitions"; + var replicaAssignment = new LinkedHashMap<Integer, List<Integer>>(); + for (int i : partitionList) { + replicaAssignment.put(i, List.of(cluster.brokers().size() - i - 1)); + } + cluster.createTopicWithAssignment(topicWithSingleRackPartitions, replicaAssignment); + + List<HashMap<String, Object>> consumerConfigs = cluster.brokers().keySet().stream() + .sorted() + .map(brokerId -> { + var config = new HashMap<String, Object>(); + config.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, cluster.bootstrapServers()); + config.put(ConsumerConfig.GROUP_ID_CONFIG, "rack-aware-group"); + config.put(ConsumerConfig.PARTITION_ASSIGNMENT_STRATEGY_CONFIG, RangeAssignor.class.getName()); + config.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest"); + config.put(ConsumerConfig.CLIENT_RACK_CONFIG, String.valueOf(cluster.brokers().get(brokerId).config().rack().orElse(null))); + config.put(ConsumerConfig.GROUP_INSTANCE_ID_CONFIG, "instance-" + brokerId); + config.put(ConsumerConfig.METADATA_MAX_AGE_CONFIG, "1000"); + config.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, "false"); + return config; + }) + .toList(); + + List<Consumer<byte[], byte[]>> consumers = consumerConfigs.stream() + .map(config -> cluster.<byte[], byte[]>consumer(config)) Review Comment: thank you for the review! updated. -- 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]
