dajac commented on code in PR #14067: URL: https://github.com/apache/kafka/pull/14067#discussion_r1270960303
########## group-coordinator/src/main/java/org/apache/kafka/coordinator/group/OffsetMetadataManager.java: ########## @@ -0,0 +1,448 @@ +/* + * 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.coordinator.group; + +import org.apache.kafka.common.KafkaException; +import org.apache.kafka.common.TopicPartition; +import org.apache.kafka.common.errors.ApiException; +import org.apache.kafka.common.errors.GroupIdNotFoundException; +import org.apache.kafka.common.message.OffsetCommitRequestData; +import org.apache.kafka.common.message.OffsetCommitResponseData; +import org.apache.kafka.common.protocol.Errors; +import org.apache.kafka.common.record.RecordBatch; +import org.apache.kafka.common.requests.OffsetCommitRequest; +import org.apache.kafka.common.requests.RequestContext; +import org.apache.kafka.common.utils.LogContext; +import org.apache.kafka.common.utils.Time; +import org.apache.kafka.coordinator.group.consumer.ConsumerGroup; +import org.apache.kafka.coordinator.group.generated.OffsetCommitKey; +import org.apache.kafka.coordinator.group.generated.OffsetCommitValue; +import org.apache.kafka.coordinator.group.generic.GenericGroup; +import org.apache.kafka.coordinator.group.runtime.CoordinatorResult; +import org.apache.kafka.image.MetadataDelta; +import org.apache.kafka.image.MetadataImage; +import org.apache.kafka.timeline.SnapshotRegistry; +import org.apache.kafka.timeline.TimelineHashMap; +import org.slf4j.Logger; + +import java.util.ArrayList; +import java.util.List; +import java.util.Map; +import java.util.Optional; +import java.util.OptionalInt; +import java.util.OptionalLong; + +import static org.apache.kafka.coordinator.group.generic.GenericGroupState.COMPLETING_REBALANCE; +import static org.apache.kafka.coordinator.group.generic.GenericGroupState.DEAD; +import static org.apache.kafka.coordinator.group.generic.GenericGroupState.EMPTY; + +/** + * The OffsetMetadataManager manages the offsets of all the groups. It basically maintains + * a mapping from group id to topic-partition to offset. This class has two kinds of methods: + * 1) The request handlers which handle the requests and generate a response and records to + * mutate the hard state. Those records will be written by the runtime and applied to the + * hard state via the replay methods. + * 2) The replay methods which apply records to the hard state. Those are used in the request + * handling as well as during the initial loading of the records from the partitions. + */ +public class OffsetMetadataManager { + public static class Builder { + private LogContext logContext = null; + private SnapshotRegistry snapshotRegistry = null; + private Time time = null; + private GroupMetadataManager groupMetadataManager = null; + private int offsetMetadataMaxSize = 4096; + private MetadataImage metadataImage = null; + + Builder withLogContext(LogContext logContext) { + this.logContext = logContext; + return this; + } + + Builder withSnapshotRegistry(SnapshotRegistry snapshotRegistry) { + this.snapshotRegistry = snapshotRegistry; + return this; + } + + Builder withTime(Time time) { + this.time = time; + return this; + } + + Builder withGroupMetadataManager(GroupMetadataManager groupMetadataManager) { + this.groupMetadataManager = groupMetadataManager; + return this; + } + + Builder withOffsetMetadataMaxSize(int offsetMetadataMaxSize) { + this.offsetMetadataMaxSize = offsetMetadataMaxSize; + return this; + } + + Builder withMetadataImage(MetadataImage metadataImage) { + this.metadataImage = metadataImage; + return this; + } + + public OffsetMetadataManager build() { + if (logContext == null) logContext = new LogContext(); + if (snapshotRegistry == null) snapshotRegistry = new SnapshotRegistry(logContext); + if (metadataImage == null) metadataImage = MetadataImage.EMPTY; + if (time == null) time = Time.SYSTEM; + + if (groupMetadataManager == null) { + throw new IllegalArgumentException("GroupMetadataManager cannot be null"); + } + + return new OffsetMetadataManager( + snapshotRegistry, + logContext, + time, + metadataImage, + groupMetadataManager, + offsetMetadataMaxSize + ); + } + } + + /** + * The logger. + */ + private final Logger log; + + /** + * The snapshot registry. + */ + private final SnapshotRegistry snapshotRegistry; + + /** + * The system time. + */ + private final Time time; + + /** + * The metadata image. + */ + private MetadataImage metadataImage; + + /** + * The group metadata manager. + */ + private final GroupMetadataManager groupMetadataManager; + + /** + * The maximum allowed metadata for any offset commit. + */ + private final int offsetMetadataMaxSize; + + /** + * The offsets keyed by topic-partition and group id. + */ + private final TimelineHashMap<String, TimelineHashMap<TopicPartition, OffsetAndMetadata>> offsetsByGroup; + + OffsetMetadataManager( + SnapshotRegistry snapshotRegistry, + LogContext logContext, + Time time, + MetadataImage metadataImage, + GroupMetadataManager groupMetadataManager, + int offsetMetadataMaxSize + ) { + this.snapshotRegistry = snapshotRegistry; + this.log = logContext.logger(OffsetMetadataManager.class); + this.time = time; + this.metadataImage = metadataImage; + this.groupMetadataManager = groupMetadataManager; + this.offsetMetadataMaxSize = offsetMetadataMaxSize; + this.offsetsByGroup = new TimelineHashMap<>(snapshotRegistry, 0); + } + + /** + * Validates an OffsetCommit request. + * + * @param context The request context. + * @param request The actual request. + */ + private void validateOffsetCommit( + RequestContext context, + OffsetCommitRequestData request + ) throws ApiException { + Group group; + try { + // If the group does not exist and generation id is -1, the request comes from + // either the admin client or a consumer which does not use the group management + // facility. In this case, a so-called simple group is created and the request + // is accepted. + group = groupMetadataManager.getOrMaybeCreateSimpleGroup( + request.groupId(), + request.generationIdOrMemberEpoch() < 0 + ); + } catch (GroupIdNotFoundException ex) { + // Maintain backward compatibility. This is a bit weird in the + // context of the new protocol though. + throw Errors.ILLEGAL_GENERATION.exception(); + } + + // Validate the request based on the group type. + switch (group.type()) { + case GENERIC: + validateOffsetCommitForGenericGroup( + (GenericGroup) group, + request + ); + break; + + case CONSUMER: + validateOffsetCommitForConsumerGroup( + (ConsumerGroup) group, + context, + request + ); + break; + } + } + + /** + * Validates an OffsetCommit request for a generic group. + * + * @param group The generic group. + * @param request The actual request. + */ + public void validateOffsetCommitForGenericGroup( + GenericGroup group, + OffsetCommitRequestData request + ) throws KafkaException { + if (group.isInState(DEAD)) { + throw Errors.COORDINATOR_NOT_AVAILABLE.exception(); + } + + if (request.generationIdOrMemberEpoch() < 0 && group.isInState(EMPTY)) { + // When the generation id is -1, the request comes from either the admin client + // or a consumer which does not use the group management facility. In this case, + // the request can commit offsets if the group is empty. + return; + } + + Optional<String> groupInstanceId = OffsetCommitRequest.groupInstanceId(request); + if (request.generationIdOrMemberEpoch() >= 0 || !request.memberId().isEmpty() || groupInstanceId.isPresent()) { + // We are validating three things: + // 1. If the `groupInstanceId` is present, then it exists and is mapped to `memberId`; + // 2. The `memberId` exists in the group; and + // 3. The `generationId` matches the current generation id. + if (groupInstanceId.isPresent()) { + String memberId = group.staticMemberId(groupInstanceId.get()); + if (memberId == null) { + throw Errors.UNKNOWN_MEMBER_ID.exception(); + } else if (!request.memberId().equals(memberId)) { + throw Errors.FENCED_INSTANCE_ID.exception(); + } + } + + if (!group.hasMemberId(request.memberId())) { + throw Errors.UNKNOWN_MEMBER_ID.exception(); + } + + if (request.generationIdOrMemberEpoch() != group.generationId()) { + throw Errors.ILLEGAL_GENERATION.exception(); + } + } else if (!group.isInState(EMPTY)) { + // If the request does not contain the member id and the generation + // id (version 0), offset commits are only accepted when the group + // is not empty. + throw Errors.UNKNOWN_MEMBER_ID.exception(); + } + + if (group.isInState(COMPLETING_REBALANCE)) { + // We should not receive a commit request if the group has not completed rebalance; + // but since the consumer's member.id and generation is valid, it means it has received + // the latest group generation information from the JoinResponse. + // So let's return a REBALANCE_IN_PROGRESS to let consumer handle it gracefully. + throw Errors.REBALANCE_IN_PROGRESS.exception(); + } + } + + /** + * Validates an OffsetCommit request for a consumer group. + * + * @param group The consumer group. + * @param context The request context. + * @param request The actual request. + */ + public void validateOffsetCommitForConsumerGroup( + ConsumerGroup group, + RequestContext context, + OffsetCommitRequestData request + ) throws KafkaException { + if (request.generationIdOrMemberEpoch() < 0 && group.members().isEmpty()) { + // When the generation id is -1, the request comes from either the admin client + // or a consumer which does not use the group management facility. In this case, + // the request can commit offsets if the group is empty. + return; + } + + if (!group.hasMember(request.memberId())) { + throw Errors.UNKNOWN_MEMBER_ID.exception(); + } + + final int memberEpoch = group.getOrMaybeCreateMember(request.memberId(), false).memberEpoch(); + if (request.generationIdOrMemberEpoch() != memberEpoch) { + // Consumers using the new consumer group protocol (KIP-848) should be using the + // OffsetCommit API >= 9. As we don't support upgrading from the old to the new + // protocol yet, we return an UNSUPPORTED_VERSION error if an older version is + // used. We will revise this when the upgrade path is implemented. + if (context.header.apiVersion() >= 9) { + throw Errors.STALE_MEMBER_EPOCH.exception(); + } else { + throw Errors.UNSUPPORTED_VERSION.exception(); + } + } + } + + /** + * Handles an OffsetCommit request. + * + * @param context The request context. + * @param request The OffsetCommit request. + * + * @return A Result containing the OffsetCommitResponseData response and + * a list of records to update the state machine. + */ + public CoordinatorResult<OffsetCommitResponseData, Record> commitOffset( + RequestContext context, + OffsetCommitRequestData request + ) throws ApiException { + validateOffsetCommit(context, request); + + final long currentTimeMs = time.milliseconds(); + // "default" expiration timestamp is defined as now + retention. The retention may be overridden + // in versions from v2 to v4. Otherwise, the retention defined on the broker is used. If an explicit + // commit timestamp is provided (v1 only), the expiration timestamp is computed based on that. + final OptionalLong expireTimestampMs = request.retentionTimeMs() == OffsetCommitRequest.DEFAULT_RETENTION_TIME ? + OptionalLong.empty() : OptionalLong.of(currentTimeMs + request.retentionTimeMs()); + final OffsetCommitResponseData response = new OffsetCommitResponseData(); + final List<Record> records = new ArrayList<>(); + + request.topics().forEach(topic -> { + final OffsetCommitResponseData.OffsetCommitResponseTopic topicResponse = + new OffsetCommitResponseData.OffsetCommitResponseTopic() + .setName(topic.name()); + response.topics().add(topicResponse); + + topic.partitions().forEach(partition -> { + if (partition.committedMetadata() != null && partition.committedMetadata().length() > offsetMetadataMaxSize) { + topicResponse.partitions().add(new OffsetCommitResponseData.OffsetCommitResponsePartition() + .setPartitionIndex(partition.partitionIndex()) + .setErrorCode(Errors.OFFSET_METADATA_TOO_LARGE.code())); + } else { + log.debug("[GroupId {}] Committing offsets {} for partition {}-{} from member {} with leader epoch {}.", + request.groupId(), partition.committedOffset(), topic.name(), partition.partitionIndex(), + request.memberId(), partition.committedLeaderEpoch()); + + topicResponse.partitions().add(new OffsetCommitResponseData.OffsetCommitResponsePartition() + .setPartitionIndex(partition.partitionIndex()) + .setErrorCode(Errors.NONE.code())); + + final OptionalInt leaderEpoch = partition.committedLeaderEpoch() == RecordBatch.NO_PARTITION_LEADER_EPOCH ? Review Comment: This one is from [here](https://github.com/apache/kafka/blob/trunk/core/src/main/scala/kafka/coordinator/group/GroupCoordinatorAdapter.scala#L480). -- 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
