Search in sources :

Example 1 with UnalignedRecords

use of org.apache.kafka.common.record.UnalignedRecords in project kafka by apache.

the class KafkaRaftClient method handleFetchSnapshotRequest.

/**
 * Handle a FetchSnapshot request, similar to the Fetch request but we use {@link UnalignedRecords}
 * in response because the records are not necessarily offset-aligned.
 *
 * This API may return the following errors:
 *
 * - {@link Errors#INCONSISTENT_CLUSTER_ID} if the cluster id is presented in request
 *     but different from this node
 * - {@link Errors#BROKER_NOT_AVAILABLE} if this node is currently shutting down
 * - {@link Errors#FENCED_LEADER_EPOCH} if the epoch is smaller than this node's epoch
 * - {@link Errors#INVALID_REQUEST} if the request epoch is larger than the leader's current epoch
 *     or if either the fetch offset or the last fetched epoch is invalid
 * - {@link Errors#SNAPSHOT_NOT_FOUND} if the request snapshot id does not exists
 * - {@link Errors#POSITION_OUT_OF_RANGE} if the request snapshot offset out of range
 */
private FetchSnapshotResponseData handleFetchSnapshotRequest(RaftRequest.Inbound requestMetadata) {
    FetchSnapshotRequestData data = (FetchSnapshotRequestData) requestMetadata.data;
    if (!hasValidClusterId(data.clusterId())) {
        return new FetchSnapshotResponseData().setErrorCode(Errors.INCONSISTENT_CLUSTER_ID.code());
    }
    if (data.topics().size() != 1 && data.topics().get(0).partitions().size() != 1) {
        return FetchSnapshotResponse.withTopLevelError(Errors.INVALID_REQUEST);
    }
    Optional<FetchSnapshotRequestData.PartitionSnapshot> partitionSnapshotOpt = FetchSnapshotRequest.forTopicPartition(data, log.topicPartition());
    if (!partitionSnapshotOpt.isPresent()) {
        // The Raft client assumes that there is only one topic partition.
        TopicPartition unknownTopicPartition = new TopicPartition(data.topics().get(0).name(), data.topics().get(0).partitions().get(0).partition());
        return FetchSnapshotResponse.singleton(unknownTopicPartition, responsePartitionSnapshot -> responsePartitionSnapshot.setErrorCode(Errors.UNKNOWN_TOPIC_OR_PARTITION.code()));
    }
    FetchSnapshotRequestData.PartitionSnapshot partitionSnapshot = partitionSnapshotOpt.get();
    Optional<Errors> leaderValidation = validateLeaderOnlyRequest(partitionSnapshot.currentLeaderEpoch());
    if (leaderValidation.isPresent()) {
        return FetchSnapshotResponse.singleton(log.topicPartition(), responsePartitionSnapshot -> addQuorumLeader(responsePartitionSnapshot).setErrorCode(leaderValidation.get().code()));
    }
    OffsetAndEpoch snapshotId = new OffsetAndEpoch(partitionSnapshot.snapshotId().endOffset(), partitionSnapshot.snapshotId().epoch());
    Optional<RawSnapshotReader> snapshotOpt = log.readSnapshot(snapshotId);
    if (!snapshotOpt.isPresent()) {
        return FetchSnapshotResponse.singleton(log.topicPartition(), responsePartitionSnapshot -> addQuorumLeader(responsePartitionSnapshot).setErrorCode(Errors.SNAPSHOT_NOT_FOUND.code()));
    }
    RawSnapshotReader snapshot = snapshotOpt.get();
    long snapshotSize = snapshot.sizeInBytes();
    if (partitionSnapshot.position() < 0 || partitionSnapshot.position() >= snapshotSize) {
        return FetchSnapshotResponse.singleton(log.topicPartition(), responsePartitionSnapshot -> addQuorumLeader(responsePartitionSnapshot).setErrorCode(Errors.POSITION_OUT_OF_RANGE.code()));
    }
    if (partitionSnapshot.position() > Integer.MAX_VALUE) {
        throw new IllegalStateException(String.format("Trying to fetch a snapshot with size (%s) and a position (%s) larger than %s", snapshotSize, partitionSnapshot.position(), Integer.MAX_VALUE));
    }
    int maxSnapshotSize;
    try {
        maxSnapshotSize = Math.toIntExact(snapshotSize);
    } catch (ArithmeticException e) {
        maxSnapshotSize = Integer.MAX_VALUE;
    }
    UnalignedRecords records = snapshot.slice(partitionSnapshot.position(), Math.min(data.maxBytes(), maxSnapshotSize));
    return FetchSnapshotResponse.singleton(log.topicPartition(), responsePartitionSnapshot -> {
        addQuorumLeader(responsePartitionSnapshot).snapshotId().setEndOffset(snapshotId.offset).setEpoch(snapshotId.epoch);
        return responsePartitionSnapshot.setSize(snapshotSize).setPosition(partitionSnapshot.position()).setUnalignedRecords(records);
    });
}
Also used : FetchSnapshotResponseData(org.apache.kafka.common.message.FetchSnapshotResponseData) RawSnapshotReader(org.apache.kafka.snapshot.RawSnapshotReader) Errors(org.apache.kafka.common.protocol.Errors) TopicPartition(org.apache.kafka.common.TopicPartition) RaftUtil.hasValidTopicPartition(org.apache.kafka.raft.RaftUtil.hasValidTopicPartition) FetchSnapshotRequestData(org.apache.kafka.common.message.FetchSnapshotRequestData) UnalignedRecords(org.apache.kafka.common.record.UnalignedRecords)

Aggregations

TopicPartition (org.apache.kafka.common.TopicPartition)1 FetchSnapshotRequestData (org.apache.kafka.common.message.FetchSnapshotRequestData)1 FetchSnapshotResponseData (org.apache.kafka.common.message.FetchSnapshotResponseData)1 Errors (org.apache.kafka.common.protocol.Errors)1 UnalignedRecords (org.apache.kafka.common.record.UnalignedRecords)1 RaftUtil.hasValidTopicPartition (org.apache.kafka.raft.RaftUtil.hasValidTopicPartition)1 RawSnapshotReader (org.apache.kafka.snapshot.RawSnapshotReader)1