consumer_partition.go

// 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 pulsar

import (
	"container/list"
	"encoding/hex"
	"errors"
	"fmt"
	"math"
	"strings"
	"sync"
	"sync/atomic"
	"time"

	"google.golang.org/protobuf/proto"

	"github.com/apache/pulsar-client-go/pulsar/crypto"
	"github.com/apache/pulsar-client-go/pulsar/internal"
	"github.com/apache/pulsar-client-go/pulsar/internal/compression"
	cryptointernal "github.com/apache/pulsar-client-go/pulsar/internal/crypto"
	pb "github.com/apache/pulsar-client-go/pulsar/internal/pulsar_proto"
	"github.com/apache/pulsar-client-go/pulsar/log"
	"github.com/bits-and-blooms/bitset"

	uAtomic "go.uber.org/atomic"
)

type consumerState int

const (
	// consumer states
	consumerInit = iota
	consumerReady
	consumerClosing
	consumerClosed
)

func (s consumerState) String() string {
	switch s {
	case consumerInit:
		return "Initializing"
	case consumerReady:
		return "Ready"
	case consumerClosing:
		return "Closing"
	case consumerClosed:
		return "Closed"
	default:
		return "Unknown"
	}
}

type subscriptionMode int

const (
	// Make the subscription to be backed by a durable cursor that will retain messages and persist the current
	// position
	durable subscriptionMode = iota

	// Lightweight subscription mode that doesn't have a durable cursor associated
	nonDurable
)

const (
	noMessageEntry = -1
)

type partitionConsumerOpts struct {
	topic                       string
	consumerName                string
	subscription                string
	subscriptionType            SubscriptionType
	subscriptionInitPos         SubscriptionInitialPosition
	partitionIdx                int
	receiverQueueSize           int
	nackRedeliveryDelay         time.Duration
	nackBackoffPolicy           NackBackoffPolicy
	metadata                    map[string]string
	subProperties               map[string]string
	replicateSubscriptionState  bool
	startMessageID              *trackingMessageID
	startMessageIDInclusive     bool
	subscriptionMode            subscriptionMode
	readCompacted               bool
	disableForceTopicCreation   bool
	interceptors                ConsumerInterceptors
	maxReconnectToBroker        *uint
	backoffPolicy               internal.BackoffPolicy
	keySharedPolicy             *KeySharedPolicy
	schema                      Schema
	decryption                  *MessageDecryptionInfo
	ackWithResponse             bool
	maxPendingChunkedMessage    int
	expireTimeOfIncompleteChunk time.Duration
	autoAckIncompleteChunk      bool
	// in failover mode, this callback will be called when consumer change
	consumerEventListener ConsumerEventListener
	enableBatchIndexAck   bool
	ackGroupingOptions    *AckGroupingOptions
}

type ConsumerEventListener interface {
	BecameActive(consumer Consumer, topicName string, partition int32)
	BecameInactive(consumer Consumer, topicName string, partition int32)
}

type partitionConsumer struct {
	client *client

	// this is needed for sending ConsumerMessage on the messageCh
	parentConsumer Consumer
	state          uAtomic.Int32
	options        *partitionConsumerOpts

	conn uAtomic.Value

	topic        string
	name         string
	consumerID   uint64
	partitionIdx int32

	// shared channel
	messageCh chan ConsumerMessage

	// the number of message slots available
	availablePermits *availablePermits

	// the size of the queue channel for buffering messages
	queueSize       int32
	queueCh         chan []*message
	startMessageID  atomicMessageID
	lastDequeuedMsg *trackingMessageID

	eventsCh        chan interface{}
	connectedCh     chan struct{}
	connectClosedCh chan connectionClosed
	closeCh         chan struct{}
	clearQueueCh    chan func(id *trackingMessageID)

	nackTracker *negativeAcksTracker
	dlq         *dlqRouter

	log                  log.Logger
	compressionProviders sync.Map //map[pb.CompressionType]compression.Provider
	metrics              *internal.LeveledMetrics
	decryptor            cryptointernal.Decryptor
	schemaInfoCache      *schemaInfoCache

	chunkedMsgCtxMap   *chunkedMsgCtxMap
	unAckChunksTracker *unAckChunksTracker
	ackGroupingTracker ackGroupingTracker
}

func (pc *partitionConsumer) ActiveConsumerChanged(isActive bool) {
	listener := pc.options.consumerEventListener
	if listener == nil {
		// didn't set a listener
		return
	}
	if isActive {
		listener.BecameActive(pc.parentConsumer, pc.topic, pc.partitionIdx)
	} else {
		listener.BecameInactive(pc.parentConsumer, pc.topic, pc.partitionIdx)
	}
}

type availablePermits struct {
	permits int32
	pc      *partitionConsumer
}

func (p *availablePermits) inc() {
	// atomic add availablePermits
	ap := atomic.AddInt32(&p.permits, 1)

	// TODO implement a better flow controller
	// send more permits if needed
	flowThreshold := int32(math.Max(float64(p.pc.queueSize/2), 1))
	if ap >= flowThreshold {
		availablePermits := ap
		requestedPermits := ap
		// check if permits changed
		if !atomic.CompareAndSwapInt32(&p.permits, ap, 0) {
			return
		}

		p.pc.log.Debugf("requesting more permits=%d available=%d", requestedPermits, availablePermits)
		if err := p.pc.internalFlow(uint32(requestedPermits)); err != nil {
			p.pc.log.WithError(err).Error("unable to send permits")
		}
	}
}

func (p *availablePermits) reset() {
	atomic.StoreInt32(&p.permits, 0)
}

// atomicMessageID is a wrapper for trackingMessageID to make get and set atomic
type atomicMessageID struct {
	msgID *trackingMessageID
	sync.RWMutex
}

func (a *atomicMessageID) get() *trackingMessageID {
	a.RLock()
	defer a.RUnlock()
	return a.msgID
}

func (a *atomicMessageID) set(msgID *trackingMessageID) {
	a.Lock()
	defer a.Unlock()
	a.msgID = msgID
}

type schemaInfoCache struct {
	lock   sync.RWMutex
	cache  map[string]Schema
	client *client
	topic  string
}

func newSchemaInfoCache(client *client, topic string) *schemaInfoCache {
	return &schemaInfoCache{
		cache:  make(map[string]Schema),
		client: client,
		topic:  topic,
	}
}

func (s *schemaInfoCache) Get(schemaVersion []byte) (schema Schema, err error) {
	key := hex.EncodeToString(schemaVersion)
	s.lock.RLock()
	schema, ok := s.cache[key]
	s.lock.RUnlock()
	if ok {
		return schema, nil
	}

	pbSchema, err := s.client.lookupService.GetSchema(s.topic, schemaVersion)
	if err != nil {
		return nil, err
	}

	if pbSchema == nil {
		err = fmt.Errorf("schema not found for topic: [ %v ], schema version : [ %v ]", s.topic, schemaVersion)
		return nil, err
	}

	var properties = internal.ConvertToStringMap(pbSchema.Properties)

	schema, err = NewSchema(SchemaType(*pbSchema.Type), pbSchema.SchemaData, properties)
	if err != nil {
		return nil, err
	}
	s.add(key, schema)
	return schema, nil
}

func (s *schemaInfoCache) add(schemaVersionHash string, schema Schema) {
	s.lock.Lock()
	defer s.lock.Unlock()

	s.cache[schemaVersionHash] = schema
}

func newPartitionConsumer(parent Consumer, client *client, options *partitionConsumerOpts,
	messageCh chan ConsumerMessage, dlq *dlqRouter,
	metrics *internal.LeveledMetrics) (*partitionConsumer, error) {
	pc := &partitionConsumer{
		parentConsumer:       parent,
		client:               client,
		options:              options,
		topic:                options.topic,
		name:                 options.consumerName,
		consumerID:           client.rpcClient.NewConsumerID(),
		partitionIdx:         int32(options.partitionIdx),
		eventsCh:             make(chan interface{}, 10),
		queueSize:            int32(options.receiverQueueSize),
		queueCh:              make(chan []*message, options.receiverQueueSize),
		startMessageID:       atomicMessageID{msgID: options.startMessageID},
		connectedCh:          make(chan struct{}),
		messageCh:            messageCh,
		connectClosedCh:      make(chan connectionClosed, 10),
		closeCh:              make(chan struct{}),
		clearQueueCh:         make(chan func(id *trackingMessageID)),
		compressionProviders: sync.Map{},
		dlq:                  dlq,
		metrics:              metrics,
		schemaInfoCache:      newSchemaInfoCache(client, options.topic),
	}
	pc.availablePermits = &availablePermits{pc: pc}
	pc.chunkedMsgCtxMap = newChunkedMsgCtxMap(options.maxPendingChunkedMessage, pc)
	pc.unAckChunksTracker = newUnAckChunksTracker(pc)
	pc.ackGroupingTracker = newAckGroupingTracker(options.ackGroupingOptions,
		func(id MessageID) { pc.sendIndividualAck(id) },
		func(id MessageID) { pc.sendCumulativeAck(id) },
		func(ids []*pb.MessageIdData) { pc.eventsCh <- ids })
	pc.setConsumerState(consumerInit)
	pc.log = client.log.SubLogger(log.Fields{
		"name":         pc.name,
		"topic":        options.topic,
		"subscription": options.subscription,
		"consumerID":   pc.consumerID,
	})

	var decryptor cryptointernal.Decryptor
	if pc.options.decryption == nil {
		decryptor = cryptointernal.NewNoopDecryptor() // default to noopDecryptor
	} else {
		decryptor = cryptointernal.NewConsumerDecryptor(
			options.decryption.KeyReader,
			options.decryption.MessageCrypto,
			pc.log,
		)
	}

	pc.decryptor = decryptor

	pc.nackTracker = newNegativeAcksTracker(pc, options.nackRedeliveryDelay, options.nackBackoffPolicy, pc.log)

	err := pc.grabConn()
	if err != nil {
		pc.log.WithError(err).Error("Failed to create consumer")
		pc.nackTracker.Close()
		return nil, err
	}
	pc.log.Info("Created consumer")
	pc.setConsumerState(consumerReady)

	startingMessageID := pc.startMessageID.get()
	if pc.options.startMessageIDInclusive && startingMessageID != nil && startingMessageID.equal(latestMessageID) {
		msgID, err := pc.requestGetLastMessageID()
		if err != nil {
			pc.nackTracker.Close()
			return nil, err
		}
		if msgID.entryID != noMessageEntry {
			pc.startMessageID.set(msgID)

			// use the WithoutClear version because the dispatcher is not started yet
			err = pc.requestSeekWithoutClear(msgID.messageID)
			if err != nil {
				pc.nackTracker.Close()
				return nil, err
			}
		}
	}

	go pc.dispatcher()

	go pc.runEventsLoop()

	return pc, nil
}

func (pc *partitionConsumer) Unsubscribe() error {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to unsubscribe closing or closed consumer")
		return nil
	}

	req := &unsubscribeRequest{doneCh: make(chan struct{})}
	pc.eventsCh <- req

	// wait for the request to complete
	<-req.doneCh
	return req.err
}

func (pc *partitionConsumer) internalUnsubscribe(unsub *unsubscribeRequest) {
	defer close(unsub.doneCh)

	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to unsubscribe closing or closed consumer")
		return
	}

	pc.setConsumerState(consumerClosing)
	requestID := pc.client.rpcClient.NewRequestID()
	cmdUnsubscribe := &pb.CommandUnsubscribe{
		RequestId:  proto.Uint64(requestID),
		ConsumerId: proto.Uint64(pc.consumerID),
	}
	_, err := pc.client.rpcClient.RequestOnCnx(pc._getConn(), requestID, pb.BaseCommand_UNSUBSCRIBE, cmdUnsubscribe)
	if err != nil {
		pc.log.WithError(err).Error("Failed to unsubscribe consumer")
		unsub.err = err
		// Set the state to ready for closing the consumer
		pc.setConsumerState(consumerReady)
		// Should'nt remove the consumer handler
		return
	}

	pc._getConn().DeleteConsumeHandler(pc.consumerID)
	if pc.nackTracker != nil {
		pc.nackTracker.Close()
	}
	pc.log.Infof("The consumer[%d] successfully unsubscribed", pc.consumerID)
	pc.setConsumerState(consumerClosed)
}

func (pc *partitionConsumer) getLastMessageID() (*trackingMessageID, error) {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to redeliver closing or closed consumer")
		return nil, errors.New("failed to redeliver closing or closed consumer")
	}
	req := &getLastMsgIDRequest{doneCh: make(chan struct{})}
	pc.eventsCh <- req

	// wait for the request to complete
	<-req.doneCh
	return req.msgID, req.err
}

func (pc *partitionConsumer) internalGetLastMessageID(req *getLastMsgIDRequest) {
	defer close(req.doneCh)
	req.msgID, req.err = pc.requestGetLastMessageID()
}

func (pc *partitionConsumer) requestGetLastMessageID() (*trackingMessageID, error) {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to getLastMessageID closing or closed consumer")
		return nil, errors.New("failed to getLastMessageID closing or closed consumer")
	}

	requestID := pc.client.rpcClient.NewRequestID()
	cmdGetLastMessageID := &pb.CommandGetLastMessageId{
		RequestId:  proto.Uint64(requestID),
		ConsumerId: proto.Uint64(pc.consumerID),
	}
	res, err := pc.client.rpcClient.RequestOnCnx(pc._getConn(), requestID,
		pb.BaseCommand_GET_LAST_MESSAGE_ID, cmdGetLastMessageID)
	if err != nil {
		pc.log.WithError(err).Error("Failed to get last message id")
		return nil, err
	}
	id := res.Response.GetLastMessageIdResponse.GetLastMessageId()
	return convertToMessageID(id), nil
}

func (pc *partitionConsumer) ackID(msgID MessageID, withResponse bool) error {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to ack by closing or closed consumer")
		return errors.New("consumer state is closed")
	}

	if cmid, ok := msgID.(*chunkMessageID); ok {
		return pc.unAckChunksTracker.ack(cmid)
	}

	trackingID := toTrackingMessageID(msgID)

	if trackingID != nil && trackingID.ack() {
		pc.metrics.AcksCounter.Inc()
		pc.metrics.ProcessingTime.Observe(float64(time.Now().UnixNano()-trackingID.receivedTime.UnixNano()) / 1.0e9)
	} else if !pc.options.enableBatchIndexAck {
		return nil
	}

	var ackReq *ackRequest
	if withResponse {
		ackReq := pc.sendIndividualAck(trackingID)
		<-ackReq.doneCh
	} else {
		pc.ackGroupingTracker.add(trackingID)
	}
	pc.options.interceptors.OnAcknowledge(pc.parentConsumer, msgID)
	if ackReq == nil {
		return nil
	}
	return ackReq.err
}

func (pc *partitionConsumer) sendIndividualAck(msgID MessageID) *ackRequest {
	ackReq := &ackRequest{
		doneCh:  make(chan struct{}),
		ackType: individualAck,
		msgID:   *msgID.(*trackingMessageID),
	}
	pc.eventsCh <- ackReq
	return ackReq
}

func (pc *partitionConsumer) AckIDWithResponse(msgID MessageID) error {
	if !checkMessageIDType(msgID) {
		pc.log.Errorf("invalid message id type %T", msgID)
		return fmt.Errorf("invalid message id type %T", msgID)
	}
	return pc.ackID(msgID, true)
}

func (pc *partitionConsumer) AckID(msgID MessageID) error {
	if !checkMessageIDType(msgID) {
		pc.log.Errorf("invalid message id type %T", msgID)
		return fmt.Errorf("invalid message id type %T", msgID)
	}
	return pc.ackID(msgID, false)
}

func (pc *partitionConsumer) AckIDCumulative(msgID MessageID) error {
	if !checkMessageIDType(msgID) {
		pc.log.Errorf("invalid message id type %T", msgID)
		return fmt.Errorf("invalid message id type %T", msgID)
	}
	return pc.internalAckIDCumulative(msgID, false)
}

func (pc *partitionConsumer) AckIDWithResponseCumulative(msgID MessageID) error {
	if !checkMessageIDType(msgID) {
		pc.log.Errorf("invalid message id type %T", msgID)
		return fmt.Errorf("invalid message id type %T", msgID)
	}
	return pc.internalAckIDCumulative(msgID, true)
}

func (pc *partitionConsumer) internalAckIDCumulative(msgID MessageID, withResponse bool) error {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to ack by closing or closed consumer")
		return errors.New("consumer state is closed")
	}

	// chunk message id will be converted to tracking message id
	trackingID := toTrackingMessageID(msgID)
	if trackingID == nil {
		return errors.New("failed to convert trackingMessageID")
	}

	var msgIDToAck *trackingMessageID
	if trackingID.ackCumulative() || pc.options.enableBatchIndexAck {
		msgIDToAck = trackingID
	} else if !trackingID.tracker.hasPrevBatchAcked() {
		// get previous batch message id
		msgIDToAck = trackingID.prev()
		trackingID.tracker.setPrevBatchAcked()
	} else {
		// waiting for all the msgs are acked in this batch
		return nil
	}

	pc.metrics.AcksCounter.Inc()
	pc.metrics.ProcessingTime.Observe(float64(time.Now().UnixNano()-trackingID.receivedTime.UnixNano()) / 1.0e9)

	var ackReq *ackRequest
	if withResponse {
		ackReq := pc.sendCumulativeAck(msgIDToAck)
		<-ackReq.doneCh
	} else {
		pc.ackGroupingTracker.addCumulative(msgIDToAck)
	}

	pc.options.interceptors.OnAcknowledge(pc.parentConsumer, msgID)

	if cmid, ok := msgID.(*chunkMessageID); ok {
		pc.unAckChunksTracker.remove(cmid)
	}

	if ackReq == nil {
		return nil
	}
	return ackReq.err
}

func (pc *partitionConsumer) sendCumulativeAck(msgID MessageID) *ackRequest {
	ackReq := &ackRequest{
		doneCh:  make(chan struct{}),
		ackType: cumulativeAck,
		msgID:   *msgID.(*trackingMessageID),
	}
	pc.eventsCh <- ackReq
	return ackReq
}

func (pc *partitionConsumer) NackID(msgID MessageID) {
	if !checkMessageIDType(msgID) {
		pc.log.Warnf("invalid message id type %T", msgID)
		return
	}

	if cmid, ok := msgID.(*chunkMessageID); ok {
		pc.unAckChunksTracker.nack(cmid)
		return
	}

	trackingID := toTrackingMessageID(msgID)

	pc.nackTracker.Add(trackingID.messageID)
	pc.metrics.NacksCounter.Inc()
}

func (pc *partitionConsumer) NackMsg(msg Message) {
	pc.nackTracker.AddMessage(msg)
	pc.metrics.NacksCounter.Inc()
}

func (pc *partitionConsumer) Redeliver(msgIds []messageID) {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to redeliver closing or closed consumer")
		return
	}
	pc.eventsCh <- &redeliveryRequest{msgIds}

	iMsgIds := make([]MessageID, len(msgIds))
	for i := range iMsgIds {
		iMsgIds[i] = &msgIds[i]
	}
	pc.options.interceptors.OnNegativeAcksSend(pc.parentConsumer, iMsgIds)
}

func (pc *partitionConsumer) internalRedeliver(req *redeliveryRequest) {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to redeliver closing or closed consumer")
		return
	}
	msgIds := req.msgIds
	pc.log.Debug("Request redelivery after negative ack for messages", msgIds)

	msgIDDataList := make([]*pb.MessageIdData, len(msgIds))
	for i := 0; i < len(msgIds); i++ {
		msgIDDataList[i] = &pb.MessageIdData{
			LedgerId: proto.Uint64(uint64(msgIds[i].ledgerID)),
			EntryId:  proto.Uint64(uint64(msgIds[i].entryID)),
		}
	}

	err := pc.client.rpcClient.RequestOnCnxNoWait(pc._getConn(),
		pb.BaseCommand_REDELIVER_UNACKNOWLEDGED_MESSAGES, &pb.CommandRedeliverUnacknowledgedMessages{
			ConsumerId: proto.Uint64(pc.consumerID),
			MessageIds: msgIDDataList,
		})
	if err != nil {
		pc.log.Error("Connection was closed when request redeliver cmd")
	}
}

func (pc *partitionConsumer) getConsumerState() consumerState {
	return consumerState(pc.state.Load())
}

func (pc *partitionConsumer) setConsumerState(state consumerState) {
	pc.state.Store(int32(state))
}

func (pc *partitionConsumer) Close() {

	if pc.getConsumerState() != consumerReady {
		return
	}

	// flush all pending ACK requests and terminate the timer goroutine
	pc.ackGroupingTracker.close()

	// close chunkedMsgCtxMap
	pc.chunkedMsgCtxMap.Close()

	req := &closeRequest{doneCh: make(chan struct{})}
	pc.eventsCh <- req

	// wait for request to finish
	<-req.doneCh
}

func (pc *partitionConsumer) Seek(msgID MessageID) error {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to seek by closing or closed consumer")
		return errors.New("failed to seek by closing or closed consumer")
	}

	if !checkMessageIDType(msgID) {
		pc.log.Errorf("invalid message id type %T", msgID)
		return fmt.Errorf("invalid message id type %T", msgID)
	}

	req := &seekRequest{
		doneCh: make(chan struct{}),
	}
	if cmid, ok := msgID.(*chunkMessageID); ok {
		req.msgID = cmid.firstChunkID
	} else {
		tmid := toTrackingMessageID(msgID)
		req.msgID = tmid.messageID
	}

	pc.ackGroupingTracker.flushAndClean()
	pc.eventsCh <- req

	// wait for the request to complete
	<-req.doneCh
	return req.err
}

func (pc *partitionConsumer) internalSeek(seek *seekRequest) {
	defer close(seek.doneCh)
	seek.err = pc.requestSeek(seek.msgID)
}
func (pc *partitionConsumer) requestSeek(msgID *messageID) error {
	if err := pc.requestSeekWithoutClear(msgID); err != nil {
		return err
	}
	pc.clearReceiverQueue()
	return nil
}

func (pc *partitionConsumer) requestSeekWithoutClear(msgID *messageID) error {
	state := pc.getConsumerState()
	if state == consumerClosing || state == consumerClosed {
		pc.log.WithField("state", state).Error("failed seek by consumer is closing or has closed")
		return nil
	}

	id := &pb.MessageIdData{}
	err := proto.Unmarshal(msgID.Serialize(), id)
	if err != nil {
		pc.log.WithError(err).Errorf("deserialize message id error: %s", err.Error())
		return err
	}

	requestID := pc.client.rpcClient.NewRequestID()
	cmdSeek := &pb.CommandSeek{
		ConsumerId: proto.Uint64(pc.consumerID),
		RequestId:  proto.Uint64(requestID),
		MessageId:  id,
	}

	_, err = pc.client.rpcClient.RequestOnCnx(pc._getConn(), requestID, pb.BaseCommand_SEEK, cmdSeek)
	if err != nil {
		pc.log.WithError(err).Error("Failed to reset to message id")
		return err
	}
	return nil
}

func (pc *partitionConsumer) SeekByTime(time time.Time) error {
	if state := pc.getConsumerState(); state == consumerClosing || state == consumerClosed {
		pc.log.WithField("state", pc.state).Error("Failed seekByTime by consumer is closing or has closed")
		return errors.New("failed seekByTime by consumer is closing or has closed")
	}
	req := &seekByTimeRequest{
		doneCh:      make(chan struct{}),
		publishTime: time,
	}
	pc.ackGroupingTracker.flushAndClean()
	pc.eventsCh <- req

	// wait for the request to complete
	<-req.doneCh
	return req.err
}

func (pc *partitionConsumer) internalSeekByTime(seek *seekByTimeRequest) {
	defer close(seek.doneCh)

	state := pc.getConsumerState()
	if state == consumerClosing || state == consumerClosed {
		pc.log.WithField("state", pc.state).Error("Failed seekByTime by consumer is closing or has closed")
		return
	}

	requestID := pc.client.rpcClient.NewRequestID()
	cmdSeek := &pb.CommandSeek{
		ConsumerId:         proto.Uint64(pc.consumerID),
		RequestId:          proto.Uint64(requestID),
		MessagePublishTime: proto.Uint64(uint64(seek.publishTime.UnixNano() / int64(time.Millisecond))),
	}

	_, err := pc.client.rpcClient.RequestOnCnx(pc._getConn(), requestID, pb.BaseCommand_SEEK, cmdSeek)
	if err != nil {
		pc.log.WithError(err).Error("Failed to reset to message publish time")
		seek.err = err
		return
	}
	pc.clearReceiverQueue()
}

func (pc *partitionConsumer) internalAck(req *ackRequest) {
	defer close(req.doneCh)
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to ack by closing or closed consumer")
		return
	}
	msgID := req.msgID

	messageIDs := make([]*pb.MessageIdData, 1)
	messageIDs[0] = &pb.MessageIdData{
		LedgerId: proto.Uint64(uint64(msgID.ledgerID)),
		EntryId:  proto.Uint64(uint64(msgID.entryID)),
	}
	if pc.options.enableBatchIndexAck && msgID.tracker != nil {
		ackSet := msgID.tracker.toAckSet()
		if ackSet != nil {
			messageIDs[0].AckSet = ackSet
		}
	}

	reqID := pc.client.rpcClient.NewRequestID()
	cmdAck := &pb.CommandAck{
		ConsumerId: proto.Uint64(pc.consumerID),
		MessageId:  messageIDs,
	}

	switch req.ackType {
	case individualAck:
		cmdAck.AckType = pb.CommandAck_Individual.Enum()
	case cumulativeAck:
		cmdAck.AckType = pb.CommandAck_Cumulative.Enum()
	}

	if pc.options.ackWithResponse {
		cmdAck.RequestId = proto.Uint64(reqID)
		_, err := pc.client.rpcClient.RequestOnCnx(pc._getConn(), reqID, pb.BaseCommand_ACK, cmdAck)
		if err != nil {
			pc.log.WithError(err).Error("Ack with response error")
			req.err = err
		}
		return
	}

	err := pc.client.rpcClient.RequestOnCnxNoWait(pc._getConn(), pb.BaseCommand_ACK, cmdAck)
	if err != nil {
		pc.log.Error("Connection was closed when request ack cmd")
		req.err = err
	}
}

func (pc *partitionConsumer) internalAckList(msgIDs []*pb.MessageIdData) {
	pc.client.rpcClient.RequestOnCnxNoWait(pc._getConn(), pb.BaseCommand_ACK, &pb.CommandAck{
		AckType:    pb.CommandAck_Individual.Enum(),
		ConsumerId: proto.Uint64(pc.consumerID),
		MessageId:  msgIDs,
	})
}

func (pc *partitionConsumer) MessageReceived(response *pb.CommandMessage, headersAndPayload internal.Buffer) error {
	pbMsgID := response.GetMessageId()

	reader := internal.NewMessageReader(headersAndPayload)
	brokerMetadata, err := reader.ReadBrokerMetadata()
	if err != nil {
		// todo optimize use more appropriate error codes
		pc.discardCorruptedMessage(pbMsgID, pb.CommandAck_BatchDeSerializeError)
		return err
	}
	msgMeta, err := reader.ReadMessageMetadata()
	if err != nil {
		pc.discardCorruptedMessage(pbMsgID, pb.CommandAck_ChecksumMismatch)
		return err
	}
	decryptedPayload, err := pc.decryptor.Decrypt(headersAndPayload.ReadableSlice(), pbMsgID, msgMeta)
	// error decrypting the payload
	if err != nil {
		// default crypto failure action
		crypToFailureAction := crypto.ConsumerCryptoFailureActionFail
		if pc.options.decryption != nil {
			crypToFailureAction = pc.options.decryption.ConsumerCryptoFailureAction
		}

		switch crypToFailureAction {
		case crypto.ConsumerCryptoFailureActionFail:
			pc.log.Errorf("consuming message failed due to decryption err :%v", err)
			pc.NackID(newTrackingMessageID(int64(pbMsgID.GetLedgerId()), int64(pbMsgID.GetEntryId()), 0, 0, 0, nil))
			return err
		case crypto.ConsumerCryptoFailureActionDiscard:
			pc.discardCorruptedMessage(pbMsgID, pb.CommandAck_DecryptionError)
			return fmt.Errorf("discarding message on decryption error :%v", err)
		case crypto.ConsumerCryptoFailureActionConsume:
			pc.log.Warnf("consuming encrypted message due to error in decryption :%v", err)
			messages := []*message{
				{
					publishTime:  timeFromUnixTimestampMillis(msgMeta.GetPublishTime()),
					eventTime:    timeFromUnixTimestampMillis(msgMeta.GetEventTime()),
					key:          msgMeta.GetPartitionKey(),
					producerName: msgMeta.GetProducerName(),
					properties:   internal.ConvertToStringMap(msgMeta.GetProperties()),
					topic:        pc.topic,
					msgID: newMessageID(
						int64(pbMsgID.GetLedgerId()),
						int64(pbMsgID.GetEntryId()),
						pbMsgID.GetBatchIndex(),
						pc.partitionIdx,
						pbMsgID.GetBatchSize(),
					),
					payLoad:             headersAndPayload.ReadableSlice(),
					schema:              pc.options.schema,
					replicationClusters: msgMeta.GetReplicateTo(),
					replicatedFrom:      msgMeta.GetReplicatedFrom(),
					redeliveryCount:     response.GetRedeliveryCount(),
					encryptionContext:   createEncryptionContext(msgMeta),
					orderingKey:         string(msgMeta.OrderingKey),
				},
			}
			pc.queueCh <- messages
			return nil
		}
	}

	isChunkedMsg := false
	if msgMeta.GetNumChunksFromMsg() > 1 {
		isChunkedMsg = true
	}

	processedPayloadBuffer := internal.NewBufferWrapper(decryptedPayload)
	if isChunkedMsg {
		processedPayloadBuffer = pc.processMessageChunk(processedPayloadBuffer, msgMeta, pbMsgID)
		if processedPayloadBuffer == nil {
			return nil
		}
	}

	// decryption is success, decompress the payload
	uncompressedHeadersAndPayload, err := pc.Decompress(msgMeta, processedPayloadBuffer)
	if err != nil {
		pc.discardCorruptedMessage(pbMsgID, pb.CommandAck_DecompressionError)
		return err
	}

	// Reset the reader on the uncompressed buffer
	reader.ResetBuffer(uncompressedHeadersAndPayload)

	numMsgs := 1
	if msgMeta.NumMessagesInBatch != nil {
		numMsgs = int(msgMeta.GetNumMessagesInBatch())
	}

	messages := make([]*message, 0)
	var ackTracker *ackTracker
	// are there multiple messages in this batch?
	if numMsgs > 1 {
		ackTracker = newAckTracker(uint(numMsgs))
	}

	var ackSet *bitset.BitSet
	if response.GetAckSet() != nil {
		ackSetFromResponse := response.GetAckSet()
		buf := make([]uint64, len(ackSetFromResponse))
		for i := 0; i < len(buf); i++ {
			buf[i] = uint64(ackSetFromResponse[i])
		}
		ackSet = bitset.From(buf)
	}

	pc.metrics.MessagesReceived.Add(float64(numMsgs))
	pc.metrics.PrefetchedMessages.Add(float64(numMsgs))

	for i := 0; i < numMsgs; i++ {
		smm, payload, err := reader.ReadMessage()
		if err != nil || payload == nil {
			pc.discardCorruptedMessage(pbMsgID, pb.CommandAck_BatchDeSerializeError)
			return err
		}
		if ackSet != nil && !ackSet.Test(uint(i)) {
			pc.log.Debugf("Ignoring message from %vth message, which has been acknowledged", i)
			continue
		}

		pc.metrics.BytesReceived.Add(float64(len(payload)))
		pc.metrics.PrefetchedBytes.Add(float64(len(payload)))

		trackingMsgID := newTrackingMessageID(
			int64(pbMsgID.GetLedgerId()),
			int64(pbMsgID.GetEntryId()),
			int32(i),
			pc.partitionIdx,
			int32(numMsgs),
			ackTracker)
		// set the consumer so we know how to ack the message id
		trackingMsgID.consumer = pc

		if pc.messageShouldBeDiscarded(trackingMsgID) {
			pc.AckID(trackingMsgID)
			continue
		}

		var msgID MessageID
		if isChunkedMsg {
			ctx := pc.chunkedMsgCtxMap.get(msgMeta.GetUuid())
			if ctx == nil {
				// chunkedMsgCtxMap has closed because of consumer closed
				pc.log.Warnf("get chunkedMsgCtx for chunk with uuid %s failed because consumer has closed",
					msgMeta.Uuid)
				return nil
			}
			cmid := newChunkMessageID(ctx.firstChunkID(), ctx.lastChunkID())
			// set the consumer so we know how to ack the message id
			cmid.consumer = pc
			// clean chunkedMsgCtxMap
			pc.chunkedMsgCtxMap.remove(msgMeta.GetUuid())
			pc.unAckChunksTracker.add(cmid, ctx.chunkedMsgIDs)
			msgID = cmid
		} else {
			msgID = trackingMsgID
		}

		if pc.ackGroupingTracker.isDuplicate(msgID) {
			continue
		}

		var messageIndex *uint64
		var brokerPublishTime *time.Time
		if brokerMetadata != nil {
			if brokerMetadata.Index != nil {
				aux := brokerMetadata.GetIndex() - uint64(numMsgs) + uint64(i) + 1
				messageIndex = &aux
			}
			if brokerMetadata.BrokerTimestamp != nil {
				aux := timeFromUnixTimestampMillis(*brokerMetadata.BrokerTimestamp)
				brokerPublishTime = &aux
			}
		}

		var msg *message
		if smm != nil {
			msg = &message{
				publishTime:         timeFromUnixTimestampMillis(msgMeta.GetPublishTime()),
				eventTime:           timeFromUnixTimestampMillis(smm.GetEventTime()),
				key:                 smm.GetPartitionKey(),
				producerName:        msgMeta.GetProducerName(),
				properties:          internal.ConvertToStringMap(smm.GetProperties()),
				topic:               pc.topic,
				msgID:               msgID,
				payLoad:             payload,
				schema:              pc.options.schema,
				replicationClusters: msgMeta.GetReplicateTo(),
				replicatedFrom:      msgMeta.GetReplicatedFrom(),
				redeliveryCount:     response.GetRedeliveryCount(),
				schemaVersion:       msgMeta.GetSchemaVersion(),
				schemaInfoCache:     pc.schemaInfoCache,
				orderingKey:         string(smm.OrderingKey),
				index:               messageIndex,
				brokerPublishTime:   brokerPublishTime,
			}
		} else {
			msg = &message{
				publishTime:         timeFromUnixTimestampMillis(msgMeta.GetPublishTime()),
				eventTime:           timeFromUnixTimestampMillis(msgMeta.GetEventTime()),
				key:                 msgMeta.GetPartitionKey(),
				producerName:        msgMeta.GetProducerName(),
				properties:          internal.ConvertToStringMap(msgMeta.GetProperties()),
				topic:               pc.topic,
				msgID:               msgID,
				payLoad:             payload,
				schema:              pc.options.schema,
				replicationClusters: msgMeta.GetReplicateTo(),
				replicatedFrom:      msgMeta.GetReplicatedFrom(),
				redeliveryCount:     response.GetRedeliveryCount(),
				schemaVersion:       msgMeta.GetSchemaVersion(),
				schemaInfoCache:     pc.schemaInfoCache,
				index:               messageIndex,
				brokerPublishTime:   brokerPublishTime,
			}
		}

		pc.options.interceptors.BeforeConsume(ConsumerMessage{
			Consumer: pc.parentConsumer,
			Message:  msg,
		})

		messages = append(messages, msg)
	}

	// send messages to the dispatcher
	pc.queueCh <- messages
	return nil
}

func (pc *partitionConsumer) processMessageChunk(compressedPayload internal.Buffer,
	msgMeta *pb.MessageMetadata,
	pbMsgID *pb.MessageIdData) internal.Buffer {
	uuid := msgMeta.GetUuid()
	numChunks := msgMeta.GetNumChunksFromMsg()
	totalChunksSize := int(msgMeta.GetTotalChunkMsgSize())
	chunkID := msgMeta.GetChunkId()
	msgID := &messageID{
		ledgerID:     int64(pbMsgID.GetLedgerId()),
		entryID:      int64(pbMsgID.GetEntryId()),
		batchIdx:     -1,
		partitionIdx: pc.partitionIdx,
	}

	if msgMeta.GetChunkId() == 0 {
		pc.chunkedMsgCtxMap.addIfAbsent(uuid,
			numChunks,
			totalChunksSize,
		)
	}

	ctx := pc.chunkedMsgCtxMap.get(uuid)

	if ctx == nil || ctx.chunkedMsgBuffer == nil || chunkID != ctx.lastChunkedMsgID+1 {
		lastChunkedMsgID := -1
		totalChunks := -1
		if ctx != nil {
			lastChunkedMsgID = int(ctx.lastChunkedMsgID)
			totalChunks = int(ctx.totalChunks)
			ctx.chunkedMsgBuffer.Clear()
		}
		pc.log.Warnf(fmt.Sprintf(
			"Received unexpected chunk messageId %s, last-chunk-id %d, chunkId = %d, total-chunks %d",
			msgID.String(), lastChunkedMsgID, chunkID, totalChunks))
		pc.chunkedMsgCtxMap.remove(uuid)
		pc.availablePermits.inc()
		return nil
	}

	ctx.append(chunkID, msgID, compressedPayload)

	if msgMeta.GetChunkId() != msgMeta.GetNumChunksFromMsg()-1 {
		pc.availablePermits.inc()
		return nil
	}

	return ctx.chunkedMsgBuffer
}

func (pc *partitionConsumer) messageShouldBeDiscarded(msgID *trackingMessageID) bool {
	if pc.startMessageID.get() == nil {
		return false
	}
	// if we start at latest message, we should never discard
	if pc.options.startMessageID != nil && pc.options.startMessageID.equal(latestMessageID) {
		return false
	}

	if pc.options.startMessageIDInclusive {
		return pc.startMessageID.get().greater(msgID.messageID)
	}

	// Non inclusive
	return pc.startMessageID.get().greaterEqual(msgID.messageID)
}

// create EncryptionContext from message metadata
// this will be used to decrypt the message payload outside of this client
// it is the responsibility of end user to decrypt the payload
// It will be used only when  crypto failure action is set to consume i.e crypto.ConsumerCryptoFailureActionConsume
func createEncryptionContext(msgMeta *pb.MessageMetadata) *EncryptionContext {
	encCtx := EncryptionContext{
		Algorithm:        msgMeta.GetEncryptionAlgo(),
		Param:            msgMeta.GetEncryptionParam(),
		UncompressedSize: int(msgMeta.GetUncompressedSize()),
		BatchSize:        int(msgMeta.GetNumMessagesInBatch()),
	}

	if msgMeta.Compression != nil {
		encCtx.CompressionType = CompressionType(*msgMeta.Compression)
	}

	keyMap := map[string]EncryptionKey{}
	for _, k := range msgMeta.GetEncryptionKeys() {
		metaMap := map[string]string{}
		for _, m := range k.GetMetadata() {
			metaMap[*m.Key] = *m.Value
		}

		keyMap[*k.Key] = EncryptionKey{
			KeyValue: k.GetValue(),
			Metadata: metaMap,
		}
	}

	encCtx.Keys = keyMap
	return &encCtx
}

func (pc *partitionConsumer) ConnectionClosed() {
	// Trigger reconnection in the consumer goroutine
	pc.log.Debug("connection closed and send to connectClosedCh")
	pc.connectClosedCh <- connectionClosed{}
}

// Flow command gives additional permits to send messages to the consumer.
// A typical consumer implementation will use a queue to accumulate these messages
// before the application is ready to consume them. After the consumer is ready,
// the client needs to give permission to the broker to push messages.
func (pc *partitionConsumer) internalFlow(permits uint32) error {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to redeliver closing or closed consumer")
		return errors.New("consumer closing or closed")
	}
	if permits == 0 {
		return fmt.Errorf("invalid number of permits requested: %d", permits)
	}

	cmdFlow := &pb.CommandFlow{
		ConsumerId:     proto.Uint64(pc.consumerID),
		MessagePermits: proto.Uint32(permits),
	}
	err := pc.client.rpcClient.RequestOnCnxNoWait(pc._getConn(), pb.BaseCommand_FLOW, cmdFlow)
	if err != nil {
		pc.log.Error("Connection was closed when request flow cmd")
		return err
	}

	return nil
}

// dispatcher manages the internal message queue channel
// and manages the flow control
func (pc *partitionConsumer) dispatcher() {
	defer func() {
		pc.log.Debug("exiting dispatch loop")
	}()
	var messages []*message
	for {
		var queueCh chan []*message
		var messageCh chan ConsumerMessage
		var nextMessage ConsumerMessage

		// are there more messages to send?
		if len(messages) > 0 {
			nextMessage = ConsumerMessage{
				Consumer: pc.parentConsumer,
				Message:  messages[0],
			}

			if pc.dlq.shouldSendToDlq(&nextMessage) {
				// pass the message to the DLQ router
				pc.metrics.DlqCounter.Inc()
				messageCh = pc.dlq.Chan()
			} else {
				// pass the message to application channel
				messageCh = pc.messageCh
			}

			pc.metrics.PrefetchedMessages.Dec()
			pc.metrics.PrefetchedBytes.Sub(float64(len(messages[0].payLoad)))
		} else {
			// we are ready for more messages
			queueCh = pc.queueCh
		}

		select {
		case <-pc.closeCh:
			return

		case _, ok := <-pc.connectedCh:
			if !ok {
				return
			}
			pc.log.Debug("dispatcher received connection event")

			messages = nil

			// reset available permits
			pc.availablePermits.reset()
			initialPermits := uint32(pc.queueSize)

			pc.log.Debugf("dispatcher requesting initial permits=%d", initialPermits)
			// send initial permits
			if err := pc.internalFlow(initialPermits); err != nil {
				pc.log.WithError(err).Error("unable to send initial permits to broker")
			}

		case msgs, ok := <-queueCh:
			if !ok {
				return
			}
			// we only read messages here after the consumer has processed all messages
			// in the previous batch
			messages = msgs

		// if the messageCh is nil or the messageCh is full this will not be selected
		case messageCh <- nextMessage:
			// allow this message to be garbage collected
			messages[0] = nil
			messages = messages[1:]

			pc.availablePermits.inc()

		case clearQueueCb := <-pc.clearQueueCh:
			// drain the message queue on any new connection by sending a
			// special nil message to the channel so we know when to stop dropping messages
			var nextMessageInQueue *trackingMessageID
			go func() {
				pc.queueCh <- nil
			}()

			for m := range pc.queueCh {
				// the queue has been drained
				if m == nil {
					break
				} else if nextMessageInQueue == nil {
					nextMessageInQueue = toTrackingMessageID(m[0].msgID)
				}
			}

			messages = nil

			clearQueueCb(nextMessageInQueue)
		}
	}
}

const (
	individualAck = iota
	cumulativeAck
)

type ackRequest struct {
	doneCh  chan struct{}
	msgID   trackingMessageID
	ackType int
	err     error
}

type unsubscribeRequest struct {
	doneCh chan struct{}
	err    error
}

type closeRequest struct {
	doneCh chan struct{}
}

type redeliveryRequest struct {
	msgIds []messageID
}

type getLastMsgIDRequest struct {
	doneCh chan struct{}
	msgID  *trackingMessageID
	err    error
}

type seekRequest struct {
	doneCh chan struct{}
	msgID  *messageID
	err    error
}

type seekByTimeRequest struct {
	doneCh      chan struct{}
	publishTime time.Time
	err         error
}

func (pc *partitionConsumer) runEventsLoop() {
	defer func() {
		pc.log.Debug("exiting events loop")
	}()
	pc.log.Debug("get into runEventsLoop")

	go func() {
		for {
			select {
			case <-pc.closeCh:
				pc.log.Info("close consumer, exit reconnect")
				return
			case <-pc.connectClosedCh:
				pc.log.Debug("runEventsLoop will reconnect")
				pc.reconnectToBroker()
			}
		}
	}()

	for {
		for i := range pc.eventsCh {
			switch v := i.(type) {
			case *ackRequest:
				pc.internalAck(v)
			case []*pb.MessageIdData:
				pc.internalAckList(v)
			case *redeliveryRequest:
				pc.internalRedeliver(v)
			case *unsubscribeRequest:
				pc.internalUnsubscribe(v)
			case *getLastMsgIDRequest:
				pc.internalGetLastMessageID(v)
			case *seekRequest:
				pc.internalSeek(v)
			case *seekByTimeRequest:
				pc.internalSeekByTime(v)
			case *closeRequest:
				pc.internalClose(v)
				return
			}
		}
	}
}

func (pc *partitionConsumer) internalClose(req *closeRequest) {
	defer close(req.doneCh)
	state := pc.getConsumerState()
	if state != consumerReady {
		// this might be redundant but to ensure nack tracker is closed
		if pc.nackTracker != nil {
			pc.nackTracker.Close()
		}
		return
	}

	if state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Consumer is closing or has closed")
		if pc.nackTracker != nil {
			pc.nackTracker.Close()
		}
		return
	}

	pc.setConsumerState(consumerClosing)
	pc.log.Infof("Closing consumer=%d", pc.consumerID)

	requestID := pc.client.rpcClient.NewRequestID()
	cmdClose := &pb.CommandCloseConsumer{
		ConsumerId: proto.Uint64(pc.consumerID),
		RequestId:  proto.Uint64(requestID),
	}
	_, err := pc.client.rpcClient.RequestOnCnx(pc._getConn(), requestID, pb.BaseCommand_CLOSE_CONSUMER, cmdClose)
	if err != nil {
		pc.log.WithError(err).Warn("Failed to close consumer")
	} else {
		pc.log.Info("Closed consumer")
	}

	pc.compressionProviders.Range(func(_, v interface{}) bool {
		if provider, ok := v.(compression.Provider); ok {
			provider.Close()
		} else {
			err := fmt.Errorf("unexpected compression provider type: %T", v)
			pc.log.WithError(err).Warn("Failed to close compression provider")
		}
		return true
	})

	pc.setConsumerState(consumerClosed)
	pc._getConn().DeleteConsumeHandler(pc.consumerID)
	if pc.nackTracker != nil {
		pc.nackTracker.Close()
	}
	close(pc.closeCh)
}

func (pc *partitionConsumer) reconnectToBroker() {
	var maxRetry int

	if pc.options.maxReconnectToBroker == nil {
		maxRetry = -1
	} else {
		maxRetry = int(*pc.options.maxReconnectToBroker)
	}

	for maxRetry != 0 {
		if pc.getConsumerState() != consumerReady {
			// Consumer is already closing
			pc.log.Info("consumer state not ready, exit reconnect")
			return
		}

		var (
			delayReconnectTime time.Duration
			defaultBackoff     = internal.DefaultBackoff{}
		)

		if pc.options.backoffPolicy == nil {
			delayReconnectTime = defaultBackoff.Next()
		} else {
			delayReconnectTime = pc.options.backoffPolicy.Next()
		}

		pc.log.Info("Reconnecting to broker in ", delayReconnectTime)
		time.Sleep(delayReconnectTime)

		err := pc.grabConn()
		if err == nil {
			// Successfully reconnected
			pc.log.Info("Reconnected consumer to broker")
			return
		}
		pc.log.WithError(err).Error("Failed to create consumer at reconnect")
		errMsg := err.Error()
		if strings.Contains(errMsg, errTopicNotFount) {
			// when topic is deleted, we should give up reconnection.
			pc.log.Warn("Topic Not Found.")
			break
		}

		if maxRetry > 0 {
			maxRetry--
		}
		pc.metrics.ConsumersReconnectFailure.Inc()
		if maxRetry == 0 || defaultBackoff.IsMaxBackoffReached() {
			pc.metrics.ConsumersReconnectMaxRetry.Inc()
		}
	}
}

func (pc *partitionConsumer) grabConn() error {
	lr, err := pc.client.lookupService.Lookup(pc.topic)
	if err != nil {
		pc.log.WithError(err).Warn("Failed to lookup topic")
		return err
	}
	pc.log.Debugf("Lookup result: %+v", lr)

	subType := toProtoSubType(pc.options.subscriptionType)
	initialPosition := toProtoInitialPosition(pc.options.subscriptionInitPos)
	keySharedMeta := toProtoKeySharedMeta(pc.options.keySharedPolicy)
	requestID := pc.client.rpcClient.NewRequestID()

	var pbSchema *pb.Schema

	if pc.options.schema != nil && pc.options.schema.GetSchemaInfo() != nil {
		tmpSchemaType := pb.Schema_Type(int32(pc.options.schema.GetSchemaInfo().Type))
		pbSchema = &pb.Schema{
			Name:       proto.String(pc.options.schema.GetSchemaInfo().Name),
			Type:       &tmpSchemaType,
			SchemaData: []byte(pc.options.schema.GetSchemaInfo().Schema),
			Properties: internal.ConvertFromStringMap(pc.options.schema.GetSchemaInfo().Properties),
		}
		pc.log.Debugf("The partition consumer schema name is: %s", pbSchema.Name)
	} else {
		pc.log.Debug("The partition consumer schema is nil")
	}

	cmdSubscribe := &pb.CommandSubscribe{
		Topic:                      proto.String(pc.topic),
		Subscription:               proto.String(pc.options.subscription),
		SubType:                    subType.Enum(),
		ConsumerId:                 proto.Uint64(pc.consumerID),
		RequestId:                  proto.Uint64(requestID),
		ConsumerName:               proto.String(pc.name),
		PriorityLevel:              nil,
		Durable:                    proto.Bool(pc.options.subscriptionMode == durable),
		Metadata:                   internal.ConvertFromStringMap(pc.options.metadata),
		SubscriptionProperties:     internal.ConvertFromStringMap(pc.options.subProperties),
		ReadCompacted:              proto.Bool(pc.options.readCompacted),
		Schema:                     pbSchema,
		InitialPosition:            initialPosition.Enum(),
		ReplicateSubscriptionState: proto.Bool(pc.options.replicateSubscriptionState),
		KeySharedMeta:              keySharedMeta,
	}

	pc.startMessageID.set(pc.clearReceiverQueue())
	if pc.options.subscriptionMode != durable {
		// For regular subscriptions the broker will determine the restarting point
		cmdSubscribe.StartMessageId = convertToMessageIDData(pc.startMessageID.get())
	}

	if len(pc.options.metadata) > 0 {
		cmdSubscribe.Metadata = toKeyValues(pc.options.metadata)
	}

	if len(pc.options.subProperties) > 0 {
		cmdSubscribe.SubscriptionProperties = toKeyValues(pc.options.subProperties)
	}

	// force topic creation is enabled by default so
	// we only need to set the flag when disabling it
	if pc.options.disableForceTopicCreation {
		cmdSubscribe.ForceTopicCreation = proto.Bool(false)
	}

	res, err := pc.client.rpcClient.Request(lr.LogicalAddr, lr.PhysicalAddr, requestID,
		pb.BaseCommand_SUBSCRIBE, cmdSubscribe)

	if err != nil {
		pc.log.WithError(err).Error("Failed to create consumer")
		return err
	}

	if res.Response.ConsumerStatsResponse != nil {
		pc.name = res.Response.ConsumerStatsResponse.GetConsumerName()
	}

	pc._setConn(res.Cnx)
	pc.log.Info("Connected consumer")
	err = pc._getConn().AddConsumeHandler(pc.consumerID, pc)
	if err != nil {
		pc.log.WithError(err).Error("Failed to add consumer handler")
		return err
	}

	msgType := res.Response.GetType()

	switch msgType {
	case pb.BaseCommand_SUCCESS:
		// notify the dispatcher we have connection
		go func() {
			pc.connectedCh <- struct{}{}
		}()
		return nil
	case pb.BaseCommand_ERROR:
		errMsg := res.Response.GetError()
		return fmt.Errorf("%s: %s", errMsg.GetError().String(), errMsg.GetMessage())
	default:
		return newUnexpectedErrMsg(msgType, requestID)
	}
}

func (pc *partitionConsumer) clearQueueAndGetNextMessage() *trackingMessageID {
	if pc.getConsumerState() != consumerReady {
		return nil
	}
	wg := &sync.WaitGroup{}
	wg.Add(1)
	var msgID *trackingMessageID

	pc.clearQueueCh <- func(id *trackingMessageID) {
		msgID = id
		wg.Done()
	}

	wg.Wait()
	return msgID
}

/**
 * Clear the internal receiver queue and returns the message id of what was the 1st message in the queue that was
 * not seen by the application
 */
func (pc *partitionConsumer) clearReceiverQueue() *trackingMessageID {
	nextMessageInQueue := pc.clearQueueAndGetNextMessage()

	if pc.startMessageID.get() == nil {
		return pc.startMessageID.get()
	}

	if nextMessageInQueue != nil {
		return getPreviousMessage(nextMessageInQueue)
	} else if pc.lastDequeuedMsg != nil {
		// If the queue was empty we need to restart from the message just after the last one that has been dequeued
		// in the past
		return pc.lastDequeuedMsg
	} else {
		// No message was received or dequeued by this consumer. Next message would still be the startMessageId
		return pc.startMessageID.get()
	}
}

func getPreviousMessage(mid *trackingMessageID) *trackingMessageID {
	if mid.batchIdx >= 0 {
		return &trackingMessageID{
			messageID: &messageID{
				ledgerID:     mid.ledgerID,
				entryID:      mid.entryID,
				batchIdx:     mid.batchIdx - 1,
				partitionIdx: mid.partitionIdx,
			},
			tracker:      mid.tracker,
			consumer:     mid.consumer,
			receivedTime: mid.receivedTime,
		}
	}

	// Get on previous message in previous entry
	return &trackingMessageID{
		messageID: &messageID{
			ledgerID:     mid.ledgerID,
			entryID:      mid.entryID - 1,
			batchIdx:     mid.batchIdx,
			partitionIdx: mid.partitionIdx,
		},
		tracker:      mid.tracker,
		consumer:     mid.consumer,
		receivedTime: mid.receivedTime,
	}
}

func (pc *partitionConsumer) Decompress(msgMeta *pb.MessageMetadata, payload internal.Buffer) (internal.Buffer, error) {
	providerEntry, ok := pc.compressionProviders.Load(msgMeta.GetCompression())
	if !ok {
		newProvider, err := pc.initializeCompressionProvider(msgMeta.GetCompression())
		if err != nil {
			pc.log.WithError(err).Error("Failed to decompress message.")
			return nil, err
		}

		var loaded bool
		providerEntry, loaded = pc.compressionProviders.LoadOrStore(msgMeta.GetCompression(), newProvider)
		if loaded {
			// another thread already loaded this provider, so close the one we just initialized
			newProvider.Close()
		}
	}
	provider, ok := providerEntry.(compression.Provider)
	if !ok {
		err := fmt.Errorf("unexpected compression provider type: %T", providerEntry)
		pc.log.WithError(err).Error("Failed to decompress message.")
		return nil, err
	}

	uncompressed, err := provider.Decompress(nil, payload.ReadableSlice(), int(msgMeta.GetUncompressedSize()))
	if err != nil {
		return nil, err
	}

	return internal.NewBufferWrapper(uncompressed), nil
}

func (pc *partitionConsumer) initializeCompressionProvider(
	compressionType pb.CompressionType) (compression.Provider, error) {
	switch compressionType {
	case pb.CompressionType_NONE:
		return compression.NewNoopProvider(), nil
	case pb.CompressionType_ZLIB:
		return compression.NewZLibProvider(), nil
	case pb.CompressionType_LZ4:
		return compression.NewLz4Provider(), nil
	case pb.CompressionType_ZSTD:
		return compression.NewZStdProvider(compression.Default), nil
	}

	return nil, fmt.Errorf("unsupported compression type: %v", compressionType)
}

func (pc *partitionConsumer) discardCorruptedMessage(msgID *pb.MessageIdData,
	validationError pb.CommandAck_ValidationError) {
	if state := pc.getConsumerState(); state == consumerClosed || state == consumerClosing {
		pc.log.WithField("state", state).Error("Failed to discardCorruptedMessage " +
			"by closing or closed consumer")
		return
	}
	pc.log.WithFields(log.Fields{
		"msgID":           msgID,
		"validationError": validationError,
	}).Error("Discarding corrupted message")

	err := pc.client.rpcClient.RequestOnCnxNoWait(pc._getConn(),
		pb.BaseCommand_ACK, &pb.CommandAck{
			ConsumerId:      proto.Uint64(pc.consumerID),
			MessageId:       []*pb.MessageIdData{msgID},
			AckType:         pb.CommandAck_Individual.Enum(),
			ValidationError: validationError.Enum(),
		})
	if err != nil {
		pc.log.Error("Connection was closed when request ack cmd")
	}
	pc.availablePermits.inc()
}

// _setConn sets the internal connection field of this partition consumer atomically.
// Note: should only be called by this partition consumer when a new connection is available.
func (pc *partitionConsumer) _setConn(conn internal.Connection) {
	pc.conn.Store(conn)
}

// _getConn returns internal connection field of this partition consumer atomically.
// Note: should only be called by this partition consumer before attempting to use the connection
func (pc *partitionConsumer) _getConn() internal.Connection {
	// Invariant: The conn must be non-nill for the lifetime of the partitionConsumer.
	//            For this reason we leave this cast unchecked and panic() if the
	//            invariant is broken
	return pc.conn.Load().(internal.Connection)
}

func convertToMessageIDData(msgID *trackingMessageID) *pb.MessageIdData {
	if msgID == nil {
		return nil
	}

	return &pb.MessageIdData{
		LedgerId: proto.Uint64(uint64(msgID.ledgerID)),
		EntryId:  proto.Uint64(uint64(msgID.entryID)),
	}
}

func convertToMessageID(id *pb.MessageIdData) *trackingMessageID {
	if id == nil {
		return nil
	}

	msgID := &trackingMessageID{
		messageID: &messageID{
			ledgerID: int64(*id.LedgerId),
			entryID:  int64(*id.EntryId),
		},
	}
	if id.BatchIndex != nil {
		msgID.batchIdx = *id.BatchIndex
	}

	return msgID
}

type chunkedMsgCtx struct {
	totalChunks      int32
	chunkedMsgBuffer internal.Buffer
	lastChunkedMsgID int32
	chunkedMsgIDs    []*messageID
	receivedTime     int64

	mu sync.Mutex
}

func newChunkedMsgCtx(numChunksFromMsg int32, totalChunkMsgSize int) *chunkedMsgCtx {
	return &chunkedMsgCtx{
		totalChunks:      numChunksFromMsg,
		chunkedMsgBuffer: internal.NewBuffer(totalChunkMsgSize),
		lastChunkedMsgID: -1,
		chunkedMsgIDs:    make([]*messageID, numChunksFromMsg),
		receivedTime:     time.Now().Unix(),
	}
}

func (c *chunkedMsgCtx) append(chunkID int32, msgID *messageID, partPayload internal.Buffer) {
	c.mu.Lock()
	defer c.mu.Unlock()
	c.chunkedMsgIDs[chunkID] = msgID
	c.chunkedMsgBuffer.Write(partPayload.ReadableSlice())
	c.lastChunkedMsgID = chunkID
}

func (c *chunkedMsgCtx) firstChunkID() *messageID {
	c.mu.Lock()
	defer c.mu.Unlock()
	if len(c.chunkedMsgIDs) == 0 {
		return nil
	}
	return c.chunkedMsgIDs[0]
}

func (c *chunkedMsgCtx) lastChunkID() *messageID {
	c.mu.Lock()
	defer c.mu.Unlock()
	if len(c.chunkedMsgIDs) == 0 {
		return nil
	}
	return c.chunkedMsgIDs[len(c.chunkedMsgIDs)-1]
}

func (c *chunkedMsgCtx) discard(pc *partitionConsumer) {
	c.mu.Lock()
	defer c.mu.Unlock()

	for _, mid := range c.chunkedMsgIDs {
		if mid == nil {
			continue
		}
		pc.log.Info("Removing chunk message-id", mid.String())
		tmid := toTrackingMessageID(mid)
		pc.AckID(tmid)
	}
}

type chunkedMsgCtxMap struct {
	chunkedMsgCtxs map[string]*chunkedMsgCtx
	pendingQueue   *list.List
	maxPending     int
	pc             *partitionConsumer
	mu             sync.Mutex
	closed         bool
}

func newChunkedMsgCtxMap(maxPending int, pc *partitionConsumer) *chunkedMsgCtxMap {
	return &chunkedMsgCtxMap{
		chunkedMsgCtxs: make(map[string]*chunkedMsgCtx, maxPending),
		pendingQueue:   list.New(),
		maxPending:     maxPending,
		pc:             pc,
		mu:             sync.Mutex{},
	}
}

func (c *chunkedMsgCtxMap) addIfAbsent(uuid string, totalChunks int32, totalChunkMsgSize int) {
	c.mu.Lock()
	defer c.mu.Unlock()
	if c.closed {
		return
	}
	if _, ok := c.chunkedMsgCtxs[uuid]; !ok {
		c.chunkedMsgCtxs[uuid] = newChunkedMsgCtx(totalChunks, totalChunkMsgSize)
		c.pendingQueue.PushBack(uuid)
		go c.discardChunkIfExpire(uuid, true, c.pc.options.expireTimeOfIncompleteChunk)
	}
	if c.maxPending > 0 && c.pendingQueue.Len() > c.maxPending {
		go c.discardOldestChunkMessage(c.pc.options.autoAckIncompleteChunk)
	}
}

func (c *chunkedMsgCtxMap) get(uuid string) *chunkedMsgCtx {
	c.mu.Lock()
	defer c.mu.Unlock()
	if c.closed {
		return nil
	}
	return c.chunkedMsgCtxs[uuid]
}

func (c *chunkedMsgCtxMap) remove(uuid string) {
	c.mu.Lock()
	defer c.mu.Unlock()
	if c.closed {
		return
	}
	delete(c.chunkedMsgCtxs, uuid)
	e := c.pendingQueue.Front()
	for ; e != nil; e = e.Next() {
		if e.Value.(string) == uuid {
			c.pendingQueue.Remove(e)
			break
		}
	}
}

func (c *chunkedMsgCtxMap) discardOldestChunkMessage(autoAck bool) {
	c.mu.Lock()
	defer c.mu.Unlock()
	if c.closed || (c.maxPending > 0 && c.pendingQueue.Len() <= c.maxPending) {
		return
	}
	oldest := c.pendingQueue.Front().Value.(string)
	ctx, ok := c.chunkedMsgCtxs[oldest]
	if !ok {
		return
	}
	if autoAck {
		ctx.discard(c.pc)
	}
	delete(c.chunkedMsgCtxs, oldest)
	c.pc.log.Infof("Chunked message [%s] has been removed from chunkedMsgCtxMap", oldest)
}

func (c *chunkedMsgCtxMap) discardChunkMessage(uuid string, autoAck bool) {
	c.mu.Lock()
	defer c.mu.Unlock()
	if c.closed {
		return
	}
	ctx, ok := c.chunkedMsgCtxs[uuid]
	if !ok {
		return
	}
	if autoAck {
		ctx.discard(c.pc)
	}
	delete(c.chunkedMsgCtxs, uuid)
	e := c.pendingQueue.Front()
	for ; e != nil; e = e.Next() {
		if e.Value.(string) == uuid {
			c.pendingQueue.Remove(e)
			break
		}
	}
	c.pc.log.Infof("Chunked message [%s] has been removed from chunkedMsgCtxMap", uuid)
}

func (c *chunkedMsgCtxMap) discardChunkIfExpire(uuid string, autoAck bool, expire time.Duration) {
	timer := time.NewTimer(expire)
	<-timer.C
	c.discardChunkMessage(uuid, autoAck)
}

func (c *chunkedMsgCtxMap) Close() {
	c.mu.Lock()
	defer c.mu.Unlock()
	c.closed = true
}

type unAckChunksTracker struct {
	// TODO: use hash code of chunkMessageID as the key
	chunkIDs map[chunkMessageID][]*messageID
	pc       *partitionConsumer
	mu       sync.Mutex
}

func newUnAckChunksTracker(pc *partitionConsumer) *unAckChunksTracker {
	return &unAckChunksTracker{
		chunkIDs: make(map[chunkMessageID][]*messageID),
		pc:       pc,
	}
}

func (u *unAckChunksTracker) add(cmid *chunkMessageID, ids []*messageID) {
	u.mu.Lock()
	defer u.mu.Unlock()

	u.chunkIDs[*cmid] = ids
}

func (u *unAckChunksTracker) get(cmid *chunkMessageID) []*messageID {
	u.mu.Lock()
	defer u.mu.Unlock()

	return u.chunkIDs[*cmid]
}

func (u *unAckChunksTracker) remove(cmid *chunkMessageID) {
	u.mu.Lock()
	defer u.mu.Unlock()

	delete(u.chunkIDs, *cmid)
}

func (u *unAckChunksTracker) ack(cmid *chunkMessageID) error {
	ids := u.get(cmid)
	for _, id := range ids {
		if err := u.pc.AckID(id); err != nil {
			return err
		}
	}
	u.remove(cmid)
	return nil
}

func (u *unAckChunksTracker) nack(cmid *chunkMessageID) {
	ids := u.get(cmid)
	for _, id := range ids {
		u.pc.NackID(id)
	}
	u.remove(cmid)
}