コード例 #1
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
//send the actual MutationStreamRequest on adminport
func (k *kvSender) sendMutationTopicRequest(ap *projClient.Client, topic string,
	reqTimestamps *protobuf.TsVbuuid,
	instances []*protobuf.Instance) (*protobuf.TopicResponse, error) {

	logging.Infof("KVSender::sendMutationTopicRequest Projector %v Topic %v %v \n\tInstances %v",
		ap, topic, reqTimestamps.GetBucket(), instances)

	logging.LazyVerbosef("KVSender::sendMutationTopicRequest RequestTS %v", reqTimestamps.Repr)

	endpointType := "dataport"

	if res, err := ap.MutationTopicRequest(topic, endpointType,
		[]*protobuf.TsVbuuid{reqTimestamps}, instances); err != nil {
		logging.Fatalf("KVSender::sendMutationTopicRequest Projector %v Topic %v %v \n\tUnexpected Error %v", ap,
			topic, reqTimestamps.GetBucket(), err)

		return res, err
	} else {
		logging.Infof("KVSender::sendMutationTopicRequest Success Projector %v Topic %v %v InstanceIds %v",
			ap, topic, reqTimestamps.GetBucket(), res.GetInstanceIds())
		if logging.IsEnabled(logging.Verbose) {
			logging.Verbosef("KVSender::sendMutationTopicRequest ActiveTs %v \n\tRollbackTs %v",
				debugPrintTs(res.GetActiveTimestamps(), reqTimestamps.GetBucket()),
				debugPrintTs(res.GetRollbackTimestamps(), reqTimestamps.GetBucket()))
		}
		return res, nil
	}
}
コード例 #2
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
//check if any vb in vbList is part of the given ts
func checkVbListInTS(vbList []uint32, ts *protobuf.TsVbuuid) bool {

	for _, vb := range vbList {
		if ts.Contains(uint16(vb)) == true {
			return true
		}
	}
	return false

}
コード例 #3
0
ファイル: dcp.go プロジェクト: prataprc/indexing
// StartVbStreams implements Feeder{} interface.
func (bdcp *bucketDcp) StartVbStreams(
	opaque uint16, reqTs *protobuf.TsVbuuid) error {

	var err error

	if bdcp.bucket != nil {
		bdcp.bucket.Refresh()
	}
	vbnos := c.Vbno32to16(reqTs.GetVbnos())
	vbuuids, seqnos := reqTs.GetVbuuids(), reqTs.GetSeqnos()
	for i, vbno := range vbnos {
		snapshots := reqTs.GetSnapshots()
		flags, vbuuid := uint32(0), vbuuids[i]
		start, end := seqnos[i], uint64(0xFFFFFFFFFFFFFFFF)
		snapStart, snapEnd := snapshots[i].GetStart(), snapshots[i].GetEnd()
		e := bdcp.dcpFeed.DcpRequestStream(
			vbno, opaque, flags, vbuuid, start, end, snapStart, snapEnd)
		if e != nil {
			err = e
		}
		// FIXME/TODO: the below sleep avoid back-to-back dispatch of
		// StreamRequest to DCP, which seem to cause some problems.
		time.Sleep(time.Millisecond)
	}
	return err
}
コード例 #4
0
ファイル: dcp.go プロジェクト: prataprc/indexing
// EndVbStreams implements Feeder{} interface.
func (bdcp *bucketDcp) EndVbStreams(
	opaque uint16, ts *protobuf.TsVbuuid) (err error) {

	if bdcp.bucket != nil {
		bdcp.bucket.Refresh()
	}
	vbnos := c.Vbno32to16(ts.GetVbnos())
	for _, vbno := range vbnos {
		if e := bdcp.dcpFeed.DcpCloseStream(vbno, opaque); e != nil {
			err = e
		}
	}
	return err
}
コード例 #5
0
ファイル: stream_admin.go プロジェクト: jchris/indexing
//
// Find the offset/index in the timestamp for the given vbucket no.  Return
// -1 if no matching vbno being found.
//
func findTimestampOffsetForVb(ts *protobuf.TsVbuuid, vbno uint32) int {

	if ts == nil {
		return -1
	}

	for i, ts_vbno := range ts.GetVbnos() {
		if ts_vbno == vbno {
			return i
		}
	}

	return -1
}
コード例 #6
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func updateCurrentTsFromResponse(bucket string,
	currentTs *protobuf.TsVbuuid, res *protobuf.TimestampResponse) *protobuf.TsVbuuid {

	currentTsList := res.GetCurrentTimestamps()
	for _, ts := range currentTsList {
		if ts != nil && !ts.IsEmpty() && ts.GetBucket() == bucket {
			if currentTs == nil {
				currentTs = ts.Clone()
			} else {
				currentTs = currentTs.Union(ts)
			}
		}
	}
	return currentTs

}
コード例 #7
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func updateActiveTsFromResponse(bucket string,
	activeTs *protobuf.TsVbuuid, res *protobuf.TopicResponse) *protobuf.TsVbuuid {

	activeTsList := res.GetActiveTimestamps()
	for _, ts := range activeTsList {
		if ts != nil && !ts.IsEmpty() && ts.GetBucket() == bucket {
			if activeTs == nil {
				activeTs = ts.Clone()
			} else {
				activeTs = activeTs.Union(ts)
			}
		}
	}
	return activeTs

}
コード例 #8
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func updateRollbackTsFromResponse(bucket string,
	rollbackTs *protobuf.TsVbuuid, res *protobuf.TopicResponse) *protobuf.TsVbuuid {

	rollbackTsList := res.GetRollbackTimestamps()
	for _, ts := range rollbackTsList {
		if ts != nil && !ts.IsEmpty() && ts.GetBucket() == bucket {
			if rollbackTs == nil {
				rollbackTs = ts.Clone()
			} else {
				rollbackTs = rollbackTs.Union(ts)
			}
		}
	}

	return rollbackTs

}
コード例 #9
0
ファイル: timer.go プロジェクト: jchris/indexing
func unmarshallTimestamp(str string) (*common.TsVbuuid, error) {

	data, err := base64.StdEncoding.DecodeString(str)
	if err != nil {
		return nil, err
	}

	source := new(protobuf.TsVbuuid)
	if err := proto.Unmarshal(data, source); err != nil {
		return nil, err
	}

	target := common.NewTsVbuuid(source.GetBucket(), NUM_VB)

	for _, vbno := range source.Vbnos {
		target.Seqnos[vbno] = source.Seqnos[vbno]
		target.Vbuuids[vbno] = source.Vbuuids[vbno]
	}

	return target, nil
}
コード例 #10
0
ファイル: stream_admin.go プロジェクト: jchris/indexing
//
// Compute a new request timestamp based on the response from projector.
// If all the vb is active for the given requestTs, then this function returns nil.
//
func recomputeRequestTimestamp(requestTs *protobuf.TsVbuuid,
	rollbackTimestamps []*protobuf.TsVbuuid) *protobuf.TsVbuuid {

	newTs := protobuf.NewTsVbuuid(DEFAULT_POOL_NAME, requestTs.GetBucket(), len(requestTs.GetVbnos()))
	rollbackTs := findTimestampForBucket(rollbackTimestamps, requestTs.GetBucket())

	for i, vbno := range requestTs.GetVbnos() {
		offset := findTimestampOffsetForVb(rollbackTs, vbno)
		if offset != -1 {
			// there is a failover Ts for this vbno.  Use that one for retry.
			newTs.Append(uint16(vbno), rollbackTs.Seqnos[offset], rollbackTs.Vbuuids[offset],
				rollbackTs.Snapshots[offset].GetStart(), rollbackTs.Snapshots[offset].GetEnd())
		} else {
			// the vb is not active, just copy from the original requestTS
			newTs.Append(uint16(vbno), requestTs.Seqnos[i], requestTs.Vbuuids[i],
				requestTs.Snapshots[i].GetStart(), requestTs.Snapshots[i].GetEnd())
		}
	}

	return newTs
}
コード例 #11
0
ファイル: kvdata.go プロジェクト: jchris/indexing
func (kvdata *KVData) scatterMutation(
	m *mc.DcpEvent, ts *protobuf.TsVbuuid) (err error) {

	vbno := m.VBucket

	switch m.Opcode {
	case mcd.DCP_STREAMREQ:
		if m.Status == mcd.ROLLBACK {
			fmsg := "%v ##%x StreamRequest ROLLBACK: %v\n"
			logging.Infof(fmsg, kvdata.logPrefix, m.Opaque, m)

		} else if m.Status != mcd.SUCCESS {
			fmsg := "%v ##%x StreamRequest %s: %v\n"
			logging.Errorf(fmsg, kvdata.logPrefix, m.Opaque, m.Status, m)

		} else if _, ok := kvdata.vrs[vbno]; ok {
			fmsg := "%v ##%x duplicate OpStreamRequest: %v\n"
			logging.Errorf(fmsg, kvdata.logPrefix, m.Opaque, m)

		} else if m.VBuuid, _, err = m.FailoverLog.Latest(); err != nil {
			panic(err)

		} else {
			fmsg := "%v ##%x StreamRequest: %v\n"
			logging.Tracef(fmsg, kvdata.logPrefix, m.Opaque, m)
			topic, bucket := kvdata.topic, kvdata.bucket
			m.Seqno, _ = ts.SeqnoFor(vbno)
			cluster, config := kvdata.feed.cluster, kvdata.config
			vr := NewVbucketRoutine(
				cluster, topic, bucket,
				m.Opaque, vbno, m.VBuuid, m.Seqno, config)
			_, err := vr.AddEngines(0xFFFF, kvdata.engines, kvdata.endpoints)
			if err != nil {
				panic(err)
			}
			if vr.Event(m) != nil {
				panic(err)
			}
			kvdata.vrs[vbno] = vr
		}
		kvdata.feed.PostStreamRequest(kvdata.bucket, m)

	case mcd.DCP_STREAMEND:
		if vr, ok := kvdata.vrs[vbno]; !ok {
			fmsg := "%v ##%x duplicate OpStreamEnd: %v\n"
			logging.Errorf(fmsg, kvdata.logPrefix, m.Opaque, m)

		} else if m.Status != mcd.SUCCESS {
			fmsg := "%v ##%x StreamEnd %s: %v\n"
			logging.Errorf(fmsg, kvdata.logPrefix, m.Opaque, m)

		} else {
			fmsg := "%v ##%x StreamEnd: %v\n"
			logging.Tracef(fmsg, kvdata.logPrefix, m.Opaque, m)
			if vr.Event(m) != nil {
				panic(err)
			}
			delete(kvdata.vrs, vbno)
		}
		kvdata.feed.PostStreamEnd(kvdata.bucket, m)

	case mcd.DCP_SNAPSHOT:
		if vr, ok := kvdata.vrs[vbno]; ok && (vr.Event(m) != nil) {
			panic(err)
		} else if !ok {
			fmsg := "%v ##%x unknown vbucket: %v\n"
			logging.Fatalf(fmsg, kvdata.logPrefix, m.Opaque, m)
		}

	case mcd.DCP_MUTATION, mcd.DCP_DELETION, mcd.DCP_EXPIRATION:
		if vr, ok := kvdata.vrs[vbno]; ok && (vr.Event(m) != nil) {
			panic(err)
		} else if !ok {
			fmsg := "%v ##%x unknown vbucket: %v\n"
			logging.Fatalf(fmsg, kvdata.logPrefix, m.Opaque, m)
		}
	}
	return
}
コード例 #12
0
ファイル: kvdata.go プロジェクト: jchris/indexing
// go-routine handles data path.
func (kvdata *KVData) runScatter(
	ts *protobuf.TsVbuuid, mutch <-chan *mc.DcpEvent) {

	// NOTE: panic will bubble up from vbucket-routine to kvdata.
	defer func() {
		if r := recover(); r != nil {
			fmsg := "%v ##%x runScatter() crashed: %v\n"
			logging.Errorf(fmsg, kvdata.logPrefix, kvdata.opaque, r)
			logging.Errorf("%s", logging.StackTrace())
		}
		kvdata.publishStreamEnd()
		kvdata.feed.PostFinKVdata(kvdata.bucket)
		close(kvdata.finch)
		logging.Infof("%v ##%x ... stopped\n", kvdata.logPrefix, kvdata.opaque)
	}()

	// stats
	eventCount, addCount, delCount := int64(0), int64(0), int64(0)
	tsCount := int64(0)
	heartBeat := time.After(kvdata.syncTimeout)
	fmsg := "%v ##%x heartbeat (%v) loaded ...\n"
	logging.Infof(fmsg, kvdata.logPrefix, kvdata.opaque, kvdata.syncTimeout)

loop:
	for {
		select {
		case m, ok := <-mutch:
			if ok == false { // upstream has closed
				break loop
			}
			kvdata.scatterMutation(m, ts)
			eventCount++

		case <-heartBeat:
			vrs := make([]*VbucketRoutine, 0, len(kvdata.vrs))
			for _, vr := range kvdata.vrs {
				vrs = append(vrs, vr)
			}

			heartBeat = nil

			// propogate the sync-pulse via separate routine so that
			// the data-path is not blocked.
			go func() {
				// during cleanup, as long as the vbucket-routines are
				// shutdown this routine will eventually exit.
				for _, vr := range vrs {
					vr.SyncPulse()
				}
				if err := kvdata.ReloadHeartbeat(); err != nil {
					fmsg := "%v ##%x ReloadHeartbeat(): %v\n"
					logging.Errorf(fmsg, kvdata.logPrefix, kvdata.opaque, err)
				}
			}()

		case msg := <-kvdata.sbch:
			cmd := msg[0].(byte)
			switch cmd {
			case kvCmdAddEngines:
				opaque := msg[1].(uint16)
				respch := msg[4].(chan []interface{})
				if msg[2] != nil {
					for uuid, engine := range msg[2].(map[uint64]*Engine) {
						if _, ok := kvdata.engines[uuid]; !ok {
							fmsg := "%v ##%x new engine added %v"
							logging.Infof(fmsg, kvdata.logPrefix, opaque, uuid)
						}
						kvdata.engines[uuid] = engine
					}
				}
				if msg[3] != nil {
					rv := msg[3].(map[string]c.RouterEndpoint)
					for raddr, endp := range rv {
						fmsg := "%v ##%x updated endpoint %q"
						logging.Infof(fmsg, kvdata.logPrefix, opaque, raddr)
						kvdata.endpoints[raddr] = endp
					}
				}
				curSeqnos := make(map[uint16]uint64)
				if kvdata.engines != nil || kvdata.endpoints != nil {
					engines, endpoints := kvdata.engines, kvdata.endpoints
					for _, vr := range kvdata.vrs {
						curSeqno, err := vr.AddEngines(opaque, engines, endpoints)
						if err != nil {
							panic(err)
						}
						curSeqnos[vr.vbno] = curSeqno
					}
				}
				addCount++
				respch <- []interface{}{curSeqnos, nil}

			case kvCmdDelEngines:
				opaque := msg[1].(uint16)
				engineKeys := msg[2].([]uint64)
				respch := msg[3].(chan []interface{})
				for _, vr := range kvdata.vrs {
					if err := vr.DeleteEngines(opaque, engineKeys); err != nil {
						panic(err)
					}
				}
				for _, engineKey := range engineKeys {
					delete(kvdata.engines, engineKey)
					fmsg := "%v ##%x deleted engine %q"
					logging.Infof(fmsg, kvdata.logPrefix, opaque, engineKey)
				}
				delCount++
				respch <- []interface{}{nil}

			case kvCmdTs:
				_ /*opaque*/ = msg[1].(uint16)
				ts = ts.Union(msg[2].(*protobuf.TsVbuuid))
				respch := msg[3].(chan []interface{})
				tsCount++
				respch <- []interface{}{nil}

			case kvCmdGetStats:
				respch := msg[1].(chan []interface{})
				stats := kvdata.newStats()
				stats.Set("events", float64(eventCount))
				stats.Set("addInsts", float64(addCount))
				stats.Set("delInsts", float64(delCount))
				stats.Set("tsCount", float64(tsCount))
				statVbuckets := make(map[string]interface{})
				for i, vr := range kvdata.vrs {
					stats, err := vr.GetStatistics()
					if err != nil {
						panic(err)
					}
					statVbuckets[strconv.Itoa(int(i))] = stats
				}
				stats.Set("vbuckets", statVbuckets)
				respch <- []interface{}{map[string]interface{}(stats)}

			case kvCmdResetConfig:
				config, respch := msg[1].(c.Config), msg[2].(chan []interface{})
				if cv, ok := config["syncTimeout"]; ok && heartBeat != nil {
					kvdata.syncTimeout = time.Duration(cv.Int())
					kvdata.syncTimeout *= time.Millisecond
					logging.Infof(
						"%v ##%x heart-beat settings reloaded: %v\n",
						kvdata.logPrefix, kvdata.opaque, kvdata.syncTimeout)
					heartBeat = time.After(kvdata.syncTimeout)
				}
				for _, vr := range kvdata.vrs {
					if err := vr.ResetConfig(config); err != nil {
						panic(err)
					}
				}
				kvdata.config = kvdata.config.Override(config)
				respch <- []interface{}{nil}

			case kvCmdReloadHeartBeat:
				respch := msg[1].(chan []interface{})
				heartBeat = time.After(kvdata.syncTimeout)
				respch <- []interface{}{nil}

			case kvCmdClose:
				for _, vr := range kvdata.vrs {
					vr.Close()
				}
				respch := msg[1].(chan []interface{})
				respch <- []interface{}{nil}
				break loop
			}
		}
	}
}
コード例 #13
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func (k *kvSender) sendRestartVbuckets(ap *projClient.Client,
	topic string, connErrVbs []Vbucket,
	restartTs *protobuf.TsVbuuid) (*protobuf.TopicResponse, error) {

	logging.Infof("KVSender::sendRestartVbuckets Projector %v Topic %v %v", ap, topic, restartTs.GetBucket())
	logging.LazyVerbosef("KVSender::sendRestartVbuckets RestartTs %v", restartTs.Repr)

	//Shutdown the vbucket before restart if there was a ConnErr. If the vbucket is already
	//running, projector will ignore the request otherwise
	if len(connErrVbs) != 0 {

		logging.Infof("KVSender::sendRestartVbuckets ShutdownVbuckets %v Topic %v %v ConnErrVbs %v",
			ap, topic, restartTs.GetBucket(), connErrVbs)

		// Only shutting down the Vb that receieve connection error.  It is probably not harmful
		// to shutdown every VB in the repairTS, including those that only receive StreamEnd.
		// But due to network / projecctor latency, a VB StreamBegin may be coming on the way
		// for those VB (especially when RepairStream has already retried a couple of times).
		// So shutting all VB in restartTs may unnecessarily causing race condition and
		// make the protocol longer to converge. ShutdownVbuckets should have no effect on
		// projector that does not own the Vb.
		shutdownTs := k.computeShutdownTs(restartTs, connErrVbs)

		logging.Infof("KVSender::sendRestartVbuckets ShutdownVbuckets Projector %v Topic %v %v \n\tShutdownTs %v",
			ap, topic, restartTs.GetBucket(), shutdownTs.Repr())

		if err := ap.ShutdownVbuckets(topic, []*protobuf.TsVbuuid{shutdownTs}); err != nil {
			logging.Errorf("KVSender::sendRestartVbuckets Unexpected Error During "+
				"ShutdownVbuckets Request for Projector %v Topic %v. Err %v.", ap,
				topic, err)

			//all shutdownVbuckets errors are treated as success as it is a best-effort call.
			//RestartVbuckets errors will be acted upon.
		}
	}

	if res, err := ap.RestartVbuckets(topic, []*protobuf.TsVbuuid{restartTs}); err != nil {
		logging.Fatalf("KVSender::sendRestartVbuckets Unexpected Error During "+
			"Restart Vbuckets Request for Projector %v Topic %v %v . Err %v.", ap,
			topic, restartTs.GetBucket(), err)

		return res, err
	} else {
		logging.Infof("KVSender::sendRestartVbuckets Success Projector %v Topic %v %v", ap, topic, restartTs.GetBucket())
		if logging.IsEnabled(logging.Verbose) {
			logging.Verbosef("KVSender::sendRestartVbuckets \nActiveTs %v \nRollbackTs %v",
				debugPrintTs(res.GetActiveTimestamps(), restartTs.GetBucket()),
				debugPrintTs(res.GetRollbackTimestamps(), restartTs.GetBucket()))
		}
		return res, nil
	}
}
コード例 #14
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func (k *kvSender) addIndexForExistingBucket(streamId c.StreamId, bucket string, indexInstList []c.IndexInst,
	respCh MsgChannel, stopCh StopChannel) {

	addrs, err := k.getAllProjectorAddrs()
	if err != nil {
		logging.Errorf("KVSender::addIndexForExistingBucket %v %v Error in fetching cluster info %v",
			streamId, bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}
		return
	}

	var currentTs *protobuf.TsVbuuid
	protoInstList := convertIndexListToProto(k.config, k.cInfoCache, indexInstList, streamId)
	topic := getTopicForStreamId(streamId)

	fn := func(r int, err error) error {

		//clear the error before every retry
		err = nil
		for _, addr := range addrs {
			execWithStopCh(func() {
				ap := newProjClient(addr)
				if res, ret := sendAddInstancesRequest(ap, topic, protoInstList); ret != nil {
					logging.Errorf("KVSender::addIndexForExistingBucket %v %v Error Received %v from %v",
						streamId, bucket, ret, addr)
					err = ret
				} else {
					currentTs = updateCurrentTsFromResponse(bucket, currentTs, res)
				}
			}, stopCh)
		}

		//check if we have received currentTs for all vbuckets
		numVbuckets := k.config["numVbuckets"].Int()
		if currentTs == nil || currentTs.Len() != numVbuckets {
			return errors.New("ErrPartialVbStart")
		} else {
			return err
		}

	}

	rh := c.NewRetryHelper(MAX_KV_REQUEST_RETRY, time.Second, BACKOFF_FACTOR, fn)
	err = rh.Run()
	if err != nil {
		logging.Errorf("KVSender::addIndexForExistingBucket %v %v Error Received %v",
			streamId, bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}
		return
	}

	numVbuckets := k.config["numVbuckets"].Int()
	nativeTs := currentTs.ToTsVbuuid(numVbuckets)

	respCh <- &MsgStreamUpdate{mType: MSG_SUCCESS,
		streamId:  streamId,
		bucket:    bucket,
		restartTs: nativeTs}
}
コード例 #15
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func (k *kvSender) restartVbuckets(streamId c.StreamId, restartTs *c.TsVbuuid,
	connErrVbs []Vbucket, respCh MsgChannel, stopCh StopChannel) {

	addrs, err := k.getProjAddrsForVbuckets(restartTs.Bucket, restartTs.GetVbnos())
	if err != nil {
		logging.Errorf("KVSender::restartVbuckets %v %v Error in fetching cluster info %v",
			streamId, restartTs.Bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}

		return
	}

	//convert TS to protobuf format
	var protoRestartTs *protobuf.TsVbuuid
	numVbuckets := k.config["numVbuckets"].Int()
	protoTs := protobuf.NewTsVbuuid(DEFAULT_POOL, restartTs.Bucket, numVbuckets)
	protoRestartTs = protoTs.FromTsVbuuid(restartTs)

	var rollbackTs *protobuf.TsVbuuid
	topic := getTopicForStreamId(streamId)
	rollback := false

	fn := func(r int, err error) error {

		for _, addr := range addrs {
			ap := newProjClient(addr)

			if res, ret := k.sendRestartVbuckets(ap, topic, connErrVbs, protoRestartTs); ret != nil {
				//retry for all errors
				logging.Errorf("KVSender::restartVbuckets %v %v Error Received %v from %v",
					streamId, restartTs.Bucket, ret, addr)
				err = ret
			} else {
				rollbackTs = updateRollbackTsFromResponse(restartTs.Bucket, rollbackTs, res)
			}
		}

		if rollbackTs != nil && checkVbListInTS(protoRestartTs.GetVbnos(), rollbackTs) {
			//if rollback, no need to retry
			rollback = true
			return nil
		} else {
			return err
		}
	}

	rh := c.NewRetryHelper(MAX_KV_REQUEST_RETRY, time.Second, BACKOFF_FACTOR, fn)
	err = rh.Run()

	//if any of the requested vb is in rollback ts, send rollback
	//msg to caller
	if rollback {
		//convert from protobuf to native format
		nativeTs := rollbackTs.ToTsVbuuid(numVbuckets)

		respCh <- &MsgRollback{streamId: streamId,
			rollbackTs: nativeTs}
	} else if err != nil {
		//if there is a topicMissing/genServer.Closed error, a fresh
		//MutationTopicRequest is required.
		if err.Error() == projClient.ErrorTopicMissing.Error() ||
			err.Error() == c.ErrorClosed.Error() ||
			err.Error() == projClient.ErrorInvalidBucket.Error() {
			respCh <- &MsgKVStreamRepair{
				streamId: streamId,
				bucket:   restartTs.Bucket,
			}
		} else {
			respCh <- &MsgError{
				err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
					severity: FATAL,
					cause:    err}}

		}
	} else {
		respCh <- &MsgSuccess{}
	}
}
コード例 #16
0
ファイル: kv_sender.go プロジェクト: jchris/indexing
func (k *kvSender) openMutationStream(streamId c.StreamId, indexInstList []c.IndexInst,
	restartTs *c.TsVbuuid, respCh MsgChannel, stopCh StopChannel) {

	if len(indexInstList) == 0 {
		logging.Warnf("KVSender::openMutationStream Empty IndexList. Nothing to do.")
		respCh <- &MsgSuccess{}
		return
	}

	protoInstList := convertIndexListToProto(k.config, k.cInfoCache, indexInstList, streamId)
	bucket := indexInstList[0].Defn.Bucket

	//use any bucket as list of vbs remain the same for all buckets
	vbnos, err := k.getAllVbucketsInCluster(bucket)
	if err != nil {
		logging.Errorf("KVSender::openMutationStream %v %v Error in fetching vbuckets info %v",
			streamId, restartTs.Bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}
		return
	}

	restartTsList, err := k.makeRestartTsForVbs(bucket, restartTs, vbnos)
	if err != nil {
		logging.Errorf("KVSender::openMutationStream %v %v Error making restart ts %v",
			streamId, bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}
		return
	}

	addrs, err := k.getAllProjectorAddrs()
	if err != nil {
		logging.Errorf("KVSender::openMutationStream %v %v Error Fetching Projector Addrs %v",
			streamId, bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}
		return
	}

	var rollbackTs *protobuf.TsVbuuid
	var activeTs *protobuf.TsVbuuid
	topic := getTopicForStreamId(streamId)

	fn := func(r int, err error) error {

		//clear the error before every retry
		err = nil
		for _, addr := range addrs {

			execWithStopCh(func() {
				ap := newProjClient(addr)
				if res, ret := k.sendMutationTopicRequest(ap, topic, restartTsList, protoInstList); ret != nil {
					//for all errors, retry
					logging.Errorf("KVSender::openMutationStream %v %v Error Received %v from %v",
						streamId, bucket, ret, addr)
					err = ret
				} else {
					activeTs = updateActiveTsFromResponse(bucket, activeTs, res)
					rollbackTs = updateRollbackTsFromResponse(bucket, rollbackTs, res)
				}
			}, stopCh)
		}

		if rollbackTs != nil {
			//no retry required for rollback
			return nil
		} else if err != nil {
			//retry for any error
			return err
		} else {
			//check if we have received activeTs for all vbuckets
			retry := false
			if activeTs == nil || activeTs.Len() != len(vbnos) {
				retry = true
			}

			if retry {
				return errors.New("ErrPartialVbStart")
			} else {
				return nil
			}

		}
	}

	rh := c.NewRetryHelper(MAX_KV_REQUEST_RETRY, time.Second, BACKOFF_FACTOR, fn)
	err = rh.Run()

	if rollbackTs != nil {
		logging.Infof("KVSender::openMutationStream %v %v Rollback Received %v",
			streamId, bucket, rollbackTs)
		//convert from protobuf to native format
		numVbuckets := k.config["numVbuckets"].Int()
		nativeTs := rollbackTs.ToTsVbuuid(numVbuckets)
		respCh <- &MsgRollback{streamId: streamId,
			bucket:     bucket,
			rollbackTs: nativeTs}
	} else if err != nil {
		logging.Errorf("KVSender::openMutationStream %v %v Error Received %v",
			streamId, bucket, err)
		respCh <- &MsgError{
			err: Error{code: ERROR_KVSENDER_STREAM_REQUEST_ERROR,
				severity: FATAL,
				cause:    err}}
	} else {
		numVbuckets := k.config["numVbuckets"].Int()
		respCh <- &MsgSuccessOpenStream{activeTs: activeTs.ToTsVbuuid(numVbuckets)}
	}
}