dist_sender.go

// Copyright 2014 The Cockroach Authors.
//
// Licensed 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 kv

import (
	"context"
	"fmt"
	"sync/atomic"
	"unsafe"

	"github.com/cockroachdb/cockroach/pkg/base"
	"github.com/cockroachdb/cockroach/pkg/gossip"
	"github.com/cockroachdb/cockroach/pkg/internal/client"
	"github.com/cockroachdb/cockroach/pkg/keys"
	"github.com/cockroachdb/cockroach/pkg/roachpb"
	"github.com/cockroachdb/cockroach/pkg/rpc"
	"github.com/cockroachdb/cockroach/pkg/rpc/nodedialer"
	"github.com/cockroachdb/cockroach/pkg/settings"
	"github.com/cockroachdb/cockroach/pkg/settings/cluster"
	"github.com/cockroachdb/cockroach/pkg/util/grpcutil"
	"github.com/cockroachdb/cockroach/pkg/util/hlc"
	"github.com/cockroachdb/cockroach/pkg/util/log"
	"github.com/cockroachdb/cockroach/pkg/util/metric"
	"github.com/cockroachdb/cockroach/pkg/util/retry"
	"github.com/cockroachdb/cockroach/pkg/util/tracing"
	"github.com/pkg/errors"
)

const (
	// The default limit for asynchronous senders.
	defaultSenderConcurrency = 500
	// The maximum number of range descriptors to prefetch during range lookups.
	rangeLookupPrefetchCount = 8
)

var (
	metaDistSenderBatchCount = metric.Metadata{
		Name:        "distsender.batches",
		Help:        "Number of batches processed",
		Measurement: "Batches",
		Unit:        metric.Unit_COUNT,
	}
	metaDistSenderPartialBatchCount = metric.Metadata{
		Name:        "distsender.batches.partial",
		Help:        "Number of partial batches processed after being divided on range boundaries",
		Measurement: "Partial Batches",
		Unit:        metric.Unit_COUNT,
	}
	metaDistSenderAsyncSentCount = metric.Metadata{
		Name:        "distsender.batches.async.sent",
		Help:        "Number of partial batches sent asynchronously",
		Measurement: "Partial Batches",
		Unit:        metric.Unit_COUNT,
	}
	metaDistSenderAsyncThrottledCount = metric.Metadata{
		Name:        "distsender.batches.async.throttled",
		Help:        "Number of partial batches not sent asynchronously due to throttling",
		Measurement: "Partial Batches",
		Unit:        metric.Unit_COUNT,
	}
	metaTransportSentCount = metric.Metadata{
		Name:        "distsender.rpc.sent",
		Help:        "Number of RPCs sent",
		Measurement: "RPCs",
		Unit:        metric.Unit_COUNT,
	}
	metaTransportLocalSentCount = metric.Metadata{
		Name:        "distsender.rpc.sent.local",
		Help:        "Number of local RPCs sent",
		Measurement: "RPCs",
		Unit:        metric.Unit_COUNT,
	}
	metaTransportSenderNextReplicaErrCount = metric.Metadata{
		Name:        "distsender.rpc.sent.nextreplicaerror",
		Help:        "Number of RPCs sent due to per-replica errors",
		Measurement: "RPCs",
		Unit:        metric.Unit_COUNT,
	}
	metaDistSenderNotLeaseHolderErrCount = metric.Metadata{
		Name:        "distsender.errors.notleaseholder",
		Help:        "Number of NotLeaseHolderErrors encountered",
		Measurement: "Errors",
		Unit:        metric.Unit_COUNT,
	}
	metaDistSenderInLeaseTransferBackoffsCount = metric.Metadata{
		Name:        "distsender.errors.inleasetransferbackoffs",
		Help:        "Number of times backed off due to NotLeaseHolderErrors during lease transfer.",
		Measurement: "Errors",
		Unit:        metric.Unit_COUNT,
	}
)

var rangeDescriptorCacheSize = settings.RegisterIntSetting(
	"kv.range_descriptor_cache.size",
	"maximum number of entries in the range descriptor and leaseholder caches",
	1e6,
)

// DistSenderMetrics is the set of metrics for a given distributed sender.
type DistSenderMetrics struct {
	BatchCount              *metric.Counter
	PartialBatchCount       *metric.Counter
	AsyncSentCount          *metric.Counter
	AsyncThrottledCount     *metric.Counter
	SentCount               *metric.Counter
	LocalSentCount          *metric.Counter
	NextReplicaErrCount     *metric.Counter
	NotLeaseHolderErrCount  *metric.Counter
	InLeaseTransferBackoffs *metric.Counter
}

func makeDistSenderMetrics() DistSenderMetrics {
	return DistSenderMetrics{
		BatchCount:              metric.NewCounter(metaDistSenderBatchCount),
		PartialBatchCount:       metric.NewCounter(metaDistSenderPartialBatchCount),
		AsyncSentCount:          metric.NewCounter(metaDistSenderAsyncSentCount),
		AsyncThrottledCount:     metric.NewCounter(metaDistSenderAsyncThrottledCount),
		SentCount:               metric.NewCounter(metaTransportSentCount),
		LocalSentCount:          metric.NewCounter(metaTransportLocalSentCount),
		NextReplicaErrCount:     metric.NewCounter(metaTransportSenderNextReplicaErrCount),
		NotLeaseHolderErrCount:  metric.NewCounter(metaDistSenderNotLeaseHolderErrCount),
		InLeaseTransferBackoffs: metric.NewCounter(metaDistSenderInLeaseTransferBackoffsCount),
	}
}

// A firstRangeMissingError indicates that the first range has not yet
// been gossiped. This will be the case for a node which hasn't yet
// joined the gossip network.
type firstRangeMissingError struct{}

// Error is part of the error interface.
func (f firstRangeMissingError) Error() string {
	return "the descriptor for the first range is not available via gossip"
}

// A DistSender provides methods to access Cockroach's monolithic,
// distributed key value store. Each method invocation triggers a
// lookup or lookups to find replica metadata for implicated key
// ranges. RPCs are sent to one or more of the replicas to satisfy
// the method invocation.
type DistSender struct {
	log.AmbientContext

	st *cluster.Settings
	// nodeDescriptor, if set, holds the descriptor of the node the
	// DistSender lives on. It should be accessed via getNodeDescriptor(),
	// which tries to obtain the value from the Gossip network if the
	// descriptor is unknown.
	nodeDescriptor unsafe.Pointer
	// clock is used to set time for some calls. E.g. read-only ops
	// which span ranges and don't require read consistency.
	clock *hlc.Clock
	// gossip provides up-to-date information about the start of the
	// key range, used to find the replica metadata for arbitrary key
	// ranges.
	gossip  *gossip.Gossip
	metrics DistSenderMetrics
	// rangeCache caches replica metadata for key ranges.
	rangeCache *RangeDescriptorCache
	// leaseHolderCache caches range lease holders by range ID.
	leaseHolderCache *LeaseHolderCache
	transportFactory TransportFactory
	rpcContext       *rpc.Context
	nodeDialer       *nodedialer.Dialer
	rpcRetryOptions  retry.Options
	asyncSenderSem   chan struct{}

	// disableFirstRangeUpdates disables updates of the first range via
	// gossip. Used by tests which want finer control of the contents of the
	// range cache.
	disableFirstRangeUpdates int32
}

var _ client.Sender = &DistSender{}

// DistSenderConfig holds configuration and auxiliary objects that can be passed
// to NewDistSender.
type DistSenderConfig struct {
	AmbientCtx log.AmbientContext

	Settings        *cluster.Settings
	Clock           *hlc.Clock
	RPCRetryOptions *retry.Options
	// nodeDescriptor, if provided, is used to describe which node the DistSender
	// lives on, for instance when deciding where to send RPCs.
	// Usually it is filled in from the Gossip network on demand.
	nodeDescriptor    *roachpb.NodeDescriptor
	RPCContext        *rpc.Context
	RangeDescriptorDB RangeDescriptorDB

	NodeDialer *nodedialer.Dialer

	TestingKnobs ClientTestingKnobs
}

// NewDistSender returns a batch.Sender instance which connects to the
// Cockroach cluster via the supplied gossip instance. Supplying a
// DistSenderContext or the fields within is optional. For omitted values, sane
// defaults will be used.
func NewDistSender(cfg DistSenderConfig, g *gossip.Gossip) *DistSender {
	ds := &DistSender{
		st:         cfg.Settings,
		clock:      cfg.Clock,
		gossip:     g,
		metrics:    makeDistSenderMetrics(),
		nodeDialer: cfg.NodeDialer,
	}
	if ds.st == nil {
		ds.st = cluster.MakeTestingClusterSettings()
	}

	ds.AmbientContext = cfg.AmbientCtx
	if ds.AmbientContext.Tracer == nil {
		panic("no tracer set in AmbientCtx")
	}

	if cfg.nodeDescriptor != nil {
		atomic.StorePointer(&ds.nodeDescriptor, unsafe.Pointer(cfg.nodeDescriptor))
	}
	rdb := cfg.RangeDescriptorDB
	if rdb == nil {
		rdb = ds
	}
	getRangeDescCacheSize := func() int64 {
		return rangeDescriptorCacheSize.Get(&ds.st.SV)
	}
	ds.rangeCache = NewRangeDescriptorCache(ds.st, rdb, getRangeDescCacheSize)
	ds.leaseHolderCache = NewLeaseHolderCache(getRangeDescCacheSize)
	if tf := cfg.TestingKnobs.TransportFactory; tf != nil {
		ds.transportFactory = tf
	} else {
		ds.transportFactory = GRPCTransportFactory
	}
	ds.rpcRetryOptions = base.DefaultRetryOptions()
	if cfg.RPCRetryOptions != nil {
		ds.rpcRetryOptions = *cfg.RPCRetryOptions
	}
	if cfg.RPCContext != nil {
		ds.rpcContext = cfg.RPCContext
		if ds.rpcRetryOptions.Closer == nil {
			ds.rpcRetryOptions.Closer = ds.rpcContext.Stopper.ShouldQuiesce()
		}
	}
	ds.nodeDialer = cfg.NodeDialer
	ds.asyncSenderSem = make(chan struct{}, defaultSenderConcurrency)

	if g != nil {
		ctx := ds.AnnotateCtx(context.Background())
		g.RegisterCallback(gossip.KeyFirstRangeDescriptor,
			func(_ string, value roachpb.Value) {
				if atomic.LoadInt32(&ds.disableFirstRangeUpdates) == 1 {
					return
				}
				if log.V(1) {
					var desc roachpb.RangeDescriptor
					if err := value.GetProto(&desc); err != nil {
						log.Errorf(ctx, "unable to parse gossiped first range descriptor: %s", err)
					} else {
						log.Infof(ctx, "gossiped first range descriptor: %+v", desc.Replicas)
					}
				}
				err := ds.rangeCache.EvictCachedRangeDescriptor(ctx, roachpb.RKeyMin, nil, false)
				if err != nil {
					log.Warningf(ctx, "failed to evict first range descriptor: %s", err)
				}
			})
	}
	return ds
}

// DisableFirstRangeUpdates disables updates of the first range via
// gossip. Used by tests which want finer control of the contents of the range
// cache.
func (ds *DistSender) DisableFirstRangeUpdates() {
	atomic.StoreInt32(&ds.disableFirstRangeUpdates, 1)
}

// Metrics returns a struct which contains metrics related to the distributed
// sender's activity.
func (ds *DistSender) Metrics() DistSenderMetrics {
	return ds.metrics
}

// RangeDescriptorCache gives access to the DistSender's range cache.
func (ds *DistSender) RangeDescriptorCache() *RangeDescriptorCache {
	return ds.rangeCache
}

// LeaseHolderCache gives access to the DistSender's lease cache.
func (ds *DistSender) LeaseHolderCache() *LeaseHolderCache {
	return ds.leaseHolderCache
}

// RangeLookup implements the RangeDescriptorDB interface. It uses LookupRange
// to perform a lookup scan for the provided key, using DistSender itself as the
// client.Sender. This means that the scan will recurse into DistSender, which
// will in turn use the RangeDescriptorCache again to lookup the RangeDescriptor
// necessary to perform the scan.
func (ds *DistSender) RangeLookup(
	ctx context.Context, key roachpb.RKey, useReverseScan bool,
) ([]roachpb.RangeDescriptor, []roachpb.RangeDescriptor, error) {
	// We perform the range lookup scan with a READ_UNCOMMITTED consistency
	// level because we want the scan to return intents as well as committed
	// values. The reason for this is because it's not clear whether the intent
	// or the previous value points to the correct location of the Range. It
	// gets even more complicated when there are split-related intents or a txn
	// record co-located with a replica involved in the split. Since we cannot
	// know the correct answer, we lookup both the pre- and post- transaction
	// values.
	rc := roachpb.READ_UNCOMMITTED
	// By using DistSender as the sender, we guarantee that even if the desired
	// RangeDescriptor is not on the first range we send the lookup too, we'll
	// still find it when we scan to the next range. This addresses the issue
	// described in #18032 and #16266, allowing us to support meta2 splits.
	return client.RangeLookup(ctx, ds, key.AsRawKey(), rc, rangeLookupPrefetchCount, useReverseScan)
}

// FirstRange implements the RangeDescriptorDB interface.
// FirstRange returns the RangeDescriptor for the first range on the cluster,
// which is retrieved from the gossip protocol instead of the datastore.
func (ds *DistSender) FirstRange() (*roachpb.RangeDescriptor, error) {
	if ds.gossip == nil {
		panic("with `nil` Gossip, DistSender must not use itself as rangeDescriptorDB")
	}
	rangeDesc := &roachpb.RangeDescriptor{}
	if err := ds.gossip.GetInfoProto(gossip.KeyFirstRangeDescriptor, rangeDesc); err != nil {
		return nil, firstRangeMissingError{}
	}
	return rangeDesc, nil
}

// getNodeDescriptor returns ds.nodeDescriptor, but makes an attempt to load
// it from the Gossip network if a nil value is found.
// We must jump through hoops here to get the node descriptor because it's not available
// until after the node has joined the gossip network and been allowed to initialize
// its stores.
func (ds *DistSender) getNodeDescriptor() *roachpb.NodeDescriptor {
	if desc := atomic.LoadPointer(&ds.nodeDescriptor); desc != nil {
		return (*roachpb.NodeDescriptor)(desc)
	}
	if ds.gossip == nil {
		return nil
	}

	ownNodeID := ds.gossip.NodeID.Get()
	if ownNodeID > 0 {
		// TODO(tschottdorf): Consider instead adding the NodeID of the
		// coordinator to the header, so we can get this from incoming
		// requests. Just in case we want to mostly eliminate gossip here.
		nodeDesc := &roachpb.NodeDescriptor{}
		if err := ds.gossip.GetInfoProto(gossip.MakeNodeIDKey(ownNodeID), nodeDesc); err == nil {
			atomic.StorePointer(&ds.nodeDescriptor, unsafe.Pointer(nodeDesc))
			return nodeDesc
		}
	}
	if log.V(1) {
		ctx := ds.AnnotateCtx(context.TODO())
		log.Infof(ctx, "unable to determine this node's attributes for replica "+
			"selection; node is most likely bootstrapping")
	}
	return nil
}

// sendRPC sends one or more RPCs to replicas from the supplied
// roachpb.Replica slice. Returns an RPC error if the request could
// not be sent. Note that the reply may contain a higher level error
// and must be checked in addition to the RPC error.
//
// The replicas are assumed to be ordered by preference, with closer
// ones (i.e. expected lowest latency) first.
func (ds *DistSender) sendRPC(
	ctx context.Context,
	rangeID roachpb.RangeID,
	replicas ReplicaSlice,
	ba roachpb.BatchRequest,
	cachedLeaseHolder roachpb.ReplicaDescriptor,
) (*roachpb.BatchResponse, error) {
	if len(replicas) == 0 {
		return nil, roachpb.NewSendError(
			fmt.Sprintf("no replica node addresses available via gossip for r%d", rangeID))
	}

	ba.RangeID = rangeID

	tracing.AnnotateTrace()
	defer tracing.AnnotateTrace()

	return ds.sendToReplicas(
		ctx,
		SendOptions{metrics: &ds.metrics},
		rangeID,
		replicas,
		ba,
		ds.nodeDialer,
		cachedLeaseHolder,
	)
}

// CountRanges returns the number of ranges that encompass the given key span.
func (ds *DistSender) CountRanges(ctx context.Context, rs roachpb.RSpan) (int64, error) {
	var count int64
	ri := NewRangeIterator(ds)
	for ri.Seek(ctx, rs.Key, Ascending); ri.Valid(); ri.Next(ctx) {
		count++
		if !ri.NeedAnother(rs) {
			break
		}
	}
	return count, ri.Error().GoError()
}

// getDescriptor looks up the range descriptor to use for a query of
// the key descKey with the given options. The lookup takes into
// consideration the last range descriptor that the caller had used
// for this key span, if any, and if the last range descriptor has
// been evicted because it was found to be stale, which is all managed
// through the EvictionToken. The function should be provided with an
// EvictionToken if one was acquired from this function on a previous
// call. If not, an empty EvictionToken can be provided.
//
// The range descriptor which contains the range in which the request should
// start its query is returned first. Next returned is an EvictionToken. In
// case the descriptor is discovered stale, the returned EvictionToken's evict
// method should be called; it evicts the cache appropriately.
func (ds *DistSender) getDescriptor(
	ctx context.Context, descKey roachpb.RKey, evictToken *EvictionToken, useReverseScan bool,
) (*roachpb.RangeDescriptor, *EvictionToken, error) {
	desc, returnToken, err := ds.rangeCache.LookupRangeDescriptor(
		ctx, descKey, evictToken, useReverseScan,
	)
	if err != nil {
		return nil, returnToken, err
	}

	return desc, returnToken, nil
}

// sendSingleRange gathers and rearranges the replicas, and makes an RPC call.
func (ds *DistSender) sendSingleRange(
	ctx context.Context, ba roachpb.BatchRequest, desc *roachpb.RangeDescriptor,
) (*roachpb.BatchResponse, *roachpb.Error) {
	// Try to send the call.
	replicas := NewReplicaSlice(ds.gossip, desc)

	// If this request needs to go to a lease holder and we know who that is, move
	// it to the front.
	var cachedLeaseHolder roachpb.ReplicaDescriptor
	if ba.RequiresLeaseHolder() {
		if storeID, ok := ds.leaseHolderCache.Lookup(ctx, desc.RangeID); ok {
			if i := replicas.FindReplica(storeID); i >= 0 {
				replicas.MoveToFront(i)
				cachedLeaseHolder = replicas[0].ReplicaDescriptor
			}
		}
	}
	if (cachedLeaseHolder == roachpb.ReplicaDescriptor{}) {
		// Rearrange the replicas so that they're ordered in expectation of
		// request latency.
		var latencyFn LatencyFunc
		if ds.rpcContext != nil {
			latencyFn = ds.rpcContext.RemoteClocks.Latency
		}
		replicas.OptimizeReplicaOrder(ds.getNodeDescriptor(), latencyFn)
	}

	br, err := ds.sendRPC(ctx, desc.RangeID, replicas, ba, cachedLeaseHolder)
	if err != nil {
		log.VErrEvent(ctx, 2, err.Error())
		return nil, roachpb.NewError(err)
	}

	// If the reply contains a timestamp, update the local HLC with it.
	if br.Error != nil && br.Error.Now != (hlc.Timestamp{}) {
		ds.clock.Update(br.Error.Now)
	} else if br.Now != (hlc.Timestamp{}) {
		ds.clock.Update(br.Now)
	}

	// Untangle the error from the received response.
	pErr := br.Error
	br.Error = nil // scrub the response error
	return br, pErr
}

// initAndVerifyBatch initializes timestamp-related information and
// verifies batch constraints before splitting.
func (ds *DistSender) initAndVerifyBatch(
	ctx context.Context, ba *roachpb.BatchRequest,
) *roachpb.Error {
	// Attach the local node ID to each request.
	if ba.Header.GatewayNodeID == 0 && ds.gossip != nil {
		ba.Header.GatewayNodeID = ds.gossip.NodeID.Get()
	}

	// In the event that timestamp isn't set and read consistency isn't
	// required, set the timestamp using the local clock.
	if ba.ReadConsistency != roachpb.CONSISTENT && ba.Timestamp == (hlc.Timestamp{}) {
		ba.Timestamp = ds.clock.Now()
	}

	if len(ba.Requests) < 1 {
		return roachpb.NewErrorf("empty batch")
	}

	if ba.MaxSpanRequestKeys != 0 {
		// Verify that the batch contains only specific range requests or the
		// Begin/EndTransactionRequest. Verify that a batch with a ReverseScan
		// only contains ReverseScan range requests.
		isReverse := ba.IsReverse()
		for _, req := range ba.Requests {
			inner := req.GetInner()
			switch inner.(type) {
			case *roachpb.ScanRequest, *roachpb.DeleteRangeRequest:
				// Accepted range requests. All other range requests are still
				// not supported. Note that ReverseScanRequest is _not_ handled here.
				// TODO(vivek): don't enumerate all range requests.
				if isReverse {
					return roachpb.NewErrorf("batch with limit contains both forward and reverse scans")
				}

			case *roachpb.QueryIntentRequest, *roachpb.ResolveIntentRangeRequest:
				continue

			case *roachpb.BeginTransactionRequest, *roachpb.EndTransactionRequest, *roachpb.ReverseScanRequest:
				continue

			default:
				return roachpb.NewErrorf("batch with limit contains %T request", inner)
			}
		}
	}

	// If ScanOptions is set the batch is only allowed to contain scans.
	if ba.ScanOptions != nil {
		for _, req := range ba.Requests {
			switch req.GetInner().(type) {
			case *roachpb.ScanRequest, *roachpb.ReverseScanRequest:
				// Scans are supported.
			case *roachpb.BeginTransactionRequest, *roachpb.EndTransactionRequest:
				// These requests are ignored.
			default:
				return roachpb.NewErrorf("batch with scan option has non-scans: %s", ba)
			}
		}
		// If both MaxSpanRequestKeys and MinResults are set, then they can't be
		// contradictory.
		if ba.Header.MaxSpanRequestKeys != 0 &&
			ba.Header.MaxSpanRequestKeys < ba.Header.ScanOptions.MinResults {
			return roachpb.NewErrorf("MaxSpanRequestKeys (%d) < MinResults (%d): %s",
				ba.Header.MaxSpanRequestKeys, ba.Header.ScanOptions.MinResults, ba)
		}
	}

	// Make sure that MVCCScan requests aren't in batch form if our cluster
	// version is too old.
	// TODO(jordan): delete this stanza after 2.1 is released.
	if !ds.st.Version.IsActive(cluster.VersionBatchResponse) {
		for i := range ba.Requests {
			switch req := ba.Requests[i].GetInner().(type) {
			case *roachpb.ScanRequest:
				req.ScanFormat = roachpb.KEY_VALUES
			case *roachpb.ReverseScanRequest:
				req.ScanFormat = roachpb.KEY_VALUES
			}
		}
	}
	return nil
}

// errNo1PCTxn indicates that a batch cannot be sent as a 1 phase
// commit because it spans multiple ranges and must be split into at
// least two parts, with the final part containing the EndTransaction
// request.
var errNo1PCTxn = roachpb.NewErrorf("cannot send 1PC txn to multiple ranges")

// splitBatchAndCheckForRefreshSpans splits the batch according to the
// canSplitET parameter and checks whether the final request is an
// EndTransaction. If so, the EndTransactionRequest.NoRefreshSpans
// flag is reset to indicate whether earlier parts of the split may
// result in refresh spans.
func splitBatchAndCheckForRefreshSpans(
	ba roachpb.BatchRequest, canSplitET bool,
) [][]roachpb.RequestUnion {
	parts := ba.Split(canSplitET)
	// If the final part contains an EndTransaction, we need to check
	// whether earlier split parts contain any refresh spans and properly
	// set the NoRefreshSpans flag on the end transaction.
	lastPart := parts[len(parts)-1]
	lastReq := lastPart[len(lastPart)-1].GetInner()
	if et, ok := lastReq.(*roachpb.EndTransactionRequest); ok && et.NoRefreshSpans {
		hasRefreshSpans := false
		for _, part := range parts[:len(parts)-1] {
			for _, req := range part {
				if roachpb.NeedsRefresh(req.GetInner()) {
					hasRefreshSpans = true
				}
			}
		}
		if hasRefreshSpans {
			etCopy := *et
			etCopy.NoRefreshSpans = false
			lastPart = append([]roachpb.RequestUnion(nil), lastPart...)
			lastPart[len(lastPart)-1].MustSetInner(&etCopy)
			parts[len(parts)-1] = lastPart
		}
	}
	return parts
}

// Send implements the batch.Sender interface. It subdivides the Batch
// into batches admissible for sending (preventing certain illegal
// mixtures of requests), executes each individual part (which may
// span multiple ranges), and recombines the response.
//
// When the request spans ranges, it is split by range and a partial
// subset of the batch request is sent to affected ranges in parallel.
//
// The first write in a transaction may not arrive before writes to
// other ranges. This is relevant in the case of a BeginTransaction
// request. Intents written to other ranges before the transaction
// record is created will cause the transaction to abort early.
//
// Note that on error, this method will return any batch responses for
// successfully processed batch requests. This allows the caller to
// deal with potential retry situations where a batch is split so that
// EndTransaction is processed alone, after earlier requests in the
// batch succeeded. Where possible, the caller may be able to update
// spans encountered in the transaction and retry just the
// EndTransaction request to avoid client-side serializable txn retries.
func (ds *DistSender) Send(
	ctx context.Context, ba roachpb.BatchRequest,
) (*roachpb.BatchResponse, *roachpb.Error) {
	ds.metrics.BatchCount.Inc(1)

	tracing.AnnotateTrace()

	if pErr := ds.initAndVerifyBatch(ctx, &ba); pErr != nil {
		return nil, pErr
	}

	ctx = ds.AnnotateCtx(ctx)
	ctx, sp := tracing.EnsureChildSpan(ctx, ds.AmbientContext.Tracer, "dist sender send")
	defer sp.Finish()

	var rplChunks []*roachpb.BatchResponse
	splitET := false
	var require1PC bool
	lastReq := ba.Requests[len(ba.Requests)-1].GetInner()
	if et, ok := lastReq.(*roachpb.EndTransactionRequest); ok && et.Require1PC {
		require1PC = true
	}
	// To ensure that we lay down intents to prevent starvation, always
	// split the end transaction request into its own batch on retries.
	// Txns requiring 1PC are an exception and should never be split.
	if ba.Txn != nil && ba.Txn.Epoch > 0 && !require1PC {
		splitET = true
	}
	parts := splitBatchAndCheckForRefreshSpans(ba, splitET)
	if len(parts) > 1 && ba.MaxSpanRequestKeys != 0 {
		// We already verified above that the batch contains only scan requests of the same type.
		// Such a batch should never need splitting.
		panic("batch with MaxSpanRequestKeys needs splitting")
	}

	var pErr *roachpb.Error
	errIdxOffset := 0
	for len(parts) > 0 {
		part := parts[0]
		ba.Requests = part
		// The minimal key range encompassing all requests contained within.
		// Local addressing has already been resolved.
		// TODO(tschottdorf): consider rudimentary validation of the batch here
		// (for example, non-range requests with EndKey, or empty key ranges).
		rs, err := keys.Range(ba)
		if err != nil {
			return nil, roachpb.NewError(err)
		}

		var rpl *roachpb.BatchResponse
		rpl, pErr = ds.divideAndSendBatchToRanges(ctx, ba, rs, 0 /* batchIdx */)

		if pErr == errNo1PCTxn {
			// If we tried to send a single round-trip EndTransaction but
			// it looks like it's going to hit multiple ranges, split it
			// here and try again.
			if len(parts) != 1 {
				panic("EndTransaction not in last chunk of batch")
			} else if require1PC {
				log.Fatalf(ctx, "required 1PC transaction cannot be split: %s", ba)
			}
			parts = splitBatchAndCheckForRefreshSpans(ba, true /* split ET */)
			// Restart transaction of the last chunk as multiple parts
			// with EndTransaction in the last part.
			continue
		}
		if pErr != nil {
			if pErr.Index != nil && pErr.Index.Index != -1 {
				pErr.Index.Index += int32(errIdxOffset)
			}
			// Break out of loop to collate batch responses received so far to
			// return with error.
			break
		}

		errIdxOffset += len(ba.Requests)

		// Propagate transaction from last reply to next request. The final
		// update is taken and put into the response's main header.
		ba.UpdateTxn(rpl.Txn)
		rplChunks = append(rplChunks, rpl)
		parts = parts[1:]
	}

	var reply *roachpb.BatchResponse
	if len(rplChunks) > 0 {
		reply = rplChunks[0]
		for _, rpl := range rplChunks[1:] {
			reply.Responses = append(reply.Responses, rpl.Responses...)
			reply.CollectedSpans = append(reply.CollectedSpans, rpl.CollectedSpans...)
		}
		lastHeader := rplChunks[len(rplChunks)-1].BatchResponse_Header
		lastHeader.CollectedSpans = reply.CollectedSpans
		reply.BatchResponse_Header = lastHeader
	}

	return reply, pErr
}

type response struct {
	reply     *roachpb.BatchResponse
	positions []int
	pErr      *roachpb.Error
}

// divideAndSendBatchToRanges sends the supplied batch to all of the
// ranges which comprise the span specified by rs. The batch request
// is trimmed against each range which is part of the span and sent
// either serially or in parallel, if possible. batchIdx indicates
// which partial fragment of the larger batch is being processed by
// this method. It's specified as non-zero when this method is invoked
// recursively.
func (ds *DistSender) divideAndSendBatchToRanges(
	ctx context.Context, ba roachpb.BatchRequest, rs roachpb.RSpan, batchIdx int,
) (br *roachpb.BatchResponse, pErr *roachpb.Error) {
	// Clone the BatchRequest's transaction so that future mutations to the
	// proto don't affect the proto in this batch.
	if ba.Txn != nil {
		txnCopy := *ba.Txn
		ba.Txn = &txnCopy
	}
	// Get initial seek key depending on direction of iteration.
	var scanDir ScanDirection
	var seekKey roachpb.RKey
	if !ba.IsReverse() {
		scanDir = Ascending
		seekKey = rs.Key
	} else {
		scanDir = Descending
		seekKey = rs.EndKey
	}
	ri := NewRangeIterator(ds)
	ri.Seek(ctx, seekKey, scanDir)
	if !ri.Valid() {
		return nil, ri.Error()
	}
	// Take the fast path if this batch fits within a single range.
	if !ri.NeedAnother(rs) {
		resp := ds.sendPartialBatch(ctx, ba, rs, ri.Desc(), ri.Token(), batchIdx, false /* needsTruncate */)
		return resp.reply, resp.pErr
	}

	if ba.IsUnsplittable() {
		mismatch := roachpb.NewRangeKeyMismatchError(rs.Key.AsRawKey(), rs.EndKey.AsRawKey(), ri.Desc())
		return nil, roachpb.NewError(mismatch)
	}

	// Make an empty slice of responses which will be populated with responses
	// as they come in via Combine().
	br = &roachpb.BatchResponse{
		Responses: make([]roachpb.ResponseUnion, len(ba.Requests)),
	}
	// This function builds a channel of responses for each range
	// implicated in the span (rs) and combines them into a single
	// BatchResponse when finished.
	var responseChs []chan response
	// couldHaveSkippedResponses is set if a ResumeSpan needs to be sent back.
	var couldHaveSkippedResponses bool
	// If couldHaveSkippedResponses is set, resumeReason indicates the reason why
	// the ResumeSpan is necessary. This reason is common to all individual
	// responses that carry a ResumeSpan.
	var resumeReason roachpb.ResponseHeader_ResumeReason
	defer func() {
		if r := recover(); r != nil {
			// If we're in the middle of a panic, don't wait on responseChs.
			panic(r)
		}
		var hadSuccessWriting bool
		// Combine all the responses.
		// It's important that we wait for all of them even if an error is caught
		// because the client.Sender() contract mandates that we don't "hold on" to
		// any part of a request after DistSender.Send() returns.
		for _, responseCh := range responseChs {
			resp := <-responseCh
			if resp.pErr != nil {
				if pErr == nil {
					pErr = resp.pErr
				}
				continue
			}
			if !hadSuccessWriting {
				for _, i := range resp.positions {
					req := ba.Requests[i].GetInner()
					if !roachpb.IsReadOnly(req) {
						hadSuccessWriting = true
						break
					}
				}
			}

			// Combine the new response with the existing one (including updating
			// the headers).
			if pErr == nil {
				if err := br.Combine(resp.reply, resp.positions); err != nil {
					pErr = roachpb.NewError(err)
				}
			}
		}

		// If we experienced an error, don't neglect to update the error's
		// attached transaction with any responses which were received.
		if pErr != nil {
			// The br.Txn != nil check looks unnecessary, but note that we might
			// be returning errNo1PCTxn which is a singleton, so we could end up
			// with data races.
			//
			// TODO(tschottdorf): get rid of the errNo1PCTxn singleton. It's
			// ugly.
			if br.Txn != nil {
				pErr.UpdateTxn(br.Txn)
			}
			// If this is a write batch with any successful responses, but
			// we're ultimately returning an error, wrap the error with a
			// MixedSuccessError.
			if hadSuccessWriting {
				// divideAndSendBatchToRanges can call sendPartialBatch, which in
				// turn can call divideAndSendBatchToRanges recursively. Therefore,
				// pErr can already be a MixedSuccessError returned from the
				// recursive call; do not wrap it in another MixedSuccessError.
				if _, ok := pErr.GetDetail().(*roachpb.MixedSuccessError); !ok {
					index := pErr.Index
					pErr = roachpb.NewError(&roachpb.MixedSuccessError{Wrapped: pErr})
					// Propagate the index to the MixedSuccessError. Note that the index
					// is shared by pointer, and when the index is modified as part of
					// the MixedSuccessError and the MixedSuccessError gets
					// unwrapped, the original error will contain the modified index.
					pErr.Index = index
				}
			}
		} else if couldHaveSkippedResponses {
			fillSkippedResponses(ba, br, seekKey, resumeReason)
		}
	}()

	stopAtRangeBoundary := ba.Header.ScanOptions != nil && ba.Header.ScanOptions.StopAtRangeBoundary
	// If min_results is set, num_results will count how many results scans have
	// accumulated so far.
	var numResults int64
	canParallelize := (ba.Header.MaxSpanRequestKeys == 0) && !stopAtRangeBoundary

	for ; ri.Valid(); ri.Seek(ctx, seekKey, scanDir) {
		responseCh := make(chan response, 1)
		responseChs = append(responseChs, responseCh)

		if batchIdx == 0 && ri.NeedAnother(rs) {
			// TODO(tschottdorf): we should have a mechanism for discovering
			// range merges (descriptor staleness will mostly go unnoticed),
			// or we'll be turning single-range queries into multi-range
			// queries for no good reason.
			//
			// If there's no transaction and op spans ranges, possibly
			// re-run as part of a transaction for consistency. The
			// case where we don't need to re-run is if the read
			// consistency is not required.
			if ba.Txn == nil && ba.IsPossibleTransaction() && ba.ReadConsistency == roachpb.CONSISTENT {
				responseCh <- response{pErr: roachpb.NewError(&roachpb.OpRequiresTxnError{})}
				return
			}
			// If the request is more than but ends with EndTransaction, we
			// want the caller to come again with the EndTransaction in an
			// extra call.
			if l := len(ba.Requests) - 1; l > 0 && ba.Requests[l].GetInner().Method() == roachpb.EndTransaction {
				responseCh <- response{pErr: errNo1PCTxn}
				return
			}
		}

		// Determine next seek key, taking a potentially sparse batch into
		// consideration.
		var err error
		nextRS := rs
		if scanDir == Descending {
			// In next iteration, query previous range.
			// We use the StartKey of the current descriptor as opposed to the
			// EndKey of the previous one since that doesn't have bugs when
			// stale descriptors come into play.
			seekKey, err = prev(ba, ri.Desc().StartKey)
			nextRS.EndKey = seekKey
		} else {
			// In next iteration, query next range.
			// It's important that we use the EndKey of the current descriptor
			// as opposed to the StartKey of the next one: if the former is stale,
			// it's possible that the next range has since merged the subsequent
			// one, and unless both descriptors are stale, the next descriptor's
			// StartKey would move us to the beginning of the current range,
			// resulting in a duplicate scan.
			seekKey, err = next(ba, ri.Desc().EndKey)
			nextRS.Key = seekKey
		}
		if err != nil {
			responseCh <- response{pErr: roachpb.NewError(err)}
			return
		}

		lastRange := !ri.NeedAnother(rs)
		// Send the next partial batch to the first range in the "rs" span.
		// If we can reserve one of the limited goroutines available for parallel
		// batch RPCs, send asynchronously.
		if canParallelize && !lastRange && ds.rpcContext != nil &&
			ds.sendPartialBatchAsync(ctx, ba, rs, ri.Desc(), ri.Token(), batchIdx, responseCh) {
			// Sent the batch asynchronously.
		} else {
			resp := ds.sendPartialBatch(ctx, ba, rs, ri.Desc(), ri.Token(), batchIdx, true /* needsTruncate */)
			responseCh <- resp
			if resp.pErr != nil {
				return
			}
			// Update the transaction from the response. Note that this wouldn't happen
			// on the asynchronous path, but if we have newer information it's good to
			// use it.
			ba.UpdateTxn(resp.reply.Txn)

			mightStopEarly := ba.MaxSpanRequestKeys > 0 || stopAtRangeBoundary
			// Check whether we've received enough responses to exit query loop.
			if mightStopEarly {
				var replyResults int64
				for _, r := range resp.reply.Responses {
					replyResults += r.GetInner().Header().NumKeys
				}
				// Do accounting for results. It's important that we update
				// MaxSpanRequestKeys and ScanOptions.MinResults, as ba might be
				// passed recursively to further divideAndSendBatchToRanges() calls.
				numResults += replyResults
				if ba.MaxSpanRequestKeys > 0 {
					if replyResults > ba.MaxSpanRequestKeys {
						log.Fatalf(ctx, "received %d results, limit was %d",
							replyResults, ba.MaxSpanRequestKeys)
					}
					ba.MaxSpanRequestKeys -= replyResults
					// Exiting; any missing responses will be filled in via defer().
					if ba.MaxSpanRequestKeys == 0 {
						couldHaveSkippedResponses = true
						resumeReason = roachpb.RESUME_KEY_LIMIT
						return
					}
				}
				var minResultsSatisfied bool
				if !stopAtRangeBoundary {
					minResultsSatisfied = true
				} else {
					if ba.Header.ScanOptions.MinResults == 0 {
						minResultsSatisfied = true
					} else {
						// We need to change ba.Header.ScanOptions, so we have to make a
						// copy so as to not mutate the one that we have already passed to
						// gRPC.
						scanOptsCopy := *ba.Header.ScanOptions
						scanOptsCopy.MinResults -= numResults
						minResultsSatisfied = scanOptsCopy.MinResults <= 0
						ba.Header.ScanOptions = &scanOptsCopy
					}
				}
				// If stopAtRangeBoundary is set, we stop unless MinResults is not
				// satisfied.
				if stopAtRangeBoundary && minResultsSatisfied {
					couldHaveSkippedResponses = true
					resumeReason = roachpb.RESUME_RANGE_BOUNDARY
					return
				}
			}
		}

		// The iteration is complete if the iterator's current range
		// encompasses the remaining span, OR if the next span has
		// inverted. This can happen if this method is invoked
		// re-entrantly due to ranges being split or merged. In that case
		// the batch request has all the original requests but the span is
		// a sub-span of the original, causing next() and prev() methods
		// to potentially return values which invert the span.
		if lastRange || !nextRS.Key.Less(nextRS.EndKey) {
			return
		}
		batchIdx++
		rs = nextRS
	}

	// We've exited early. Return the range iterator error.
	responseCh := make(chan response, 1)
	responseCh <- response{pErr: ri.Error()}
	responseChs = append(responseChs, responseCh)
	return
}

// sendPartialBatchAsync sends the partial batch asynchronously if
// there aren't currently more than the allowed number of concurrent
// async requests outstanding. Returns whether the partial batch was
// sent.
func (ds *DistSender) sendPartialBatchAsync(
	ctx context.Context,
	ba roachpb.BatchRequest,
	rs roachpb.RSpan,
	desc *roachpb.RangeDescriptor,
	evictToken *EvictionToken,
	batchIdx int,
	responseCh chan response,
) bool {
	if err := ds.rpcContext.Stopper.RunLimitedAsyncTask(
		ctx, "kv.DistSender: sending partial batch",
		ds.asyncSenderSem, false, /* wait */
		func(ctx context.Context) {
			ds.metrics.AsyncSentCount.Inc(1)
			responseCh <- ds.sendPartialBatch(ctx, ba, rs, desc, evictToken, batchIdx, true /* needsTruncate */)
		},
	); err != nil {
		ds.metrics.AsyncThrottledCount.Inc(1)
		return false
	}
	return true
}

// sendPartialBatch sends the supplied batch to the range specified by
// desc. The batch request is first truncated so that it contains only
// requests which intersect the range descriptor and keys for each
// request are limited to the range's key span. The send occurs in a
// retry loop to handle send failures. On failure to send to any
// replicas, we backoff and retry by refetching the range
// descriptor. If the underlying range seems to have split, we
// recursively invoke divideAndSendBatchToRanges to re-enumerate the
// ranges in the span and resend to each. If needsTruncate is true,
// the supplied batch and span must be truncated to the supplied range
// descriptor.
func (ds *DistSender) sendPartialBatch(
	ctx context.Context,
	ba roachpb.BatchRequest,
	rs roachpb.RSpan,
	desc *roachpb.RangeDescriptor,
	evictToken *EvictionToken,
	batchIdx int,
	needsTruncate bool,
) response {
	if batchIdx == 1 {
		ds.metrics.PartialBatchCount.Inc(2) // account for first batch
	} else if batchIdx > 1 {
		ds.metrics.PartialBatchCount.Inc(1)
	}
	var reply *roachpb.BatchResponse
	var pErr *roachpb.Error
	var err error
	var positions []int

	isReverse := ba.IsReverse()

	if needsTruncate {
		// Truncate the request to range descriptor.
		rs, err = rs.Intersect(desc)
		if err != nil {
			return response{pErr: roachpb.NewError(err)}
		}
		ba, positions, err = truncate(ba, rs)
		if len(positions) == 0 && err == nil {
			// This shouldn't happen in the wild, but some tests exercise it.
			return response{
				pErr: roachpb.NewErrorf("truncation resulted in empty batch on %s: %s", rs, ba),
			}
		}
		if err != nil {
			return response{pErr: roachpb.NewError(err)}
		}
	}

	// Start a retry loop for sending the batch to the range.
	for r := retry.StartWithCtx(ctx, ds.rpcRetryOptions); r.Next(); {
		// If we've cleared the descriptor on a send failure, re-lookup.
		if desc == nil {
			var descKey roachpb.RKey
			if isReverse {
				descKey = rs.EndKey
			} else {
				descKey = rs.Key
			}
			// TODO(nvanbenschoten): shouldn't we be passing an eviction token
			// here from the previous iteration? See #28967.
			desc, evictToken, err = ds.getDescriptor(ctx, descKey, nil, isReverse)
			if err != nil {
				log.VErrEventf(ctx, 1, "range descriptor re-lookup failed: %s", err)
				continue
			}
		}

		reply, pErr = ds.sendSingleRange(ctx, ba, desc)

		// If sending succeeded, return immediately.
		if pErr == nil {
			return response{reply: reply, positions: positions}
		}

		// Re-map the error index within this partial batch back
		// to its position in the encompassing batch.
		if pErr.Index != nil && pErr.Index.Index != -1 && positions != nil {
			pErr.Index.Index = int32(positions[pErr.Index.Index])
		}

		log.VErrEventf(ctx, 2, "reply error %s: %s", ba, pErr)

		// Error handling: If the error indicates that our range
		// descriptor is out of date, evict it from the cache and try
		// again. Errors that apply only to a single replica were
		// handled in send().
		//
		// TODO(bdarnell): Don't retry endlessly. If we fail twice in a
		// row and the range descriptor hasn't changed, return the error
		// to our caller.
		switch tErr := pErr.GetDetail().(type) {
		case *roachpb.SendError:
			// We've tried all the replicas without success. Either
			// they're all down, or we're using an out-of-date range
			// descriptor. Invalidate the cache and try again with the new
			// metadata.
			log.VEventf(ctx, 1, "evicting range descriptor on %T and backoff for re-lookup: %+v", tErr, desc)
			if err := evictToken.Evict(ctx); err != nil {
				return response{pErr: roachpb.NewError(err)}
			}
			// Clear the descriptor to reload on the next attempt.
			desc = nil
			continue
		case *roachpb.RangeKeyMismatchError:
			// Range descriptor might be out of date - evict it. This is
			// likely the result of a range split. If we have new range
			// descriptors, insert them instead as long as they are different
			// from the last descriptor to avoid endless loops.
			var replacements []roachpb.RangeDescriptor
			different := func(rd *roachpb.RangeDescriptor) bool {
				return !desc.RSpan().Equal(rd.RSpan())
			}
			if tErr.MismatchedRange != nil && different(tErr.MismatchedRange) {
				replacements = append(replacements, *tErr.MismatchedRange)
			}
			if tErr.SuggestedRange != nil && different(tErr.SuggestedRange) {
				if includesFrontOfCurSpan(isReverse, tErr.SuggestedRange, rs) {
					replacements = append(replacements, *tErr.SuggestedRange)
				}
			}
			// Same as Evict() if replacements is empty.
			if err := evictToken.EvictAndReplace(ctx, replacements...); err != nil {
				return response{pErr: roachpb.NewError(err)}
			}
			// On addressing errors (likely a split), we need to re-invoke
			// the range descriptor lookup machinery, so we recurse by
			// sending batch to just the partial span this descriptor was
			// supposed to cover. Note that for the resending, we use the
			// already truncated batch, so that we know that the response
			// to it matches the positions into our batch (using the full
			// batch here would give a potentially larger response slice
			// with unknown mapping to our truncated reply).
			log.VEventf(ctx, 1, "likely split; resending batch to span: %s", tErr)
			reply, pErr = ds.divideAndSendBatchToRanges(ctx, ba, rs, batchIdx)
			return response{reply: reply, positions: positions, pErr: pErr}
		}
		break
	}

	// Propagate error if either the retry closer or context done
	// channels were closed.
	if pErr == nil {
		if pErr = ds.deduceRetryEarlyExitError(ctx); pErr == nil {
			log.Fatal(ctx, "exited retry loop without an error")
		}
	}

	return response{pErr: pErr}
}

func (ds *DistSender) deduceRetryEarlyExitError(ctx context.Context) *roachpb.Error {
	select {
	case <-ds.rpcRetryOptions.Closer:
		// Typically happens during shutdown.
		return roachpb.NewError(&roachpb.NodeUnavailableError{})
	case <-ctx.Done():
		// Happens when the client request is canceled.
		return roachpb.NewError(errors.Wrap(ctx.Err(), "aborted in distSender"))
	default:
	}
	return nil
}

func includesFrontOfCurSpan(isReverse bool, rd *roachpb.RangeDescriptor, rs roachpb.RSpan) bool {
	if isReverse {
		return rd.ContainsKeyInverted(rs.EndKey)
	}
	return rd.ContainsKey(rs.Key)
}

// fillSkippedResponses fills in responses and ResumeSpans for requests
// when a batch finished without fully processing the requested key spans for
// (some of) the requests in the batch. This can happen when processing has met
// the batch key max limit for range requests, or some other stop condition
// based on ScanOptions.
//
// nextKey is the first key that was not processed. This will be used when
// filling up the ResumeSpan's.
func fillSkippedResponses(
	ba roachpb.BatchRequest,
	br *roachpb.BatchResponse,
	nextKey roachpb.RKey,
	resumeReason roachpb.ResponseHeader_ResumeReason,
) {
	// Some requests might have no response at all if we used a batch-wide
	// limit; simply create trivial responses for those. Note that any type
	// of request can crop up here - simply take a batch that exceeds the
	// limit, and add any other requests at higher keys at the end of the
	// batch -- they'll all come back without any response since they never
	// execute.
	var scratchBA roachpb.BatchRequest
	for i := range br.Responses {
		if br.Responses[i] != (roachpb.ResponseUnion{}) {
			continue
		}
		req := ba.Requests[i].GetInner()
		// We need to summon an empty response. The most convenient (but not
		// most efficient) way is to use (*BatchRequest).CreateReply.
		//
		// TODO(tschottdorf): can autogenerate CreateReply for individual
		// requests, see roachpb/gen_batch.go.
		if scratchBA.Requests == nil {
			scratchBA.Requests = make([]roachpb.RequestUnion, 1)
		}
		scratchBA.Requests[0].MustSetInner(req)
		br.Responses[i] = scratchBA.CreateReply().Responses[0]
	}
	// Set the ResumeSpan for future batch requests.
	isReverse := ba.IsReverse()
	for i, resp := range br.Responses {
		req := ba.Requests[i].GetInner()
		if !roachpb.IsRange(req) {
			continue
		}
		hdr := resp.GetInner().Header()
		hdr.ResumeReason = resumeReason
		origSpan := req.Header().Span()
		if isReverse {
			if hdr.ResumeSpan != nil {
				// The ResumeSpan.Key might be set to the StartKey of a range;
				// correctly set it to the Key of the original request span.
				hdr.ResumeSpan.Key = origSpan.Key
			} else if roachpb.RKey(origSpan.Key).Less(nextKey) {
				// Some keys have yet to be processed.
				hdr.ResumeSpan = &origSpan
				if nextKey.Less(roachpb.RKey(origSpan.EndKey)) {
					// The original span has been partially processed.
					hdr.ResumeSpan.EndKey = nextKey.AsRawKey()
				}
			}
		} else {
			if hdr.ResumeSpan != nil {
				// The ResumeSpan.EndKey might be set to the EndKey of a range because
				// that's what a store will set it to when the limit is reached; it
				// doesn't know any better). In that case, we correct it to the EndKey
				// of the original request span. Note that this doesn't touch
				// ResumeSpan.Key, which is really the important part of the ResumeSpan.
				hdr.ResumeSpan.EndKey = origSpan.EndKey
			} else {
				// The request might have been fully satisfied, in which case it doesn't
				// need a ResumeSpan, or it might not have. Figure out if we're in the
				// latter case.
				if nextKey.Less(roachpb.RKey(origSpan.EndKey)) {
					// Some keys have yet to be processed.
					hdr.ResumeSpan = &origSpan
					if roachpb.RKey(origSpan.Key).Less(nextKey) {
						// The original span has been partially processed.
						hdr.ResumeSpan.Key = nextKey.AsRawKey()
					}
				}
			}
		}
		br.Responses[i].GetInner().SetHeader(hdr)
	}
}

// sendToReplicas sends one or more RPCs to clients specified by the
// slice of replicas. On success, Send returns the first successful
// reply. If an error occurs which is not specific to a single
// replica, it's returned immediately. Otherwise, when all replicas
// have been tried and failed, returns a send error.
func (ds *DistSender) sendToReplicas(
	ctx context.Context,
	opts SendOptions,
	rangeID roachpb.RangeID,
	replicas ReplicaSlice,
	ba roachpb.BatchRequest,
	nodeDialer *nodedialer.Dialer,
	cachedLeaseHolder roachpb.ReplicaDescriptor,
) (*roachpb.BatchResponse, error) {
	var ambiguousError error
	var haveCommit bool
	// We only check for committed txns, not aborts because aborts may
	// be retried without any risk of inconsistencies.
	if etArg, ok := ba.GetArg(roachpb.EndTransaction); ok {
		haveCommit = etArg.(*roachpb.EndTransactionRequest).Commit
	}

	transport, err := ds.transportFactory(opts, nodeDialer, replicas)
	if err != nil {
		return nil, err
	}
	if transport.IsExhausted() {
		return nil, roachpb.NewSendError(
			fmt.Sprintf("sending to all %d replicas failed", len(replicas)))
	}

	curReplica := transport.NextReplica()
	if log.ExpensiveLogEnabled(ctx, 2) {
		log.VEventf(ctx, 2, "r%d: sending batch %s to %s", rangeID, ba.Summary(), curReplica)
	}
	br, err := transport.SendNext(ctx, ba)

	// maxSeenLeaseSequence tracks the maximum LeaseSequence seen in a
	// NotLeaseHolderError. If we encounter a sequence number less than or equal
	// to maxSeenLeaseSequence number in a subsequent NotLeaseHolderError then
	// the range must be experiencing a least transfer and the client should back
	// off using inTransferRetry.
	maxSeenLeaseSequence := roachpb.LeaseSequence(-1)
	inTransferRetry := retry.StartWithCtx(ctx, ds.rpcRetryOptions)
	inTransferRetry.Next() // The first call to Next does not block.

	// This loop will retry operations that fail with errors that reflect
	// per-replica state and may succeed on other replicas.
	for {
		if err != nil {
			// For most connection errors, we cannot tell whether or not
			// the request may have succeeded on the remote server, so we
			// set the ambiguous commit flag (exceptions are captured in
			// the grpcutil.RequestDidNotStart function).
			//
			// We retry ambiguous commit batches to avoid returning the
			// unrecoverable AmbiguousResultError. This is safe because
			// repeating an already-successfully applied batch is
			// guaranteed to return an error. If the original attempt merely timed out
			// or was lost, then the batch will succeed and we can be assured the
			// commit was applied just once.
			if haveCommit && !grpcutil.RequestDidNotStart(err) {
				ambiguousError = err
			}
			log.VErrEventf(ctx, 2, "RPC error: %s", err)

			// If the error wasn't just a context cancellation and the down replica
			// is cached as the lease holder, evict it. The only other eviction
			// happens below on NotLeaseHolderError, but if the next replica is the
			// actual lease holder, we're never going to receive one of those and
			// will thus pay the price of trying the down node first forever.
			//
			// NB: we should consider instead adding a successful reply from the next
			// replica into the cache, but without a leaseholder (and taking into
			// account that the local node can't be down) it won't take long until we
			// talk to a replica that tells us who the leaseholder is.
			if ctx.Err() == nil {
				if storeID, ok := ds.leaseHolderCache.Lookup(ctx, rangeID); ok && curReplica.StoreID == storeID {
					ds.leaseHolderCache.Update(ctx, rangeID, 0 /* evict */)
				}
			}
		} else {
			// NB: This section of code may have unfortunate performance implications. If we
			// exit the below type switch with propagateError remaining at `false`, we'll try
			// more replicas. That may succeed and future requests might do the same thing over
			// and over again, adding needless round-trips to the earlier replicas.
			propagateError := false
			switch tErr := br.Error.GetDetail().(type) {
			case nil:
				// When a request that we know could only succeed on the leaseholder comes
				// back as successful, make sure the leaseholder cache reflects this
				// replica. In steady state, this is almost always the case, and so we
				// gate the update on whether the response comes from a node that we didn't
				// know held the lease.
				if cachedLeaseHolder != curReplica && ba.RequiresLeaseHolder() {
					ds.leaseHolderCache.Update(ctx, rangeID, curReplica.StoreID)
				}
				return br, nil
			case *roachpb.StoreNotFoundError, *roachpb.NodeUnavailableError:
				// These errors are likely to be unique to the replica that reported
				// them, so no action is required before the next retry.
			case *roachpb.RangeNotFoundError:
				// The store we routed to doesn't have this replica. This can happen when
				// our descriptor is outright outdated, but it can also be caused by a
				// replica that has just been added but needs a snapshot to be caught up.
				//
				// We'll try other replicas which typically gives us the leaseholder, either
				// via the NotLeaseHolderError or nil error paths, both of which update the
				// leaseholder cache.
			case *roachpb.NotLeaseHolderError:
				ds.metrics.NotLeaseHolderErrCount.Inc(1)
				if lh := tErr.LeaseHolder; lh != nil {
					// Update the leaseholder cache. Naively this would also happen when the
					// next RPC comes back, but we don't want to wait out the additional RPC
					// latency.
					ds.leaseHolderCache.Update(ctx, rangeID, lh.StoreID)
					// Avoid an extra update to the leaseholder cache if the next RPC succeeds.
					cachedLeaseHolder = *lh

					// If the implicated leaseholder is not a known replica, return a SendError
					// to signal eviction of the cached RangeDescriptor and re-send.
					if replicas.FindReplica(lh.StoreID) == -1 {
						br.Error = roachpb.NewError(roachpb.NewSendError(fmt.Sprintf(
							"leaseholder s%d (via %+v) not in cached replicas %v", lh.StoreID, curReplica, replicas,
						)))
						propagateError = true
					} else {
						// Move the new lease holder to the head of the queue for the next retry.
						transport.MoveToFront(*lh)
					}
				}
				if l := tErr.Lease; !propagateError && l != nil {
					// Check whether we've seen this lease or a prior lease before and
					// backoff if so or update maxSeenLeaseSequence if not.
					if l.Sequence > maxSeenLeaseSequence {
						maxSeenLeaseSequence = l.Sequence
						inTransferRetry.Reset() // The following Next call will not block.
					} else {
						ds.metrics.InLeaseTransferBackoffs.Inc(1)
						log.VErrEventf(ctx, 2, "backing off due to NotLeaseHolderErr at "+
							"LeaseSequence %d <= %d", l.Sequence, maxSeenLeaseSequence)
					}
					inTransferRetry.Next()
				}
			default:
				propagateError = true
			}

			if propagateError {
				if ambiguousError != nil {
					return nil, roachpb.NewAmbiguousResultError(fmt.Sprintf("error=%s [propagate]", ambiguousError))
				}

				// The error received is likely not specific to this
				// replica, so we should return it instead of trying other
				// replicas.
				return br, nil
			}

			log.VErrEventf(ctx, 1, "application error: %s", br.Error)
		}

		// Has the caller given up?
		if ctx.Err() != nil {
			errMsg := fmt.Sprintf("context done during DistSender.Send: %s", ctx.Err())
			log.Eventf(ctx, errMsg)
			if ambiguousError != nil {
				return nil, roachpb.NewAmbiguousResultError(errMsg)
			}
			// Don't consider this a SendError, because SendErrors indicate that we
			// were unable to reach a replica that could serve the request, and they
			// cause range cache evictions. Context cancellations just mean the
			// sender changed its mind or the request timed out.
			return nil, errors.Wrap(ctx.Err(), "aborted during DistSender.Send")
		}

		if transport.IsExhausted() {
			if ambiguousError != nil {
				return nil, roachpb.NewAmbiguousResultError(fmt.Sprintf("error=%s [exhausted]", ambiguousError))
			}

			// TODO(bdarnell): The last error is not necessarily the best
			// one to return; we may want to remember the "best" error
			// we've seen (for example, a NotLeaseHolderError conveys more
			// information than a RangeNotFound).
			return nil, roachpb.NewSendError(
				fmt.Sprintf("sending to all %d replicas failed; last error: %v %v", len(replicas), br, err),
			)
		}

		ds.metrics.NextReplicaErrCount.Inc(1)
		curReplica = transport.NextReplica()
		log.VEventf(ctx, 2, "error: %v %v; trying next peer %s", br, err, curReplica)
		br, err = transport.SendNext(ctx, ba)
	}
}