Merge #94692

94692: kvserver: use MaxInflightBytes raft config option r=erikgrinaker,tbg a=pavelkalinnikov

This change starts using the raft's `MaxInflightBytes` option which limits the inflight total byte size of the log entries sent via `MsgApp`. The default limit is set to a conservatively large value, to not change the current behaviour for now.

Part of #90314
Epic: none
Release note (ops change): Added COCKROACH_RAFT_MAX_INFLIGHT_BYTES env variable which helps strictly limiting inflight traffic from a Raft leader to followers, particularly in situations when many large messages are sent and significantly exceed COCKROACH_RAFT_MAX_SIZE_PER_MSG * COCKROACH_RAFT_MAX_INFLIGHT_MSGS which is a softer limit.

Co-authored-by: Pavel Kalinnikov <[email protected]>

pkg/base/config.go

@@ -12,6 +12,8 @@ package base
 
 import (
 	"context"
+	"math"
+	"math/big"
 	"net/url"
 	"os"
 	"time"
@@ -252,6 +254,24 @@ var (
 	// will send to a given follower without hearing a response.
 	defaultRaftMaxInflightMsgs = envutil.EnvOrDefaultInt(
 		"COCKROACH_RAFT_MAX_INFLIGHT_MSGS", 128)
+
+	// defaultRaftMaxInflightBytes specifies the maximum aggregate byte size of
+	// Raft log entries that a leader will send to a follower without hearing
+	// responses.
+	//
+	// Previously it was assumed that RaftMaxInflightMsgs * RaftMaxSizePerMsg is a
+	// proxy for the actual max inflight traffic. However, RaftMaxSizePerMsg is
+	// not a hard limit, it's rather a "target" size for the message, so the
+	// actual inflight bytes could exceed this product by a large factor.
+	// RaftMaxInflightBytes is a more accurate limit, and should be used in
+	// conjunction with the two.
+	//
+	// TODO(#90314): lower this limit to something close to max rates observed in
+	// healthy clusters. Currently, this is a conservatively large multiple of
+	// defaultRaftMaxInflightMsgs * defaultRaftMaxSizePerMsg, so that we don't
+	// abruptly break the previous assumption and cut off traffic.
+	defaultRaftMaxInflightBytes = envutil.EnvOrDefaultBytes(
+		"COCKROACH_RAFT_MAX_INFLIGHT_BYTES", 256<<20 /* 256 MB */)
 )
 
 // Config is embedded by server.Config. A base config is not meant to be used
@@ -469,25 +489,45 @@ type RaftConfig struct {
 	// value lowers the raft recovery cost (during initial probing and after
 	// message loss during normal operation). On the other hand, it limits the
 	// throughput during normal replication.
+	//
+	// Used in combination with RaftMaxInflightMsgs and RaftMaxInflightBytes.
 	RaftMaxSizePerMsg uint64
 
 	// RaftMaxCommittedSizePerReady controls the maximum aggregate byte size of
 	// committed Raft log entries a replica will receive in a single Ready.
 	RaftMaxCommittedSizePerReady uint64
 
-	// RaftMaxInflightMsgs controls how many "inflight" MsgApps Raft will send
-	// to a follower without hearing a response. The total number of Raft log
-	// entries is a combination of this setting and RaftMaxSizePerMsg. The
-	// current default settings provide for up to 4 MB of raft log to be sent
-	// without acknowledgement. With an average entry size of 1 KB that
-	// translates to ~4096 commands that might be executed in the handling of a
-	// single raft.Ready operation.
+	// RaftMaxInflightMsgs controls how many "inflight" MsgApps Raft will send to
+	// a follower without hearing a response. The total size of inflight Raft log
+	// entries is thus roughly limited by RaftMaxInflightMsgs * RaftMaxSizePerMsg,
+	// but also by RaftMaxInflightBytes. The current default settings provide for
+	// up to 4 MB of Raft log to be sent without acknowledgement. With an average
+	// entry size of 1 KB that translates to ~4096 commands that might be executed
+	// in the handling of a single raft.Ready operation.
 	//
 	// This setting is used both by sending and receiving end of Raft messages. To
 	// minimize dropped messages on the receiver, its size should at least match
 	// the sender's (being it the default size, or taken from the env variables).
 	RaftMaxInflightMsgs int
 
+	// RaftMaxInflightBytes controls the maximum aggregate byte size of Raft log
+	// entries that a leader will send to a follower without hearing responses.
+	//
+	// Normally RaftMaxSizePerMsg * RaftMaxInflightMsgs is the actual limit. But
+	// the RaftMaxSizePerMsg is soft, and Raft may send individual messages
+	// arbitrarily larger than that (e.g. with a large write, or AddSST command),
+	// so it's possible that the overall limit is exceeded by a large multiple.
+	// RaftMaxInflightBytes is a stricter limit which can only be slightly
+	// exceeded (by a single message).
+	//
+	// This effectively bounds the bandwidth-delay product. Note that especially
+	// in high-latency deployments setting this too low can lead to a dramatic
+	// reduction in throughput. For example, with a peer that has a round-trip
+	// latency of 100ms to the leader and this setting is set to 1 MB, there is a
+	// throughput limit of 10 MB/s for this group. With RTT of 400ms, this drops
+	// to 2.5 MB/s. See Little's law to understand the maths behind.
+	RaftMaxInflightBytes uint64
+
 	// Splitting a range which has a replica needing a snapshot results in two
 	// ranges in that state. The delay configured here slows down splits when in
 	// that situation (limiting to those splits not run through the split
@@ -545,6 +585,17 @@ func (cfg *RaftConfig) SetDefaults() {
 	if cfg.RaftMaxInflightMsgs == 0 {
 		cfg.RaftMaxInflightMsgs = defaultRaftMaxInflightMsgs
 	}
+
+	if cfg.RaftMaxInflightBytes == 0 {
+		cfg.RaftMaxInflightBytes = uint64(defaultRaftMaxInflightBytes)
+	}
+	// Fixup RaftMaxInflightBytes if it is lower than reasonable.
+	if other := maxInflightBytesFrom(
+		cfg.RaftMaxInflightMsgs, cfg.RaftMaxSizePerMsg,
+	); cfg.RaftMaxInflightBytes < other {
+		cfg.RaftMaxInflightBytes = other
+	}
+
 	if cfg.RaftDelaySplitToSuppressSnapshotTicks == 0 {
 		// The Raft Ticks interval defaults to 200ms, and an election is 10
 		// ticks. Add a generous amount of ticks to make sure even a backed up
@@ -641,6 +692,19 @@ func DefaultRetryOptions() retry.Options {
 	}
 }
 
+// maxInflightBytesFrom returns the minimal value for RaftMaxInflightBytes
+// config option based on RaftMaxInflightMsgs and RaftMaxSizePerMsg.
+func maxInflightBytesFrom(maxInflightMsgs int, maxSizePerMsg uint64) uint64 {
+	// Compute min(maxInflightMsgs * maxSizePerMsg, MaxUint64) safely.
+	if mul := new(big.Int).Mul(
+		big.NewInt(int64(maxInflightMsgs)),
+		new(big.Int).SetUint64(maxSizePerMsg),
+	); mul.IsUint64() {
+		return mul.Uint64()
+	}
+	return math.MaxUint64
+}
+
 // StorageConfig contains storage configs for all storage engine.
 type StorageConfig struct {
 	Attrs roachpb.Attributes

pkg/base/config_test.go

@@ -12,6 +12,7 @@ package base_test
 
 import (
 	"fmt"
+	"math"
 	"testing"
 	"time"
 
@@ -20,6 +21,7 @@ import (
 	"github.com/cockroachdb/cockroach/pkg/testutils/echotest"
 	"github.com/cockroachdb/cockroach/pkg/util/leaktest"
 	"github.com/davecgh/go-spew/spew"
+	"github.com/stretchr/testify/require"
 )
 
 func TestDefaultRaftConfig(t *testing.T) {
@@ -122,3 +124,36 @@ func TestDefaultRaftConfig(t *testing.T) {
 
 	echotest.Require(t, s, datapathutils.TestDataPath(t, "raft_config_recovery"))
 }
+
+func TestRaftMaxInflightBytes(t *testing.T) {
+	defer leaktest.AfterTest(t)()
+	for i, tc := range []struct {
+		msgSize uint64
+		maxMsgs int
+		maxInfl uint64
+		want    uint64
+	}{
+		// If any of these tests fail, sync the corresponding default values with
+		// config.go, and update the comments that reason about default values.
+		{want: 256 << 20},                    // assert 255 MB is still default
+		{maxMsgs: 128, want: 256 << 20},      // assert 128 is still default
+		{msgSize: 32 << 10, want: 256 << 20}, // assert 32 KB is still default
+
+		{maxMsgs: 1 << 30, want: 1 << 45}, // large maxMsgs
+		{msgSize: 1 << 50, want: 1 << 57}, // large msgSize
+
+		{msgSize: 100, maxMsgs: 10, maxInfl: 1000000, want: 1000000}, // reasonable
+		{msgSize: 100, maxMsgs: 10, maxInfl: 5, want: 1000},          // fixup applied
+		{msgSize: 1 << 50, maxMsgs: 1 << 20, want: math.MaxUint64},   // overflow
+	} {
+		t.Run(fmt.Sprintf("%d", i), func(t *testing.T) {
+			cfg := base.RaftConfig{
+				RaftMaxInflightMsgs:  tc.maxMsgs,
+				RaftMaxSizePerMsg:    tc.msgSize,
+				RaftMaxInflightBytes: tc.maxInfl,
+			}
+			cfg.SetDefaults()
+			require.Equal(t, tc.want, cfg.RaftMaxInflightBytes)
+		})
+	}
+}

pkg/base/testdata/raft_config

@@ -12,6 +12,7 @@ echo
  RaftMaxSizePerMsg: (uint64) 32768,
  RaftMaxCommittedSizePerReady: (uint64) 67108864,
  RaftMaxInflightMsgs: (int) 128,
+ RaftMaxInflightBytes: (uint64) 268435456,
  RaftDelaySplitToSuppressSnapshotTicks: (int) 230
 }
 RaftHeartbeatInterval: 1s

pkg/kv/kvserver/store.go

@@ -289,6 +289,7 @@ func newRaftConfig(
 		MaxCommittedSizePerReady:  storeCfg.RaftMaxCommittedSizePerReady,
 		MaxSizePerMsg:             storeCfg.RaftMaxSizePerMsg,
 		MaxInflightMsgs:           storeCfg.RaftMaxInflightMsgs,
+		MaxInflightBytes:          storeCfg.RaftMaxInflightBytes,
 		Storage:                   strg,
 		Logger:                    logger,