Merge pull request #68147 from timothysc/etcd-3.2-latest

staging/src/k8s.io/kube-aggregator/Godeps/Godeps.json

@@ -36,51 +36,51 @@
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/auth/authpb",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/client",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/clientv3",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/etcdserver/etcdserverpb",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/mvcc/mvccpb",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/pathutil",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/srv",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/tlsutil",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/transport",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/types",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/version",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/go-semver/semver",

staging/src/k8s.io/sample-apiserver/Godeps/Godeps.json

@@ -36,51 +36,51 @@
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/auth/authpb",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/client",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/clientv3",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/etcdserver/etcdserverpb",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/mvcc/mvccpb",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/pathutil",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/srv",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/tlsutil",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/transport",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/pkg/types",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/etcd/version",
-			"Rev": "95a726a27e09030f9ccbd9982a1508f5a6d25ada"
+			"Rev": "420a452267a7ce45b3fcbed04d54030d69964fc1"
 		},
 		{
 			"ImportPath": "github.com/coreos/go-semver/semver",

vendor/github.com/coreos/etcd/auth/simple_token.go

@@ -118,6 +118,11 @@ func (t *tokenSimple) genTokenPrefix() (string, error) {
 
 func (t *tokenSimple) assignSimpleTokenToUser(username, token string) {
 	t.simpleTokensMu.Lock()
+	defer t.simpleTokensMu.Unlock()
+	if t.simpleTokenKeeper == nil {
+		return
+	}
+
 	_, ok := t.simpleTokens[token]
 	if ok {
 		plog.Panicf("token %s is alredy used", token)
@@ -125,7 +130,6 @@ func (t *tokenSimple) assignSimpleTokenToUser(username, token string) {
 
 	t.simpleTokens[token] = username
 	t.simpleTokenKeeper.addSimpleToken(token)
-	t.simpleTokensMu.Unlock()
 }
 
 func (t *tokenSimple) invalidateUser(username string) {

vendor/github.com/coreos/etcd/clientv3/BUILD

@@ -10,13 +10,13 @@ go_library(
         "compare.go",
         "config.go",
         "doc.go",
-        "grpc_options.go",
         "health_balancer.go",
         "kv.go",
         "lease.go",
         "logger.go",
         "maintenance.go",
         "op.go",
+        "options.go",
         "ready_wait.go",
         "retry.go",
         "sort.go",
@@ -31,6 +31,7 @@ go_library(
         "//vendor/github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes:go_default_library",
         "//vendor/github.com/coreos/etcd/etcdserver/etcdserverpb:go_default_library",
         "//vendor/github.com/coreos/etcd/mvcc/mvccpb:go_default_library",
+        "//vendor/github.com/coreos/etcd/pkg/types:go_default_library",
         "//vendor/golang.org/x/net/context:go_default_library",
         "//vendor/google.golang.org/grpc:go_default_library",
         "//vendor/google.golang.org/grpc/codes:go_default_library",

vendor/github.com/coreos/etcd/clientv3/client.go

@@ -529,6 +529,20 @@ func isHaltErr(ctx context.Context, err error) bool {
 	return ev.Code() != codes.Unavailable && ev.Code() != codes.Internal
 }
 
+// isUnavailableErr returns true if the given error is an unavailable error
+func isUnavailableErr(ctx context.Context, err error) bool {
+	if ctx != nil && ctx.Err() != nil {
+		return false
+	}
+	if err == nil {
+		return false
+	}
+	ev, _ := status.FromError(err)
+	// Unavailable codes mean the system will be right back.
+	// (e.g., can't connect, lost leader)
+	return ev.Code() == codes.Unavailable
+}
+
 func toErr(ctx context.Context, err error) error {
 	if err == nil {
 		return nil

vendor/github.com/coreos/etcd/clientv3/cluster.go

@@ -16,6 +16,7 @@ package clientv3
 
 import (
 	pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
+	"github.com/coreos/etcd/pkg/types"
 
 	"golang.org/x/net/context"
 	"google.golang.org/grpc"
@@ -65,6 +66,11 @@ func NewClusterFromClusterClient(remote pb.ClusterClient, c *Client) Cluster {
 }
 
 func (c *cluster) MemberAdd(ctx context.Context, peerAddrs []string) (*MemberAddResponse, error) {
+	// fail-fast before panic in rafthttp
+	if _, err := types.NewURLs(peerAddrs); err != nil {
+		return nil, err
+	}
+
 	r := &pb.MemberAddRequest{PeerURLs: peerAddrs}
 	resp, err := c.remote.MemberAdd(ctx, r, c.callOpts...)
 	if err != nil {
@@ -83,6 +89,11 @@ func (c *cluster) MemberRemove(ctx context.Context, id uint64) (*MemberRemoveRes
 }
 
 func (c *cluster) MemberUpdate(ctx context.Context, id uint64, peerAddrs []string) (*MemberUpdateResponse, error) {
+	// fail-fast before panic in rafthttp
+	if _, err := types.NewURLs(peerAddrs); err != nil {
+		return nil, err
+	}
+
 	// it is safe to retry on update.
 	r := &pb.MemberUpdateRequest{ID: id, PeerURLs: peerAddrs}
 	resp, err := c.remote.MemberUpdate(ctx, r, c.callOpts...)

vendor/github.com/coreos/etcd/clientv3/lease.go

@@ -71,8 +71,6 @@ const (
 	// defaultTTL is the assumed lease TTL used for the first keepalive
 	// deadline before the actual TTL is known to the client.
 	defaultTTL = 5 * time.Second
-	// a small buffer to store unsent lease responses.
-	leaseResponseChSize = 16
 	// NoLease is a lease ID for the absence of a lease.
 	NoLease LeaseID = 0
 
@@ -80,6 +78,11 @@ const (
 	retryConnWait = 500 * time.Millisecond
 )
 
+// LeaseResponseChSize is the size of buffer to store unsent lease responses.
+// WARNING: DO NOT UPDATE.
+// Only for testing purposes.
+var LeaseResponseChSize = 16
+
 // ErrKeepAliveHalted is returned if client keep alive loop halts with an unexpected error.
 //
 // This usually means that automatic lease renewal via KeepAlive is broken, but KeepAliveOnce will still work as expected.
@@ -219,7 +222,7 @@ func (l *lessor) TimeToLive(ctx context.Context, id LeaseID, opts ...LeaseOption
 }
 
 func (l *lessor) KeepAlive(ctx context.Context, id LeaseID) (<-chan *LeaseKeepAliveResponse, error) {
-	ch := make(chan *LeaseKeepAliveResponse, leaseResponseChSize)
+	ch := make(chan *LeaseKeepAliveResponse, LeaseResponseChSize)
 
 	l.mu.Lock()
 	// ensure that recvKeepAliveLoop is still running
@@ -475,9 +478,10 @@ func (l *lessor) recvKeepAlive(resp *pb.LeaseKeepAliveResponse) {
 	for _, ch := range ka.chs {
 		select {
 		case ch <- karesp:
-			ka.nextKeepAlive = nextKeepAlive
 		default:
 		}
+		// still advance in order to rate-limit keep-alive sends
+		ka.nextKeepAlive = nextKeepAlive
 	}
 }
 

vendor/github.com/coreos/etcd/clientv3/grpc_options.go → vendor/github.com/coreos/etcd/clientv3/options.go

@@ -44,3 +44,6 @@ var (
 // Some options are exposed to "clientv3.Config".
 // Defaults will be overridden by the settings in "clientv3.Config".
 var defaultCallOpts = []grpc.CallOption{defaultFailFast, defaultMaxCallSendMsgSize, defaultMaxCallRecvMsgSize}
+
+// MaxLeaseTTL is the maximum lease TTL value
+const MaxLeaseTTL = 9000000000

vendor/github.com/coreos/etcd/clientv3/watch.go

@@ -769,10 +769,13 @@ func (w *watchGrpcStream) joinSubstreams() {
 	}
 }
 
+var maxBackoff = 100 * time.Millisecond
+
 // openWatchClient retries opening a watch client until success or halt.
 // manually retry in case "ws==nil && err==nil"
 // TODO: remove FailFast=false
 func (w *watchGrpcStream) openWatchClient() (ws pb.Watch_WatchClient, err error) {
+	backoff := time.Millisecond
 	for {
 		select {
 		case <-w.ctx.Done():
@@ -788,6 +791,17 @@ func (w *watchGrpcStream) openWatchClient() (ws pb.Watch_WatchClient, err error)
 		if isHaltErr(w.ctx, err) {
 			return nil, v3rpc.Error(err)
 		}
+		if isUnavailableErr(w.ctx, err) {
+			// retry, but backoff
+			if backoff < maxBackoff {
+				// 25% backoff factor
+				backoff = backoff + backoff/4
+				if backoff > maxBackoff {
+					backoff = maxBackoff
+				}
+			}
+			time.Sleep(backoff)
+		}
 	}
 	return ws, nil
 }

vendor/github.com/coreos/etcd/embed/BUILD

@@ -31,6 +31,7 @@ go_library(
         "//vendor/github.com/coreos/etcd/pkg/netutil:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/runtime:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/srv:go_default_library",
+        "//vendor/github.com/coreos/etcd/pkg/tlsutil:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/transport:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/types:go_default_library",
         "//vendor/github.com/coreos/etcd/rafthttp:go_default_library",

vendor/github.com/coreos/etcd/embed/config.go

@@ -20,6 +20,7 @@ import (
 	"net"
 	"net/http"
 	"net/url"
+	"path/filepath"
 	"strings"
 	"time"
 
@@ -27,6 +28,7 @@ import (
 	"github.com/coreos/etcd/pkg/cors"
 	"github.com/coreos/etcd/pkg/netutil"
 	"github.com/coreos/etcd/pkg/srv"
+	"github.com/coreos/etcd/pkg/tlsutil"
 	"github.com/coreos/etcd/pkg/transport"
 	"github.com/coreos/etcd/pkg/types"
 
@@ -87,8 +89,38 @@ type Config struct {
 	// TickMs is the number of milliseconds between heartbeat ticks.
 	// TODO: decouple tickMs and heartbeat tick (current heartbeat tick = 1).
 	// make ticks a cluster wide configuration.
-	TickMs            uint  `json:"heartbeat-interval"`
-	ElectionMs        uint  `json:"election-timeout"`
+	TickMs     uint `json:"heartbeat-interval"`
+	ElectionMs uint `json:"election-timeout"`
+
+	// InitialElectionTickAdvance is true, then local member fast-forwards
+	// election ticks to speed up "initial" leader election trigger. This
+	// benefits the case of larger election ticks. For instance, cross
+	// datacenter deployment may require longer election timeout of 10-second.
+	// If true, local node does not need wait up to 10-second. Instead,
+	// forwards its election ticks to 8-second, and have only 2-second left
+	// before leader election.
+	//
+	// Major assumptions are that:
+	//  - cluster has no active leader thus advancing ticks enables faster
+	//    leader election, or
+	//  - cluster already has an established leader, and rejoining follower
+	//    is likely to receive heartbeats from the leader after tick advance
+	//    and before election timeout.
+	//
+	// However, when network from leader to rejoining follower is congested,
+	// and the follower does not receive leader heartbeat within left election
+	// ticks, disruptive election has to happen thus affecting cluster
+	// availabilities.
+	//
+	// Disabling this would slow down initial bootstrap process for cross
+	// datacenter deployments. Make your own tradeoffs by configuring
+	// --initial-election-tick-advance at the cost of slow initial bootstrap.
+	//
+	// If single-node, it advances ticks regardless.
+	//
+	// See https://github.com/coreos/etcd/issues/9333 for more detail.
+	InitialElectionTickAdvance bool `json:"initial-election-tick-advance"`
+
 	QuotaBackendBytes int64 `json:"quota-backend-bytes"`
 	MaxRequestBytes   uint  `json:"max-request-bytes"`
 
@@ -128,6 +160,11 @@ type Config struct {
 	PeerTLSInfo   transport.TLSInfo
 	PeerAutoTLS   bool
 
+	// CipherSuites is a list of supported TLS cipher suites between
+	// client/server and peers. If empty, Go auto-populates the list.
+	// Note that cipher suites are prioritized in the given order.
+	CipherSuites []string `json:"cipher-suites"`
+
 	// debug
 
 	Debug        bool   `json:"debug"`
@@ -190,27 +227,28 @@ func NewConfig() *Config {
 	lcurl, _ := url.Parse(DefaultListenClientURLs)
 	acurl, _ := url.Parse(DefaultAdvertiseClientURLs)
 	cfg := &Config{
-		CorsInfo:              &cors.CORSInfo{},
-		MaxSnapFiles:          DefaultMaxSnapshots,
-		MaxWalFiles:           DefaultMaxWALs,
-		Name:                  DefaultName,
-		SnapCount:             etcdserver.DefaultSnapCount,
-		MaxRequestBytes:       DefaultMaxRequestBytes,
-		GRPCKeepAliveMinTime:  DefaultGRPCKeepAliveMinTime,
-		GRPCKeepAliveInterval: DefaultGRPCKeepAliveInterval,
-		GRPCKeepAliveTimeout:  DefaultGRPCKeepAliveTimeout,
-		TickMs:                100,
-		ElectionMs:            1000,
-		LPUrls:                []url.URL{*lpurl},
-		LCUrls:                []url.URL{*lcurl},
-		APUrls:                []url.URL{*apurl},
-		ACUrls:                []url.URL{*acurl},
-		ClusterState:          ClusterStateFlagNew,
-		InitialClusterToken:   "etcd-cluster",
-		StrictReconfigCheck:   true,
-		Metrics:               "basic",
-		EnableV2:              true,
-		AuthToken:             "simple",
+		CorsInfo:                   &cors.CORSInfo{},
+		MaxSnapFiles:               DefaultMaxSnapshots,
+		MaxWalFiles:                DefaultMaxWALs,
+		Name:                       DefaultName,
+		SnapCount:                  etcdserver.DefaultSnapCount,
+		MaxRequestBytes:            DefaultMaxRequestBytes,
+		GRPCKeepAliveMinTime:       DefaultGRPCKeepAliveMinTime,
+		GRPCKeepAliveInterval:      DefaultGRPCKeepAliveInterval,
+		GRPCKeepAliveTimeout:       DefaultGRPCKeepAliveTimeout,
+		TickMs:                     100,
+		ElectionMs:                 1000,
+		InitialElectionTickAdvance: true,
+		LPUrls:              []url.URL{*lpurl},
+		LCUrls:              []url.URL{*lcurl},
+		APUrls:              []url.URL{*apurl},
+		ACUrls:              []url.URL{*acurl},
+		ClusterState:        ClusterStateFlagNew,
+		InitialClusterToken: "etcd-cluster",
+		StrictReconfigCheck: true,
+		Metrics:             "basic",
+		EnableV2:            true,
+		AuthToken:           "simple",
 	}
 	cfg.InitialCluster = cfg.InitialClusterFromName(cfg.Name)
 	return cfg
@@ -298,6 +336,25 @@ func (cfg *configYAML) configFromFile(path string) error {
 	return cfg.Validate()
 }
 
+func updateCipherSuites(tls *transport.TLSInfo, ss []string) error {
+	if len(tls.CipherSuites) > 0 && len(ss) > 0 {
+		return fmt.Errorf("TLSInfo.CipherSuites is already specified (given %v)", ss)
+	}
+	if len(ss) > 0 {
+		cs := make([]uint16, len(ss))
+		for i, s := range ss {
+			var ok bool
+			cs[i], ok = tlsutil.GetCipherSuite(s)
+			if !ok {
+				return fmt.Errorf("unexpected TLS cipher suite %q", s)
+			}
+		}
+		tls.CipherSuites = cs
+	}
+	return nil
+}
+
+// Validate ensures that '*embed.Config' fields are properly configured.
 func (cfg *Config) Validate() error {
 	if err := checkBindURLs(cfg.LPUrls); err != nil {
 		return err
@@ -400,6 +457,44 @@ func (cfg Config) defaultClientHost() bool {
 	return len(cfg.ACUrls) == 1 && cfg.ACUrls[0].String() == DefaultAdvertiseClientURLs
 }
 
+func (cfg *Config) ClientSelfCert() (err error) {
+	if !cfg.ClientAutoTLS {
+		return nil
+	}
+	if !cfg.ClientTLSInfo.Empty() {
+		plog.Warningf("ignoring client auto TLS since certs given")
+		return nil
+	}
+	chosts := make([]string, len(cfg.LCUrls))
+	for i, u := range cfg.LCUrls {
+		chosts[i] = u.Host
+	}
+	cfg.ClientTLSInfo, err = transport.SelfCert(filepath.Join(cfg.Dir, "fixtures", "client"), chosts)
+	if err != nil {
+		return err
+	}
+	return updateCipherSuites(&cfg.ClientTLSInfo, cfg.CipherSuites)
+}
+
+func (cfg *Config) PeerSelfCert() (err error) {
+	if !cfg.PeerAutoTLS {
+		return nil
+	}
+	if !cfg.PeerTLSInfo.Empty() {
+		plog.Warningf("ignoring peer auto TLS since certs given")
+		return nil
+	}
+	phosts := make([]string, len(cfg.LPUrls))
+	for i, u := range cfg.LPUrls {
+		phosts[i] = u.Host
+	}
+	cfg.PeerTLSInfo, err = transport.SelfCert(filepath.Join(cfg.Dir, "fixtures", "peer"), phosts)
+	if err != nil {
+		return err
+	}
+	return updateCipherSuites(&cfg.PeerTLSInfo, cfg.CipherSuites)
+}
+
 // UpdateDefaultClusterFromName updates cluster advertise URLs with, if available, default host,
 // if advertise URLs are default values(localhost:2379,2380) AND if listen URL is 0.0.0.0.
 // e.g. advertise peer URL localhost:2380 or listen peer URL 0.0.0.0:2380

vendor/github.com/coreos/etcd/embed/etcd.go

@@ -22,7 +22,6 @@ import (
 	defaultLog "log"
 	"net"
 	"net/http"
-	"path/filepath"
 	"sync"
 	"time"
 
@@ -128,30 +127,31 @@ func StartEtcd(inCfg *Config) (e *Etcd, err error) {
 	}
 
 	srvcfg := &etcdserver.ServerConfig{
-		Name:                    cfg.Name,
-		ClientURLs:              cfg.ACUrls,
-		PeerURLs:                cfg.APUrls,
-		DataDir:                 cfg.Dir,
-		DedicatedWALDir:         cfg.WalDir,
-		SnapCount:               cfg.SnapCount,
-		MaxSnapFiles:            cfg.MaxSnapFiles,
-		MaxWALFiles:             cfg.MaxWalFiles,
-		InitialPeerURLsMap:      urlsmap,
-		InitialClusterToken:     token,
-		DiscoveryURL:            cfg.Durl,
-		DiscoveryProxy:          cfg.Dproxy,
-		NewCluster:              cfg.IsNewCluster(),
-		ForceNewCluster:         cfg.ForceNewCluster,
-		PeerTLSInfo:             cfg.PeerTLSInfo,
-		TickMs:                  cfg.TickMs,
-		ElectionTicks:           cfg.ElectionTicks(),
-		AutoCompactionRetention: cfg.AutoCompactionRetention,
-		QuotaBackendBytes:       cfg.QuotaBackendBytes,
-		MaxRequestBytes:         cfg.MaxRequestBytes,
-		StrictReconfigCheck:     cfg.StrictReconfigCheck,
-		ClientCertAuthEnabled:   cfg.ClientTLSInfo.ClientCertAuth,
-		AuthToken:               cfg.AuthToken,
-		Debug:                   cfg.Debug,
+		Name:                       cfg.Name,
+		ClientURLs:                 cfg.ACUrls,
+		PeerURLs:                   cfg.APUrls,
+		DataDir:                    cfg.Dir,
+		DedicatedWALDir:            cfg.WalDir,
+		SnapCount:                  cfg.SnapCount,
+		MaxSnapFiles:               cfg.MaxSnapFiles,
+		MaxWALFiles:                cfg.MaxWalFiles,
+		InitialPeerURLsMap:         urlsmap,
+		InitialClusterToken:        token,
+		DiscoveryURL:               cfg.Durl,
+		DiscoveryProxy:             cfg.Dproxy,
+		NewCluster:                 cfg.IsNewCluster(),
+		ForceNewCluster:            cfg.ForceNewCluster,
+		PeerTLSInfo:                cfg.PeerTLSInfo,
+		TickMs:                     cfg.TickMs,
+		ElectionTicks:              cfg.ElectionTicks(),
+		InitialElectionTickAdvance: cfg.InitialElectionTickAdvance,
+		AutoCompactionRetention:    cfg.AutoCompactionRetention,
+		QuotaBackendBytes:          cfg.QuotaBackendBytes,
+		MaxRequestBytes:            cfg.MaxRequestBytes,
+		StrictReconfigCheck:        cfg.StrictReconfigCheck,
+		ClientCertAuthEnabled:      cfg.ClientTLSInfo.ClientCertAuth,
+		AuthToken:                  cfg.AuthToken,
+		Debug:                      cfg.Debug,
 	}
 
 	if e.Server, err = etcdserver.NewServer(srvcfg); err != nil {
@@ -263,17 +263,11 @@ func stopServers(ctx context.Context, ss *servers) {
 func (e *Etcd) Err() <-chan error { return e.errc }
 
 func startPeerListeners(cfg *Config) (peers []*peerListener, err error) {
-	if cfg.PeerAutoTLS && cfg.PeerTLSInfo.Empty() {
-		phosts := make([]string, len(cfg.LPUrls))
-		for i, u := range cfg.LPUrls {
-			phosts[i] = u.Host
-		}
-		cfg.PeerTLSInfo, err = transport.SelfCert(filepath.Join(cfg.Dir, "fixtures", "peer"), phosts)
-		if err != nil {
-			plog.Fatalf("could not get certs (%v)", err)
-		}
-	} else if cfg.PeerAutoTLS {
-		plog.Warningf("ignoring peer auto TLS since certs given")
+	if err = updateCipherSuites(&cfg.PeerTLSInfo, cfg.CipherSuites); err != nil {
+		return nil, err
+	}
+	if err = cfg.PeerSelfCert(); err != nil {
+		plog.Fatalf("could not get certs (%v)", err)
 	}
 
 	if !cfg.PeerTLSInfo.Empty() {
@@ -358,17 +352,11 @@ func (e *Etcd) servePeers() (err error) {
 }
 
 func startClientListeners(cfg *Config) (sctxs map[string]*serveCtx, err error) {
-	if cfg.ClientAutoTLS && cfg.ClientTLSInfo.Empty() {
-		chosts := make([]string, len(cfg.LCUrls))
-		for i, u := range cfg.LCUrls {
-			chosts[i] = u.Host
-		}
-		cfg.ClientTLSInfo, err = transport.SelfCert(filepath.Join(cfg.Dir, "fixtures", "client"), chosts)
-		if err != nil {
-			plog.Fatalf("could not get certs (%v)", err)
-		}
-	} else if cfg.ClientAutoTLS {
-		plog.Warningf("ignoring client auto TLS since certs given")
+	if err = updateCipherSuites(&cfg.ClientTLSInfo, cfg.CipherSuites); err != nil {
+		return nil, err
+	}
+	if err = cfg.ClientSelfCert(); err != nil {
+		plog.Fatalf("could not get certs (%v)", err)
 	}
 
 	if cfg.EnablePprof {

vendor/github.com/coreos/etcd/etcdserver/BUILD

@@ -61,6 +61,7 @@ go_library(
         "//vendor/github.com/coreos/go-semver/semver:go_default_library",
         "//vendor/github.com/coreos/pkg/capnslog:go_default_library",
         "//vendor/github.com/gogo/protobuf/proto:go_default_library",
+        "//vendor/github.com/golang/protobuf/proto:go_default_library",
         "//vendor/github.com/prometheus/client_golang/prometheus:go_default_library",
         "//vendor/golang.org/x/net/context:go_default_library",
     ],

vendor/github.com/coreos/etcd/etcdserver/api/v3election/election.go

@@ -15,6 +15,8 @@
 package v3election
 
 import (
+	"errors"
+
 	"golang.org/x/net/context"
 
 	"github.com/coreos/etcd/clientv3"
@@ -22,6 +24,10 @@ import (
 	epb "github.com/coreos/etcd/etcdserver/api/v3election/v3electionpb"
 )
 
+// ErrMissingLeaderKey is returned when election API request
+// is missing the "leader" field.
+var ErrMissingLeaderKey = errors.New(`"leader" field must be provided`)
+
 type electionServer struct {
 	c *clientv3.Client
 }
@@ -51,6 +57,9 @@ func (es *electionServer) Campaign(ctx context.Context, req *epb.CampaignRequest
 }
 
 func (es *electionServer) Proclaim(ctx context.Context, req *epb.ProclaimRequest) (*epb.ProclaimResponse, error) {
+	if req.Leader == nil {
+		return nil, ErrMissingLeaderKey
+	}
 	s, err := es.session(ctx, req.Leader.Lease)
 	if err != nil {
 		return nil, err
@@ -98,6 +107,9 @@ func (es *electionServer) Leader(ctx context.Context, req *epb.LeaderRequest) (*
 }
 
 func (es *electionServer) Resign(ctx context.Context, req *epb.ResignRequest) (*epb.ResignResponse, error) {
+	if req.Leader == nil {
+		return nil, ErrMissingLeaderKey
+	}
 	s, err := es.session(ctx, req.Leader.Lease)
 	if err != nil {
 		return nil, err

vendor/github.com/coreos/etcd/etcdserver/api/v3rpc/BUILD

@@ -44,6 +44,7 @@ go_library(
         "//vendor/google.golang.org/grpc/credentials:go_default_library",
         "//vendor/google.golang.org/grpc/grpclog:go_default_library",
         "//vendor/google.golang.org/grpc/metadata:go_default_library",
+        "//vendor/google.golang.org/grpc/status:go_default_library",
     ],
 )
 

vendor/github.com/coreos/etcd/etcdserver/api/v3rpc/lease.go

@@ -92,7 +92,11 @@ func (ls *LeaseServer) leaseKeepAlive(stream pb.Lease_LeaseKeepAliveServer) erro
 			return nil
 		}
 		if err != nil {
-			plog.Debugf("failed to receive lease keepalive request from gRPC stream (%q)", err.Error())
+			if isClientCtxErr(stream.Context().Err(), err) {
+				plog.Debugf("failed to receive lease keepalive request from gRPC stream (%q)", err.Error())
+			} else {
+				plog.Warningf("failed to receive lease keepalive request from gRPC stream (%q)", err.Error())
+			}
 			return err
 		}
 
@@ -118,7 +122,11 @@ func (ls *LeaseServer) leaseKeepAlive(stream pb.Lease_LeaseKeepAliveServer) erro
 		resp.TTL = ttl
 		err = stream.Send(resp)
 		if err != nil {
-			plog.Debugf("failed to send lease keepalive response to gRPC stream (%q)", err.Error())
+			if isClientCtxErr(stream.Context().Err(), err) {
+				plog.Debugf("failed to send lease keepalive response to gRPC stream (%q)", err.Error())
+			} else {
+				plog.Warningf("failed to send lease keepalive response to gRPC stream (%q)", err.Error())
+			}
 			return err
 		}
 	}

vendor/github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes/error.go

@@ -32,8 +32,9 @@ var (
 	ErrGRPCFutureRev     = grpc.Errorf(codes.OutOfRange, "etcdserver: mvcc: required revision is a future revision")
 	ErrGRPCNoSpace       = grpc.Errorf(codes.ResourceExhausted, "etcdserver: mvcc: database space exceeded")
 
-	ErrGRPCLeaseNotFound = grpc.Errorf(codes.NotFound, "etcdserver: requested lease not found")
-	ErrGRPCLeaseExist    = grpc.Errorf(codes.FailedPrecondition, "etcdserver: lease already exists")
+	ErrGRPCLeaseNotFound    = grpc.Errorf(codes.NotFound, "etcdserver: requested lease not found")
+	ErrGRPCLeaseExist       = grpc.Errorf(codes.FailedPrecondition, "etcdserver: lease already exists")
+	ErrGRPCLeaseTTLTooLarge = grpc.Errorf(codes.OutOfRange, "etcdserver: too large lease TTL")
 
 	ErrGRPCMemberExist            = grpc.Errorf(codes.FailedPrecondition, "etcdserver: member ID already exist")
 	ErrGRPCPeerURLExist           = grpc.Errorf(codes.FailedPrecondition, "etcdserver: Peer URLs already exists")
@@ -79,8 +80,9 @@ var (
 		grpc.ErrorDesc(ErrGRPCFutureRev):    ErrGRPCFutureRev,
 		grpc.ErrorDesc(ErrGRPCNoSpace):      ErrGRPCNoSpace,
 
-		grpc.ErrorDesc(ErrGRPCLeaseNotFound): ErrGRPCLeaseNotFound,
-		grpc.ErrorDesc(ErrGRPCLeaseExist):    ErrGRPCLeaseExist,
+		grpc.ErrorDesc(ErrGRPCLeaseNotFound):    ErrGRPCLeaseNotFound,
+		grpc.ErrorDesc(ErrGRPCLeaseExist):       ErrGRPCLeaseExist,
+		grpc.ErrorDesc(ErrGRPCLeaseTTLTooLarge): ErrGRPCLeaseTTLTooLarge,
 
 		grpc.ErrorDesc(ErrGRPCMemberExist):            ErrGRPCMemberExist,
 		grpc.ErrorDesc(ErrGRPCPeerURLExist):           ErrGRPCPeerURLExist,
@@ -126,8 +128,9 @@ var (
 	ErrFutureRev     = Error(ErrGRPCFutureRev)
 	ErrNoSpace       = Error(ErrGRPCNoSpace)
 
-	ErrLeaseNotFound = Error(ErrGRPCLeaseNotFound)
-	ErrLeaseExist    = Error(ErrGRPCLeaseExist)
+	ErrLeaseNotFound    = Error(ErrGRPCLeaseNotFound)
+	ErrLeaseExist       = Error(ErrGRPCLeaseExist)
+	ErrLeaseTTLTooLarge = Error(ErrGRPCLeaseTTLTooLarge)
 
 	ErrMemberExist            = Error(ErrGRPCMemberExist)
 	ErrPeerURLExist           = Error(ErrGRPCPeerURLExist)

vendor/github.com/coreos/etcd/etcdserver/api/v3rpc/util.go

@@ -15,14 +15,18 @@
 package v3rpc
 
 import (
+	"strings"
+
 	"github.com/coreos/etcd/auth"
 	"github.com/coreos/etcd/etcdserver"
 	"github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes"
 	"github.com/coreos/etcd/etcdserver/membership"
 	"github.com/coreos/etcd/lease"
 	"github.com/coreos/etcd/mvcc"
+
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/codes"
+	"google.golang.org/grpc/status"
 )
 
 func togRPCError(err error) error {
@@ -68,6 +72,8 @@ func togRPCError(err error) error {
 		return rpctypes.ErrGRPCLeaseNotFound
 	case lease.ErrLeaseExists:
 		return rpctypes.ErrGRPCLeaseExist
+	case lease.ErrLeaseTTLTooLarge:
+		return rpctypes.ErrGRPCLeaseTTLTooLarge
 
 	case auth.ErrRootUserNotExist:
 		return rpctypes.ErrGRPCRootUserNotExist
@@ -101,3 +107,35 @@ func togRPCError(err error) error {
 		return grpc.Errorf(codes.Unknown, err.Error())
 	}
 }
+
+func isClientCtxErr(ctxErr error, err error) bool {
+	if ctxErr != nil {
+		return true
+	}
+
+	ev, ok := status.FromError(err)
+	if !ok {
+		return false
+	}
+
+	switch ev.Code() {
+	case codes.Canceled, codes.DeadlineExceeded:
+		// client-side context cancel or deadline exceeded
+		// "rpc error: code = Canceled desc = context canceled"
+		// "rpc error: code = DeadlineExceeded desc = context deadline exceeded"
+		return true
+	case codes.Unavailable:
+		msg := ev.Message()
+		// client-side context cancel or deadline exceeded with TLS ("http2.errClientDisconnected")
+		// "rpc error: code = Unavailable desc = client disconnected"
+		if msg == "client disconnected" {
+			return true
+		}
+		// "grpc/transport.ClientTransport.CloseStream" on canceled streams
+		// "rpc error: code = Unavailable desc = stream error: stream ID 21; CANCEL")
+		if strings.HasPrefix(msg, "stream error: ") && strings.HasSuffix(msg, "; CANCEL") {
+			return true
+		}
+	}
+	return false
+}

vendor/github.com/coreos/etcd/etcdserver/api/v3rpc/watch.go

@@ -141,7 +141,11 @@ func (ws *watchServer) Watch(stream pb.Watch_WatchServer) (err error) {
 	// deadlock when calling sws.close().
 	go func() {
 		if rerr := sws.recvLoop(); rerr != nil {
-			plog.Debugf("failed to receive watch request from gRPC stream (%q)", rerr.Error())
+			if isClientCtxErr(stream.Context().Err(), rerr) {
+				plog.Debugf("failed to receive watch request from gRPC stream (%q)", rerr.Error())
+			} else {
+				plog.Warningf("failed to receive watch request from gRPC stream (%q)", rerr.Error())
+			}
 			errc <- rerr
 		}
 	}()
@@ -338,7 +342,11 @@ func (sws *serverWatchStream) sendLoop() {
 
 			mvcc.ReportEventReceived(len(evs))
 			if err := sws.gRPCStream.Send(wr); err != nil {
-				plog.Debugf("failed to send watch response to gRPC stream (%q)", err.Error())
+				if isClientCtxErr(sws.gRPCStream.Context().Err(), err) {
+					plog.Debugf("failed to send watch response to gRPC stream (%q)", err.Error())
+				} else {
+					plog.Warningf("failed to send watch response to gRPC stream (%q)", err.Error())
+				}
 				return
 			}
 
@@ -355,7 +363,11 @@ func (sws *serverWatchStream) sendLoop() {
 			}
 
 			if err := sws.gRPCStream.Send(c); err != nil {
-				plog.Debugf("failed to send watch control response to gRPC stream (%q)", err.Error())
+				if isClientCtxErr(sws.gRPCStream.Context().Err(), err) {
+					plog.Debugf("failed to send watch control response to gRPC stream (%q)", err.Error())
+				} else {
+					plog.Warningf("failed to send watch control response to gRPC stream (%q)", err.Error())
+				}
 				return
 			}
 
@@ -371,7 +383,11 @@ func (sws *serverWatchStream) sendLoop() {
 				for _, v := range pending[wid] {
 					mvcc.ReportEventReceived(len(v.Events))
 					if err := sws.gRPCStream.Send(v); err != nil {
-						plog.Debugf("failed to send pending watch response to gRPC stream (%q)", err.Error())
+						if isClientCtxErr(sws.gRPCStream.Context().Err(), err) {
+							plog.Debugf("failed to send pending watch response to gRPC stream (%q)", err.Error())
+						} else {
+							plog.Warningf("failed to send pending watch response to gRPC stream (%q)", err.Error())
+						}
 						return
 					}
 				}

vendor/github.com/coreos/etcd/etcdserver/apply.go

@@ -89,6 +89,9 @@ func (s *EtcdServer) newApplierV3() applierV3 {
 
 func (a *applierV3backend) Apply(r *pb.InternalRaftRequest) *applyResult {
 	ar := &applyResult{}
+	defer func(start time.Time) {
+		warnOfExpensiveRequest(start, &pb.InternalRaftStringer{Request: r}, ar.resp, ar.err)
+	}(time.Now())
 
 	// call into a.s.applyV3.F instead of a.F so upper appliers can check individual calls
 	switch {

vendor/github.com/coreos/etcd/etcdserver/apply_v2.go

@@ -105,10 +105,12 @@ func (a *applierV2store) Sync(r *pb.Request) Response {
 	return Response{}
 }
 
-// applyV2Request interprets r as a call to store.X and returns a Response interpreted
-// from store.Event
+// applyV2Request interprets r as a call to v2store.X
+// and returns a Response interpreted from v2store.Event
 func (s *EtcdServer) applyV2Request(r *pb.Request) Response {
+	defer warnOfExpensiveRequest(time.Now(), r, nil, nil)
 	toTTLOptions(r)
+
 	switch r.Method {
 	case "POST":
 		return s.applyV2.Post(r)

vendor/github.com/coreos/etcd/etcdserver/config.go

@@ -48,8 +48,38 @@ type ServerConfig struct {
 	ForceNewCluster     bool
 	PeerTLSInfo         transport.TLSInfo
 
-	TickMs           uint
-	ElectionTicks    int
+	TickMs        uint
+	ElectionTicks int
+
+	// InitialElectionTickAdvance is true, then local member fast-forwards
+	// election ticks to speed up "initial" leader election trigger. This
+	// benefits the case of larger election ticks. For instance, cross
+	// datacenter deployment may require longer election timeout of 10-second.
+	// If true, local node does not need wait up to 10-second. Instead,
+	// forwards its election ticks to 8-second, and have only 2-second left
+	// before leader election.
+	//
+	// Major assumptions are that:
+	//  - cluster has no active leader thus advancing ticks enables faster
+	//    leader election, or
+	//  - cluster already has an established leader, and rejoining follower
+	//    is likely to receive heartbeats from the leader after tick advance
+	//    and before election timeout.
+	//
+	// However, when network from leader to rejoining follower is congested,
+	// and the follower does not receive leader heartbeat within left election
+	// ticks, disruptive election has to happen thus affecting cluster
+	// availabilities.
+	//
+	// Disabling this would slow down initial bootstrap process for cross
+	// datacenter deployments. Make your own tradeoffs by configuring
+	// --initial-election-tick-advance at the cost of slow initial bootstrap.
+	//
+	// If single-node, it advances ticks regardless.
+	//
+	// See https://github.com/coreos/etcd/issues/9333 for more detail.
+	InitialElectionTickAdvance bool
+
 	BootstrapTimeout time.Duration
 
 	AutoCompactionRetention int

vendor/github.com/coreos/etcd/etcdserver/etcdserverpb/BUILD

@@ -5,6 +5,7 @@ go_library(
     srcs = [
         "etcdserver.pb.go",
         "raft_internal.pb.go",
+        "raft_internal_stringer.go",
         "rpc.pb.go",
     ],
     importmap = "k8s.io/kubernetes/vendor/github.com/coreos/etcd/etcdserver/etcdserverpb",

vendor/github.com/coreos/etcd/etcdserver/etcdserverpb/raft_internal_stringer.go

+// Copyright 2018 The etcd 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 etcdserverpb
+
+import (
+	"fmt"
+	"strings"
+
+	proto "github.com/golang/protobuf/proto"
+)
+
+// InternalRaftStringer implements custom proto Stringer:
+// redact password, replace value fields with value_size fields.
+type InternalRaftStringer struct {
+	Request *InternalRaftRequest
+}
+
+func (as *InternalRaftStringer) String() string {
+	switch {
+	case as.Request.LeaseGrant != nil:
+		return fmt.Sprintf("header:<%s> lease_grant:<ttl:%d-second id:%016x>",
+			as.Request.Header.String(),
+			as.Request.LeaseGrant.TTL,
+			as.Request.LeaseGrant.ID,
+		)
+	case as.Request.LeaseRevoke != nil:
+		return fmt.Sprintf("header:<%s> lease_revoke:<id:%016x>",
+			as.Request.Header.String(),
+			as.Request.LeaseRevoke.ID,
+		)
+	case as.Request.Authenticate != nil:
+		return fmt.Sprintf("header:<%s> authenticate:<name:%s simple_token:%s>",
+			as.Request.Header.String(),
+			as.Request.Authenticate.Name,
+			as.Request.Authenticate.SimpleToken,
+		)
+	case as.Request.AuthUserAdd != nil:
+		return fmt.Sprintf("header:<%s> auth_user_add:<name:%s>",
+			as.Request.Header.String(),
+			as.Request.AuthUserAdd.Name,
+		)
+	case as.Request.AuthUserChangePassword != nil:
+		return fmt.Sprintf("header:<%s> auth_user_change_password:<name:%s>",
+			as.Request.Header.String(),
+			as.Request.AuthUserChangePassword.Name,
+		)
+	case as.Request.Put != nil:
+		return fmt.Sprintf("header:<%s> put:<%s>",
+			as.Request.Header.String(),
+			newLoggablePutRequest(as.Request.Put).String(),
+		)
+	case as.Request.Txn != nil:
+		return fmt.Sprintf("header:<%s> txn:<%s>",
+			as.Request.Header.String(),
+			NewLoggableTxnRequest(as.Request.Txn).String(),
+		)
+	default:
+		// nothing to redact
+	}
+	return as.Request.String()
+}
+
+// txnRequestStringer implements a custom proto String to replace value bytes fields with value size
+// fields in any nested txn and put operations.
+type txnRequestStringer struct {
+	Request *TxnRequest
+}
+
+func NewLoggableTxnRequest(request *TxnRequest) *txnRequestStringer {
+	return &txnRequestStringer{request}
+}
+
+func (as *txnRequestStringer) String() string {
+	var compare []string
+	for _, c := range as.Request.Compare {
+		switch cv := c.TargetUnion.(type) {
+		case *Compare_Value:
+			compare = append(compare, newLoggableValueCompare(c, cv).String())
+		default:
+			// nothing to redact
+			compare = append(compare, c.String())
+		}
+	}
+	var success []string
+	for _, s := range as.Request.Success {
+		success = append(success, newLoggableRequestOp(s).String())
+	}
+	var failure []string
+	for _, f := range as.Request.Failure {
+		failure = append(failure, newLoggableRequestOp(f).String())
+	}
+	return fmt.Sprintf("compare:<%s> success:<%s> failure:<%s>",
+		strings.Join(compare, " "),
+		strings.Join(success, " "),
+		strings.Join(failure, " "),
+	)
+}
+
+// requestOpStringer implements a custom proto String to replace value bytes fields with value
+// size fields in any nested txn and put operations.
+type requestOpStringer struct {
+	Op *RequestOp
+}
+
+func newLoggableRequestOp(op *RequestOp) *requestOpStringer {
+	return &requestOpStringer{op}
+}
+
+func (as *requestOpStringer) String() string {
+	switch op := as.Op.Request.(type) {
+	case *RequestOp_RequestPut:
+		return fmt.Sprintf("request_put:<%s>", newLoggablePutRequest(op.RequestPut).String())
+	default:
+		// nothing to redact
+	}
+	return as.Op.String()
+}
+
+// loggableValueCompare implements a custom proto String for Compare.Value union member types to
+// replace the value bytes field with a value size field.
+// To preserve proto encoding of the key and range_end bytes, a faked out proto type is used here.
+type loggableValueCompare struct {
+	Result    Compare_CompareResult `protobuf:"varint,1,opt,name=result,proto3,enum=etcdserverpb.Compare_CompareResult"`
+	Target    Compare_CompareTarget `protobuf:"varint,2,opt,name=target,proto3,enum=etcdserverpb.Compare_CompareTarget"`
+	Key       []byte                `protobuf:"bytes,3,opt,name=key,proto3"`
+	ValueSize int                   `protobuf:"bytes,7,opt,name=value_size,proto3"`
+}
+
+func newLoggableValueCompare(c *Compare, cv *Compare_Value) *loggableValueCompare {
+	return &loggableValueCompare{
+		c.Result,
+		c.Target,
+		c.Key,
+		len(cv.Value),
+	}
+}
+
+func (m *loggableValueCompare) Reset()         { *m = loggableValueCompare{} }
+func (m *loggableValueCompare) String() string { return proto.CompactTextString(m) }
+func (*loggableValueCompare) ProtoMessage()    {}
+
+// loggablePutRequest implements a custom proto String to replace value bytes field with a value
+// size field.
+// To preserve proto encoding of the key bytes, a faked out proto type is used here.
+type loggablePutRequest struct {
+	Key         []byte `protobuf:"bytes,1,opt,name=key,proto3"`
+	ValueSize   int    `protobuf:"varint,2,opt,name=value_size,proto3"`
+	Lease       int64  `protobuf:"varint,3,opt,name=lease,proto3"`
+	PrevKv      bool   `protobuf:"varint,4,opt,name=prev_kv,proto3"`
+	IgnoreValue bool   `protobuf:"varint,5,opt,name=ignore_value,proto3"`
+	IgnoreLease bool   `protobuf:"varint,6,opt,name=ignore_lease,proto3"`
+}
+
+func newLoggablePutRequest(request *PutRequest) *loggablePutRequest {
+	return &loggablePutRequest{
+		request.Key,
+		len(request.Value),
+		request.Lease,
+		request.PrevKv,
+		request.IgnoreValue,
+		request.IgnoreLease,
+	}
+}
+
+func (m *loggablePutRequest) Reset()         { *m = loggablePutRequest{} }
+func (m *loggablePutRequest) String() string { return proto.CompactTextString(m) }
+func (*loggablePutRequest) ProtoMessage()    {}

vendor/github.com/coreos/etcd/etcdserver/metrics.go

@@ -15,9 +15,11 @@
 package etcdserver
 
 import (
+	goruntime "runtime"
 	"time"
 
 	"github.com/coreos/etcd/pkg/runtime"
+	"github.com/coreos/etcd/version"
 	"github.com/prometheus/client_golang/prometheus"
 )
 
@@ -28,12 +30,30 @@ var (
 		Name:      "has_leader",
 		Help:      "Whether or not a leader exists. 1 is existence, 0 is not.",
 	})
+	isLeader = prometheus.NewGauge(prometheus.GaugeOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "is_leader",
+		Help:      "Whether or not this member is a leader. 1 if is, 0 otherwise.",
+	})
 	leaderChanges = prometheus.NewCounter(prometheus.CounterOpts{
 		Namespace: "etcd",
 		Subsystem: "server",
 		Name:      "leader_changes_seen_total",
 		Help:      "The number of leader changes seen.",
 	})
+	heartbeatSendFailures = prometheus.NewCounter(prometheus.CounterOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "heartbeat_send_failures_total",
+		Help:      "The total number of leader heartbeat send failures (likely overloaded from slow disk).",
+	})
+	slowApplies = prometheus.NewCounter(prometheus.CounterOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "slow_apply_total",
+		Help:      "The total number of slow apply requests (likely overloaded from slow disk).",
+	})
 	proposalsCommitted = prometheus.NewGauge(prometheus.GaugeOpts{
 		Namespace: "etcd",
 		Subsystem: "server",
@@ -64,16 +84,56 @@ var (
 		Name:      "lease_expired_total",
 		Help:      "The total number of expired leases.",
 	})
+	slowReadIndex = prometheus.NewCounter(prometheus.CounterOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "slow_read_indexes_total",
+		Help:      "The total number of pending read indexes not in sync with leader's or timed out read index requests.",
+	})
+	quotaBackendBytes = prometheus.NewGauge(prometheus.GaugeOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "quota_backend_bytes",
+		Help:      "Current backend storage quota size in bytes.",
+	})
+	currentVersion = prometheus.NewGaugeVec(prometheus.GaugeOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "version",
+		Help:      "Which version is running. 1 for 'server_version' label with current version.",
+	},
+		[]string{"server_version"})
+	currentGoVersion = prometheus.NewGaugeVec(prometheus.GaugeOpts{
+		Namespace: "etcd",
+		Subsystem: "server",
+		Name:      "go_version",
+		Help:      "Which Go version server is running with. 1 for 'server_go_version' label with current version.",
+	},
+		[]string{"server_go_version"})
 )
 
 func init() {
 	prometheus.MustRegister(hasLeader)
+	prometheus.MustRegister(isLeader)
 	prometheus.MustRegister(leaderChanges)
+	prometheus.MustRegister(heartbeatSendFailures)
+	prometheus.MustRegister(slowApplies)
 	prometheus.MustRegister(proposalsCommitted)
 	prometheus.MustRegister(proposalsApplied)
 	prometheus.MustRegister(proposalsPending)
 	prometheus.MustRegister(proposalsFailed)
 	prometheus.MustRegister(leaseExpired)
+	prometheus.MustRegister(slowReadIndex)
+	prometheus.MustRegister(quotaBackendBytes)
+	prometheus.MustRegister(currentVersion)
+	prometheus.MustRegister(currentGoVersion)
+
+	currentVersion.With(prometheus.Labels{
+		"server_version": version.Version,
+	}).Set(1)
+	currentGoVersion.With(prometheus.Labels{
+		"server_go_version": goruntime.Version(),
+	}).Set(1)
 }
 
 func monitorFileDescriptor(done <-chan struct{}) {

vendor/github.com/coreos/etcd/etcdserver/quota.go

@@ -14,9 +14,7 @@
 
 package etcdserver
 
-import (
-	pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
-)
+import pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
 
 const (
 	// DefaultQuotaBytes is the number of bytes the backend Size may
@@ -58,15 +56,20 @@ const (
 )
 
 func NewBackendQuota(s *EtcdServer) Quota {
+	quotaBackendBytes.Set(float64(s.Cfg.QuotaBackendBytes))
+
 	if s.Cfg.QuotaBackendBytes < 0 {
 		// disable quotas if negative
 		plog.Warningf("disabling backend quota")
 		return &passthroughQuota{}
 	}
+
 	if s.Cfg.QuotaBackendBytes == 0 {
 		// use default size if no quota size given
+		quotaBackendBytes.Set(float64(DefaultQuotaBytes))
 		return &backendQuota{s, DefaultQuotaBytes}
 	}
+
 	if s.Cfg.QuotaBackendBytes > MaxQuotaBytes {
 		plog.Warningf("backend quota %v exceeds maximum recommended quota %v", s.Cfg.QuotaBackendBytes, MaxQuotaBytes)
 	}

vendor/github.com/coreos/etcd/etcdserver/raft.go

@@ -95,6 +95,7 @@ type raftNode struct {
 	term  uint64
 	lead  uint64
 
+	tickMu *sync.Mutex
 	raftNodeConfig
 
 	// a chan to send/receive snapshot
@@ -131,6 +132,7 @@ type raftNodeConfig struct {
 
 func newRaftNode(cfg raftNodeConfig) *raftNode {
 	r := &raftNode{
+		tickMu:         new(sync.Mutex),
 		raftNodeConfig: cfg,
 		// set up contention detectors for raft heartbeat message.
 		// expect to send a heartbeat within 2 heartbeat intervals.
@@ -149,6 +151,13 @@ func newRaftNode(cfg raftNodeConfig) *raftNode {
 	return r
 }
 
+// raft.Node does not have locks in Raft package
+func (r *raftNode) tick() {
+	r.tickMu.Lock()
+	r.Tick()
+	r.tickMu.Unlock()
+}
+
 // start prepares and starts raftNode in a new goroutine. It is no longer safe
 // to modify the fields after it has been started.
 func (r *raftNode) start(rh *raftReadyHandler) {
@@ -161,7 +170,7 @@ func (r *raftNode) start(rh *raftReadyHandler) {
 		for {
 			select {
 			case <-r.ticker.C:
-				r.Tick()
+				r.tick()
 			case rd := <-r.Ready():
 				if rd.SoftState != nil {
 					newLeader := rd.SoftState.Lead != raft.None && atomic.LoadUint64(&r.lead) != rd.SoftState.Lead
@@ -177,6 +186,11 @@ func (r *raftNode) start(rh *raftReadyHandler) {
 
 					atomic.StoreUint64(&r.lead, rd.SoftState.Lead)
 					islead = rd.RaftState == raft.StateLeader
+					if islead {
+						isLeader.Set(1)
+					} else {
+						isLeader.Set(0)
+					}
 					rh.updateLeadership(newLeader)
 					r.td.Reset()
 				}
@@ -332,6 +346,7 @@ func (r *raftNode) processMessages(ms []raftpb.Message) []raftpb.Message {
 				// TODO: limit request rate.
 				plog.Warningf("failed to send out heartbeat on time (exceeded the %v timeout for %v)", r.heartbeat, exceed)
 				plog.Warningf("server is likely overloaded")
+				heartbeatSendFailures.Inc()
 			}
 		}
 	}
@@ -368,13 +383,13 @@ func (r *raftNode) resumeSending() {
 	p.Resume()
 }
 
-// advanceTicksForElection advances ticks to the node for fast election.
-// This reduces the time to wait for first leader election if bootstrapping the whole
-// cluster, while leaving at least 1 heartbeat for possible existing leader
-// to contact it.
-func advanceTicksForElection(n raft.Node, electionTicks int) {
-	for i := 0; i < electionTicks-1; i++ {
-		n.Tick()
+// advanceTicks advances ticks of Raft node.
+// This can be used for fast-forwarding election
+// ticks in multi data-center deployments, thus
+// speeding up election process.
+func (r *raftNode) advanceTicks(ticks int) {
+	for i := 0; i < ticks; i++ {
+		r.tick()
 	}
 }
 
@@ -415,8 +430,8 @@ func startNode(cfg *ServerConfig, cl *membership.RaftCluster, ids []types.ID) (i
 	raftStatusMu.Lock()
 	raftStatus = n.Status
 	raftStatusMu.Unlock()
-	advanceTicksForElection(n, c.ElectionTick)
-	return
+
+	return id, n, s, w
 }
 
 func restartNode(cfg *ServerConfig, snapshot *raftpb.Snapshot) (types.ID, *membership.RaftCluster, raft.Node, *raft.MemoryStorage, *wal.WAL) {
@@ -449,7 +464,6 @@ func restartNode(cfg *ServerConfig, snapshot *raftpb.Snapshot) (types.ID, *membe
 	raftStatusMu.Lock()
 	raftStatus = n.Status
 	raftStatusMu.Unlock()
-	advanceTicksForElection(n, c.ElectionTick)
 	return id, cl, n, s, w
 }
 
@@ -498,6 +512,7 @@ func restartAsStandaloneNode(cfg *ServerConfig, snapshot *raftpb.Snapshot) (type
 		Storage:         s,
 		MaxSizePerMsg:   maxSizePerMsg,
 		MaxInflightMsgs: maxInflightMsgs,
+		CheckQuorum:     true,
 	}
 	n := raft.RestartNode(c)
 	raftStatus = n.Status

vendor/github.com/coreos/etcd/etcdserver/server.go

@@ -513,11 +513,56 @@ func NewServer(cfg *ServerConfig) (srv *EtcdServer, err error) {
 	return srv, nil
 }
 
-// Start prepares and starts server in a new goroutine. It is no longer safe to
-// modify a server's fields after it has been sent to Start.
-// It also starts a goroutine to publish its server information.
+func (s *EtcdServer) adjustTicks() {
+	clusterN := len(s.cluster.Members())
+
+	// single-node fresh start, or single-node recovers from snapshot
+	if clusterN == 1 {
+		ticks := s.Cfg.ElectionTicks - 1
+		plog.Infof("%s as single-node; fast-forwarding %d ticks (election ticks %d)", s.ID(), ticks, s.Cfg.ElectionTicks)
+		s.r.advanceTicks(ticks)
+		return
+	}
+
+	if !s.Cfg.InitialElectionTickAdvance {
+		plog.Infof("skipping initial election tick advance (election tick %d)", s.Cfg.ElectionTicks)
+		return
+	}
+
+	// retry up to "rafthttp.ConnReadTimeout", which is 5-sec
+	// until peer connection reports; otherwise:
+	// 1. all connections failed, or
+	// 2. no active peers, or
+	// 3. restarted single-node with no snapshot
+	// then, do nothing, because advancing ticks would have no effect
+	waitTime := rafthttp.ConnReadTimeout
+	itv := 50 * time.Millisecond
+	for i := int64(0); i < int64(waitTime/itv); i++ {
+		select {
+		case <-time.After(itv):
+		case <-s.stopping:
+			return
+		}
+
+		peerN := s.r.transport.ActivePeers()
+		if peerN > 1 {
+			// multi-node received peer connection reports
+			// adjust ticks, in case slow leader message receive
+			ticks := s.Cfg.ElectionTicks - 2
+			plog.Infof("%s initialzed peer connection; fast-forwarding %d ticks (election ticks %d) with %d active peer(s)", s.ID(), ticks, s.Cfg.ElectionTicks, peerN)
+			s.r.advanceTicks(ticks)
+			return
+		}
+	}
+}
+
+// Start performs any initialization of the Server necessary for it to
+// begin serving requests. It must be called before Do or Process.
+// Start must be non-blocking; any long-running server functionality
+// should be implemented in goroutines.
 func (s *EtcdServer) Start() {
 	s.start()
+	s.goAttach(func() { s.adjustTicks() })
 	s.goAttach(func() { s.publish(s.Cfg.ReqTimeout()) })
 	s.goAttach(s.purgeFile)
 	s.goAttach(func() { monitorFileDescriptor(s.stopping) })
@@ -552,18 +597,21 @@ func (s *EtcdServer) start() {
 }
 
 func (s *EtcdServer) purgeFile() {
-	var serrc, werrc <-chan error
+	var dberrc, serrc, werrc <-chan error
 	if s.Cfg.MaxSnapFiles > 0 {
+		dberrc = fileutil.PurgeFile(s.Cfg.SnapDir(), "snap.db", s.Cfg.MaxSnapFiles, purgeFileInterval, s.done)
 		serrc = fileutil.PurgeFile(s.Cfg.SnapDir(), "snap", s.Cfg.MaxSnapFiles, purgeFileInterval, s.done)
 	}
 	if s.Cfg.MaxWALFiles > 0 {
 		werrc = fileutil.PurgeFile(s.Cfg.WALDir(), "wal", s.Cfg.MaxWALFiles, purgeFileInterval, s.done)
 	}
 	select {
-	case e := <-werrc:
-		plog.Fatalf("failed to purge wal file %v", e)
+	case e := <-dberrc:
+		plog.Fatalf("failed to purge snap db file %v", e)
 	case e := <-serrc:
 		plog.Fatalf("failed to purge snap file %v", e)
+	case e := <-werrc:
+		plog.Fatalf("failed to purge wal file %v", e)
 	case <-s.stopping:
 		return
 	}
@@ -743,8 +791,13 @@ func (s *EtcdServer) run() {
 					}
 					lid := lease.ID
 					s.goAttach(func() {
-						s.LeaseRevoke(s.ctx, &pb.LeaseRevokeRequest{ID: int64(lid)})
-						leaseExpired.Inc()
+						_, lerr := s.LeaseRevoke(s.ctx, &pb.LeaseRevokeRequest{ID: int64(lid)})
+						if lerr == nil {
+							leaseExpired.Inc()
+						} else {
+							plog.Warningf("failed to revoke %016x (%q)", lid, lerr.Error())
+						}
+
 						<-c
 					})
 				}
@@ -765,14 +818,8 @@ func (s *EtcdServer) run() {
 
 func (s *EtcdServer) applyAll(ep *etcdProgress, apply *apply) {
 	s.applySnapshot(ep, apply)
-	st := time.Now()
 	s.applyEntries(ep, apply)
-	d := time.Since(st)
-	entriesNum := len(apply.entries)
-	if entriesNum != 0 && d > time.Duration(entriesNum)*warnApplyDuration {
-		plog.Warningf("apply entries took too long [%v for %d entries]", d, len(apply.entries))
-		plog.Warningf("avoid queries with large range/delete range!")
-	}
+
 	proposalsApplied.Set(float64(ep.appliedi))
 	s.applyWait.Trigger(ep.appliedi)
 	// wait for the raft routine to finish the disk writes before triggering a

vendor/github.com/coreos/etcd/etcdserver/stats/server.go

@@ -74,10 +74,10 @@ type serverStats struct {
 func (ss *ServerStats) JSON() []byte {
 	ss.Lock()
 	stats := ss.serverStats
-	ss.Unlock()
-	stats.LeaderInfo.Uptime = time.Since(stats.LeaderInfo.StartTime).String()
 	stats.SendingPkgRate, stats.SendingBandwidthRate = stats.sendRateQueue.Rate()
 	stats.RecvingPkgRate, stats.RecvingBandwidthRate = stats.recvRateQueue.Rate()
+	stats.LeaderInfo.Uptime = time.Since(stats.LeaderInfo.StartTime).String()
+	ss.Unlock()
 	b, err := json.Marshal(stats)
 	// TODO(jonboulle): appropriate error handling?
 	if err != nil {

vendor/github.com/coreos/etcd/etcdserver/util.go

@@ -15,11 +15,16 @@
 package etcdserver
 
 import (
+	"fmt"
+	"reflect"
+	"strings"
 	"time"
 
+	pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
 	"github.com/coreos/etcd/etcdserver/membership"
 	"github.com/coreos/etcd/pkg/types"
 	"github.com/coreos/etcd/rafthttp"
+	"github.com/golang/protobuf/proto"
 )
 
 // isConnectedToQuorumSince checks whether the local member is connected to the
@@ -95,3 +100,56 @@ func (nc *notifier) notify(err error) {
 	nc.err = err
 	close(nc.c)
 }
+
+func warnOfExpensiveRequest(now time.Time, reqStringer fmt.Stringer, respMsg proto.Message, err error) {
+	var resp string
+	if !isNil(respMsg) {
+		resp = fmt.Sprintf("size:%d", proto.Size(respMsg))
+	}
+	warnOfExpensiveGenericRequest(now, reqStringer, "", resp, err)
+}
+
+func warnOfExpensiveReadOnlyTxnRequest(now time.Time, r *pb.TxnRequest, txnResponse *pb.TxnResponse, err error) {
+	reqStringer := pb.NewLoggableTxnRequest(r)
+	var resp string
+	if !isNil(txnResponse) {
+		var resps []string
+		for _, r := range txnResponse.Responses {
+			switch op := r.Response.(type) {
+			case *pb.ResponseOp_ResponseRange:
+				resps = append(resps, fmt.Sprintf("range_response_count:%d", len(op.ResponseRange.Kvs)))
+			default:
+				// only range responses should be in a read only txn request
+			}
+		}
+		resp = fmt.Sprintf("responses:<%s> size:%d", strings.Join(resps, " "), proto.Size(txnResponse))
+	}
+	warnOfExpensiveGenericRequest(now, reqStringer, "read-only range ", resp, err)
+}
+
+func warnOfExpensiveReadOnlyRangeRequest(now time.Time, reqStringer fmt.Stringer, rangeResponse *pb.RangeResponse, err error) {
+	var resp string
+	if !isNil(rangeResponse) {
+		resp = fmt.Sprintf("range_response_count:%d size:%d", len(rangeResponse.Kvs), proto.Size(rangeResponse))
+	}
+	warnOfExpensiveGenericRequest(now, reqStringer, "read-only range ", resp, err)
+}
+
+func warnOfExpensiveGenericRequest(now time.Time, reqStringer fmt.Stringer, prefix string, resp string, err error) {
+	// TODO: add metrics
+	d := time.Since(now)
+	if d > warnApplyDuration {
+		var result string
+		if err != nil {
+			result = fmt.Sprintf("error:%v", err)
+		} else {
+			result = resp
+		}
+		plog.Warningf("%srequest %q with result %q took too long (%v) to execute", prefix, reqStringer.String(), result, d)
+		slowApplies.Inc()
+	}
+}
+
+func isNil(msg proto.Message) bool {
+	return msg == nil || reflect.ValueOf(msg).IsNil()
+}

vendor/github.com/coreos/etcd/etcdserver/v3_server.go

@@ -19,8 +19,6 @@ import (
 	"encoding/binary"
 	"time"
 
-	"github.com/gogo/protobuf/proto"
-
 	"github.com/coreos/etcd/auth"
 	pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
 	"github.com/coreos/etcd/etcdserver/membership"
@@ -28,7 +26,7 @@ import (
 	"github.com/coreos/etcd/lease/leasehttp"
 	"github.com/coreos/etcd/mvcc"
 	"github.com/coreos/etcd/raft"
-
+	"github.com/gogo/protobuf/proto"
 	"golang.org/x/net/context"
 )
 
@@ -82,20 +80,26 @@ type Authenticator interface {
 }
 
 func (s *EtcdServer) Range(ctx context.Context, r *pb.RangeRequest) (*pb.RangeResponse, error) {
+	var resp *pb.RangeResponse
+	var err error
+	defer func(start time.Time) {
+		warnOfExpensiveReadOnlyRangeRequest(start, r, resp, err)
+	}(time.Now())
+
 	if !r.Serializable {
-		err := s.linearizableReadNotify(ctx)
+		err = s.linearizableReadNotify(ctx)
 		if err != nil {
 			return nil, err
 		}
 	}
-	var resp *pb.RangeResponse
-	var err error
 	chk := func(ai *auth.AuthInfo) error {
 		return s.authStore.IsRangePermitted(ai, r.Key, r.RangeEnd)
 	}
+
 	get := func() { resp, err = s.applyV3Base.Range(nil, r) }
 	if serr := s.doSerialize(ctx, chk, get); serr != nil {
-		return nil, serr
+		err = serr
+		return nil, err
 	}
 	return resp, err
 }
@@ -129,12 +133,18 @@ func (s *EtcdServer) Txn(ctx context.Context, r *pb.TxnRequest) (*pb.TxnResponse
 		chk := func(ai *auth.AuthInfo) error {
 			return checkTxnAuth(s.authStore, ai, r)
 		}
+
+		defer func(start time.Time) {
+			warnOfExpensiveReadOnlyTxnRequest(start, r, resp, err)
+		}(time.Now())
+
 		get := func() { resp, err = s.applyV3Base.Txn(r) }
 		if serr := s.doSerialize(ctx, chk, get); serr != nil {
 			return nil, serr
 		}
 		return resp, err
 	}
+
 	resp, err := s.raftRequest(ctx, pb.InternalRaftRequest{Txn: r})
 	if err != nil {
 		return nil, err
@@ -587,8 +597,9 @@ func (s *EtcdServer) linearizableReadLoop() {
 	var rs raft.ReadState
 
 	for {
-		ctx := make([]byte, 8)
-		binary.BigEndian.PutUint64(ctx, s.reqIDGen.Next())
+		ctxToSend := make([]byte, 8)
+		id1 := s.reqIDGen.Next()
+		binary.BigEndian.PutUint64(ctxToSend, id1)
 
 		select {
 		case <-s.readwaitc:
@@ -604,7 +615,7 @@ func (s *EtcdServer) linearizableReadLoop() {
 		s.readMu.Unlock()
 
 		cctx, cancel := context.WithTimeout(context.Background(), s.Cfg.ReqTimeout())
-		if err := s.r.ReadIndex(cctx, ctx); err != nil {
+		if err := s.r.ReadIndex(cctx, ctxToSend); err != nil {
 			cancel()
 			if err == raft.ErrStopped {
 				return
@@ -622,16 +633,24 @@ func (s *EtcdServer) linearizableReadLoop() {
 		for !timeout && !done {
 			select {
 			case rs = <-s.r.readStateC:
-				done = bytes.Equal(rs.RequestCtx, ctx)
+				done = bytes.Equal(rs.RequestCtx, ctxToSend)
 				if !done {
 					// a previous request might time out. now we should ignore the response of it and
 					// continue waiting for the response of the current requests.
-					plog.Warningf("ignored out-of-date read index response (want %v, got %v)", rs.RequestCtx, ctx)
+					id2 := uint64(0)
+					if len(rs.RequestCtx) == 8 {
+						id2 = binary.BigEndian.Uint64(rs.RequestCtx)
+					}
+					plog.Warningf("ignored out-of-date read index response; local node read indexes queueing up and waiting to be in sync with leader (request ID want %d, got %d)", id1, id2)
+					slowReadIndex.Inc()
 				}
+
 			case <-time.After(s.Cfg.ReqTimeout()):
 				plog.Warningf("timed out waiting for read index response")
 				nr.notify(ErrTimeout)
 				timeout = true
+				slowReadIndex.Inc()
+
 			case <-s.stopping:
 				return
 			}
@@ -681,6 +700,5 @@ func (s *EtcdServer) AuthInfoFromCtx(ctx context.Context) (*auth.AuthInfo, error
 			return authInfo, nil
 		}
 	}
-
 	return s.AuthStore().AuthInfoFromCtx(ctx)
 }

vendor/github.com/coreos/etcd/integration/BUILD

@@ -26,6 +26,7 @@ go_library(
         "//vendor/github.com/coreos/etcd/etcdserver/api/v3rpc:go_default_library",
         "//vendor/github.com/coreos/etcd/etcdserver/etcdserverpb:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/testutil:go_default_library",
+        "//vendor/github.com/coreos/etcd/pkg/tlsutil:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/transport:go_default_library",
         "//vendor/github.com/coreos/etcd/pkg/types:go_default_library",
         "//vendor/github.com/coreos/etcd/rafthttp:go_default_library",

vendor/github.com/coreos/etcd/integration/cluster.go

@@ -45,6 +45,7 @@ import (
 	"github.com/coreos/etcd/etcdserver/api/v3rpc"
 	pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
 	"github.com/coreos/etcd/pkg/testutil"
+	"github.com/coreos/etcd/pkg/tlsutil"
 	"github.com/coreos/etcd/pkg/transport"
 	"github.com/coreos/etcd/pkg/types"
 	"github.com/coreos/etcd/rafthttp"
@@ -562,6 +563,7 @@ func mustNewMember(t *testing.T, mcfg memberConfig) *member {
 		m.ServerConfig.PeerTLSInfo = *m.PeerTLSInfo
 	}
 	m.ElectionTicks = electionTicks
+	m.InitialElectionTickAdvance = true
 	m.TickMs = uint(tickDuration / time.Millisecond)
 	m.QuotaBackendBytes = mcfg.quotaBackendBytes
 	m.MaxRequestBytes = mcfg.maxRequestBytes
@@ -695,10 +697,16 @@ func (m *member) Launch() error {
 		if m.PeerTLSInfo == nil {
 			hs.Start()
 		} else {
-			hs.TLS, err = m.PeerTLSInfo.ServerConfig()
+			info := m.PeerTLSInfo
+			hs.TLS, err = info.ServerConfig()
 			if err != nil {
 				return err
 			}
+			tlsCert, err := tlsutil.NewCert(info.CertFile, info.KeyFile, nil)
+			if err != nil {
+				return err
+			}
+			hs.TLS.Certificates = []tls.Certificate{*tlsCert}
 			hs.StartTLS()
 		}
 		m.hss = append(m.hss, hs)
@@ -711,10 +719,45 @@ func (m *member) Launch() error {
 		if m.ClientTLSInfo == nil {
 			hs.Start()
 		} else {
-			hs.TLS, err = m.ClientTLSInfo.ServerConfig()
+			info := m.ClientTLSInfo
+			hs.TLS, err = info.ServerConfig()
+			if err != nil {
+				return err
+			}
+
+			// baseConfig is called on initial TLS handshake start.
+			//
+			// Previously,
+			// 1. Server has non-empty (*tls.Config).Certificates on client hello
+			// 2. Server calls (*tls.Config).GetCertificate iff:
+			//    - Server's (*tls.Config).Certificates is not empty, or
+			//    - Client supplies SNI; non-empty (*tls.ClientHelloInfo).ServerName
+			//
+			// When (*tls.Config).Certificates is always populated on initial handshake,
+			// client is expected to provide a valid matching SNI to pass the TLS
+			// verification, thus trigger server (*tls.Config).GetCertificate to reload
+			// TLS assets. However, a cert whose SAN field does not include domain names
+			// but only IP addresses, has empty (*tls.ClientHelloInfo).ServerName, thus
+			// it was never able to trigger TLS reload on initial handshake; first
+			// ceritifcate object was being used, never being updated.
+			//
+			// Now, (*tls.Config).Certificates is created empty on initial TLS client
+			// handshake, in order to trigger (*tls.Config).GetCertificate and populate
+			// rest of the certificates on every new TLS connection, even when client
+			// SNI is empty (e.g. cert only includes IPs).
+			//
+			// This introduces another problem with "httptest.Server":
+			// when server initial certificates are empty, certificates
+			// are overwritten by Go's internal test certs, which have
+			// different SAN fields (e.g. example.com). To work around,
+			// re-overwrite (*tls.Config).Certificates before starting
+			// test server.
+			tlsCert, err := tlsutil.NewCert(info.CertFile, info.KeyFile, nil)
 			if err != nil {
 				return err
 			}
+			hs.TLS.Certificates = []tls.Certificate{*tlsCert}
+
 			hs.StartTLS()
 		}
 		m.hss = append(m.hss, hs)

vendor/github.com/coreos/etcd/lease/lessor.go

@@ -35,15 +35,19 @@ const (
 	forever = monotime.Time(math.MaxInt64)
 )
 
+// MaxLeaseTTL is the maximum lease TTL value
+const MaxLeaseTTL = 9000000000
+
 var (
 	leaseBucketName = []byte("lease")
 
 	// maximum number of leases to revoke per second; configurable for tests
 	leaseRevokeRate = 1000
 
-	ErrNotPrimary    = errors.New("not a primary lessor")
-	ErrLeaseNotFound = errors.New("lease not found")
-	ErrLeaseExists   = errors.New("lease already exists")
+	ErrNotPrimary       = errors.New("not a primary lessor")
+	ErrLeaseNotFound    = errors.New("lease not found")
+	ErrLeaseExists      = errors.New("lease already exists")
+	ErrLeaseTTLTooLarge = errors.New("too large lease TTL")
 )
 
 // TxnDelete is a TxnWrite that only permits deletes. Defined here
@@ -199,6 +203,10 @@ func (le *lessor) Grant(id LeaseID, ttl int64) (*Lease, error) {
 		return nil, ErrLeaseNotFound
 	}
 
+	if ttl > MaxLeaseTTL {
+		return nil, ErrLeaseTTLTooLarge
+	}
+
 	// TODO: when lessor is under high load, it should give out lease
 	// with longer TTL to reduce renew load.
 	l := &Lease{

vendor/github.com/coreos/etcd/mvcc/backend/backend.go

@@ -54,6 +54,10 @@ type Backend interface {
 	Hash(ignores map[IgnoreKey]struct{}) (uint32, error)
 	// Size returns the current size of the backend.
 	Size() int64
+	// SizeInUse returns the current size of the backend logically in use.
+	// Since the backend can manage free space in a non-byte unit such as
+	// number of pages, the returned value can be not exactly accurate in bytes.
+	SizeInUse() int64
 	Defrag() error
 	ForceCommit()
 	Close() error
@@ -74,6 +78,10 @@ type backend struct {
 
 	// size is the number of bytes in the backend
 	size int64
+
+	// sizeInUse is the number of bytes actually used in the backend
+	sizeInUse int64
+
 	// commits counts number of commits since start
 	commits int64
 
@@ -244,6 +252,10 @@ func (b *backend) Size() int64 {
 	return atomic.LoadInt64(&b.size)
 }
 
+func (b *backend) SizeInUse() int64 {
+	return atomic.LoadInt64(&b.sizeInUse)
+}
+
 func (b *backend) run() {
 	defer close(b.donec)
 	t := time.NewTimer(b.batchInterval)
@@ -272,18 +284,12 @@ func (b *backend) Commits() int64 {
 }
 
 func (b *backend) Defrag() error {
-	err := b.defrag()
-	if err != nil {
-		return err
-	}
-
-	// commit to update metadata like db.size
-	b.batchTx.Commit()
-
-	return nil
+	return b.defrag()
 }
 
 func (b *backend) defrag() error {
+	now := time.Now()
+
 	// TODO: make this non-blocking?
 	// lock batchTx to ensure nobody is using previous tx, and then
 	// close previous ongoing tx.
@@ -341,7 +347,14 @@ func (b *backend) defrag() error {
 
 	b.readTx.buf.reset()
 	b.readTx.tx = b.unsafeBegin(false)
-	atomic.StoreInt64(&b.size, b.readTx.tx.Size())
+
+	size := b.readTx.tx.Size()
+	db := b.db
+	atomic.StoreInt64(&b.size, size)
+	atomic.StoreInt64(&b.sizeInUse, size-(int64(db.Stats().FreePageN)*int64(db.Info().PageSize)))
+
+	took := time.Since(now)
+	defragDurations.Observe(took.Seconds())
 
 	return nil
 }
@@ -370,10 +383,10 @@ func defragdb(odb, tmpdb *bolt.DB, limit int) error {
 		}
 
 		tmpb, berr := tmptx.CreateBucketIfNotExists(next)
-		tmpb.FillPercent = 0.9 // for seq write in for each
 		if berr != nil {
 			return berr
 		}
+		tmpb.FillPercent = 0.9 // for seq write in for each
 
 		b.ForEach(func(k, v []byte) error {
 			count++
@@ -402,7 +415,12 @@ func (b *backend) begin(write bool) *bolt.Tx {
 	b.mu.RLock()
 	tx := b.unsafeBegin(write)
 	b.mu.RUnlock()
-	atomic.StoreInt64(&b.size, tx.Size())
+
+	size := tx.Size()
+	db := tx.DB()
+	atomic.StoreInt64(&b.size, size)
+	atomic.StoreInt64(&b.sizeInUse, size-(int64(db.Stats().FreePageN)*int64(db.Info().PageSize)))
+
 	return tx
 }
 

vendor/github.com/coreos/etcd/mvcc/backend/batch_tx.go

@@ -141,15 +141,15 @@ func unsafeForEach(tx *bolt.Tx, bucket []byte, visitor func(k, v []byte) error)
 // Commit commits a previous tx and begins a new writable one.
 func (t *batchTx) Commit() {
 	t.Lock()
-	defer t.Unlock()
 	t.commit(false)
+	t.Unlock()
 }
 
 // CommitAndStop commits the previous tx and does not create a new one.
 func (t *batchTx) CommitAndStop() {
 	t.Lock()
-	defer t.Unlock()
 	t.commit(true)
+	t.Unlock()
 }
 
 func (t *batchTx) Unlock() {
@@ -163,21 +163,15 @@ func (t *batchTx) commit(stop bool) {
 	// commit the last tx
 	if t.tx != nil {
 		if t.pending == 0 && !stop {
-			t.backend.mu.RLock()
-			defer t.backend.mu.RUnlock()
-
-			// t.tx.DB()==nil if 'CommitAndStop' calls 'batchTx.commit(true)',
-			// which initializes *bolt.Tx.db and *bolt.Tx.meta as nil; panics t.tx.Size().
-			// Server must make sure 'batchTx.commit(false)' does not follow
-			// 'batchTx.commit(true)' (e.g. stopping backend, and inflight Hash call).
-			atomic.StoreInt64(&t.backend.size, t.tx.Size())
 			return
 		}
 
 		start := time.Now()
+
 		// gofail: var beforeCommit struct{}
 		err := t.tx.Commit()
 		// gofail: var afterCommit struct{}
+
 		commitDurations.Observe(time.Since(start).Seconds())
 		atomic.AddInt64(&t.backend.commits, 1)
 
@@ -222,21 +216,21 @@ func (t *batchTxBuffered) Unlock() {
 
 func (t *batchTxBuffered) Commit() {
 	t.Lock()
-	defer t.Unlock()
 	t.commit(false)
+	t.Unlock()
 }
 
 func (t *batchTxBuffered) CommitAndStop() {
 	t.Lock()
-	defer t.Unlock()
 	t.commit(true)
+	t.Unlock()
 }
 
 func (t *batchTxBuffered) commit(stop bool) {
 	// all read txs must be closed to acquire boltdb commit rwlock
 	t.backend.readTx.mu.Lock()
-	defer t.backend.readTx.mu.Unlock()
 	t.unsafeCommit(stop)
+	t.backend.readTx.mu.Unlock()
 }
 
 func (t *batchTxBuffered) unsafeCommit(stop bool) {

vendor/github.com/coreos/etcd/mvcc/backend/metrics.go

@@ -22,7 +22,22 @@ var (
 		Subsystem: "disk",
 		Name:      "backend_commit_duration_seconds",
 		Help:      "The latency distributions of commit called by backend.",
-		Buckets:   prometheus.ExponentialBuckets(0.001, 2, 14),
+
+		// lowest bucket start of upper bound 0.001 sec (1 ms) with factor 2
+		// highest bucket start of 0.001 sec * 2^13 == 8.192 sec
+		Buckets: prometheus.ExponentialBuckets(0.001, 2, 14),
+	})
+
+	defragDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
+		Namespace: "etcd",
+		Subsystem: "disk",
+		Name:      "backend_defrag_duration_seconds",
+		Help:      "The latency distribution of backend defragmentation.",
+
+		// 100 MB usually takes 1 sec, so start with 10 MB of 100 ms
+		// lowest bucket start of upper bound 0.1 sec (100 ms) with factor 2
+		// highest bucket start of 0.1 sec * 2^12 == 409.6 sec
+		Buckets: prometheus.ExponentialBuckets(.1, 2, 13),
 	})
 
 	snapshotDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
@@ -30,12 +45,15 @@ var (
 		Subsystem: "disk",
 		Name:      "backend_snapshot_duration_seconds",
 		Help:      "The latency distribution of backend snapshots.",
-		// 10 ms -> 655 seconds
+
+		// lowest bucket start of upper bound 0.01 sec (10 ms) with factor 2
+		// highest bucket start of 0.01 sec * 2^16 == 655.36 sec
 		Buckets: prometheus.ExponentialBuckets(.01, 2, 17),
 	})
 )
 
 func init() {
 	prometheus.MustRegister(commitDurations)
+	prometheus.MustRegister(defragDurations)
 	prometheus.MustRegister(snapshotDurations)
 }

vendor/github.com/coreos/etcd/mvcc/kvstore.go

@@ -150,8 +150,12 @@ func (s *store) compactBarrier(ctx context.Context, ch chan struct{}) {
 }
 
 func (s *store) Hash() (hash uint32, revision int64, err error) {
+	start := time.Now()
+
 	s.b.ForceCommit()
 	h, err := s.b.Hash(DefaultIgnores)
+
+	hashDurations.Observe(time.Since(start).Seconds())
 	return h, s.currentRev, err
 }
 
@@ -245,10 +249,14 @@ func (s *store) Restore(b backend.Backend) error {
 }
 
 func (s *store) restore() error {
-	reportDbTotalSizeInBytesMu.Lock()
 	b := s.b
+
+	reportDbTotalSizeInBytesMu.Lock()
 	reportDbTotalSizeInBytes = func() float64 { return float64(b.Size()) }
 	reportDbTotalSizeInBytesMu.Unlock()
+	reportDbTotalSizeInUseInBytesMu.Lock()
+	reportDbTotalSizeInUseInBytes = func() float64 { return float64(b.SizeInUse()) }
+	reportDbTotalSizeInUseInBytesMu.Unlock()
 
 	min, max := newRevBytes(), newRevBytes()
 	revToBytes(revision{main: 1}, min)

vendor/github.com/coreos/etcd/mvcc/metrics.go

@@ -131,11 +131,23 @@ var (
 			Buckets: prometheus.ExponentialBuckets(100, 2, 14),
 		})
 
-	dbTotalSize = prometheus.NewGaugeFunc(prometheus.GaugeOpts{
+	dbTotalSizeDebugging = prometheus.NewGaugeFunc(prometheus.GaugeOpts{
 		Namespace: "etcd_debugging",
 		Subsystem: "mvcc",
 		Name:      "db_total_size_in_bytes",
-		Help:      "Total size of the underlying database in bytes.",
+		Help:      "Total size of the underlying database physically allocated in bytes. Use etcd_mvcc_db_total_size_in_bytes",
+	},
+		func() float64 {
+			reportDbTotalSizeInBytesMu.RLock()
+			defer reportDbTotalSizeInBytesMu.RUnlock()
+			return reportDbTotalSizeInBytes()
+		},
+	)
+	dbTotalSize = prometheus.NewGaugeFunc(prometheus.GaugeOpts{
+		Namespace: "etcd",
+		Subsystem: "mvcc",
+		Name:      "db_total_size_in_bytes",
+		Help:      "Total size of the underlying database physically allocated in bytes.",
 	},
 		func() float64 {
 			reportDbTotalSizeInBytesMu.RLock()
@@ -145,7 +157,35 @@ var (
 	)
 	// overridden by mvcc initialization
 	reportDbTotalSizeInBytesMu sync.RWMutex
-	reportDbTotalSizeInBytes   func() float64 = func() float64 { return 0 }
+	reportDbTotalSizeInBytes   = func() float64 { return 0 }
+
+	dbTotalSizeInUse = prometheus.NewGaugeFunc(prometheus.GaugeOpts{
+		Namespace: "etcd",
+		Subsystem: "mvcc",
+		Name:      "db_total_size_in_use_in_bytes",
+		Help:      "Total size of the underlying database logically in use in bytes.",
+	},
+		func() float64 {
+			reportDbTotalSizeInUseInBytesMu.RLock()
+			defer reportDbTotalSizeInUseInBytesMu.RUnlock()
+			return reportDbTotalSizeInUseInBytes()
+		},
+	)
+	// overridden by mvcc initialization
+	reportDbTotalSizeInUseInBytesMu sync.RWMutex
+	reportDbTotalSizeInUseInBytes   func() float64 = func() float64 { return 0 }
+
+	hashDurations = prometheus.NewHistogram(prometheus.HistogramOpts{
+		Namespace: "etcd",
+		Subsystem: "mvcc",
+		Name:      "hash_duration_seconds",
+		Help:      "The latency distribution of storage hash operation.",
+
+		// 100 MB usually takes 100 ms, so start with 10 MB of 10 ms
+		// lowest bucket start of upper bound 0.01 sec (10 ms) with factor 2
+		// highest bucket start of 0.01 sec * 2^14 == 163.84 sec
+		Buckets: prometheus.ExponentialBuckets(.01, 2, 15),
+	})
 )
 
 func init() {
@@ -162,7 +202,10 @@ func init() {
 	prometheus.MustRegister(indexCompactionPauseDurations)
 	prometheus.MustRegister(dbCompactionPauseDurations)
 	prometheus.MustRegister(dbCompactionTotalDurations)
+	prometheus.MustRegister(dbTotalSizeDebugging)
 	prometheus.MustRegister(dbTotalSize)
+	prometheus.MustRegister(dbTotalSizeInUse)
+	prometheus.MustRegister(hashDurations)
 }
 
 // ReportEventReceived reports that an event is received.

vendor/github.com/coreos/etcd/mvcc/watchable_store.go

@@ -188,7 +188,8 @@ func (s *watchableStore) Restore(b backend.Backend) error {
 	}
 
 	for wa := range s.synced.watchers {
-		s.unsynced.watchers.add(wa)
+		wa.restore = true
+		s.unsynced.add(wa)
 	}
 	s.synced = newWatcherGroup()
 	return nil
@@ -479,6 +480,14 @@ type watcher struct {
 	// compacted is set when the watcher is removed because of compaction
 	compacted bool
 
+	// restore is true when the watcher is being restored from leader snapshot
+	// which means that this watcher has just been moved from "synced" to "unsynced"
+	// watcher group, possibly with a future revision when it was first added
+	// to the synced watcher
+	// "unsynced" watcher revision must always be <= current revision,
+	// except when the watcher were to be moved from "synced" watcher group
+	restore bool
+
 	// minRev is the minimum revision update the watcher will accept
 	minRev int64
 	id     WatchID

vendor/github.com/coreos/etcd/mvcc/watcher_group.go

@@ -15,6 +15,7 @@
 package mvcc
 
 import (
+	"fmt"
 	"math"
 
 	"github.com/coreos/etcd/mvcc/mvccpb"
@@ -238,7 +239,15 @@ func (wg *watcherGroup) chooseAll(curRev, compactRev int64) int64 {
 	minRev := int64(math.MaxInt64)
 	for w := range wg.watchers {
 		if w.minRev > curRev {
-			panic("watcher current revision should not exceed current revision")
+			// after network partition, possibly choosing future revision watcher from restore operation
+			// with watch key "proxy-namespace__lostleader" and revision "math.MaxInt64 - 2"
+			// do not panic when such watcher had been moved from "synced" watcher during restore operation
+			if !w.restore {
+				panic(fmt.Errorf("watcher minimum revision %d should not exceed current revision %d", w.minRev, curRev))
+			}
+
+			// mark 'restore' done, since it's chosen
+			w.restore = false
 		}
 		if w.minRev < compactRev {
 			select {

vendor/github.com/coreos/etcd/pkg/tlsutil/BUILD

@@ -3,6 +3,7 @@ load("@io_bazel_rules_go//go:def.bzl", "go_library")
 go_library(
     name = "go_default_library",
     srcs = [
+        "cipher_suites.go",
         "doc.go",
         "tlsutil.go",
     ],

vendor/github.com/coreos/etcd/pkg/tlsutil/cipher_suites.go

+// Copyright 2018 The etcd 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 tlsutil
+
+import "crypto/tls"
+
+// cipher suites implemented by Go
+// https://github.com/golang/go/blob/dev.boringcrypto.go1.10/src/crypto/tls/cipher_suites.go
+var cipherSuites = map[string]uint16{
+	"TLS_RSA_WITH_RC4_128_SHA":                tls.TLS_RSA_WITH_RC4_128_SHA,
+	"TLS_RSA_WITH_3DES_EDE_CBC_SHA":           tls.TLS_RSA_WITH_3DES_EDE_CBC_SHA,
+	"TLS_RSA_WITH_AES_128_CBC_SHA":            tls.TLS_RSA_WITH_AES_128_CBC_SHA,
+	"TLS_RSA_WITH_AES_256_CBC_SHA":            tls.TLS_RSA_WITH_AES_256_CBC_SHA,
+	"TLS_RSA_WITH_AES_128_CBC_SHA256":         tls.TLS_RSA_WITH_AES_128_CBC_SHA256,
+	"TLS_RSA_WITH_AES_128_GCM_SHA256":         tls.TLS_RSA_WITH_AES_128_GCM_SHA256,
+	"TLS_RSA_WITH_AES_256_GCM_SHA384":         tls.TLS_RSA_WITH_AES_256_GCM_SHA384,
+	"TLS_ECDHE_ECDSA_WITH_RC4_128_SHA":        tls.TLS_ECDHE_ECDSA_WITH_RC4_128_SHA,
+	"TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA":    tls.TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA,
+	"TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA":    tls.TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA,
+	"TLS_ECDHE_RSA_WITH_RC4_128_SHA":          tls.TLS_ECDHE_RSA_WITH_RC4_128_SHA,
+	"TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA":     tls.TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA,
+	"TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA":      tls.TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,
+	"TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA":      tls.TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA,
+	"TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256": tls.TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256,
+	"TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256":   tls.TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256,
+	"TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256":   tls.TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256,
+	"TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256": tls.TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256,
+	"TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384":   tls.TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384,
+	"TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384": tls.TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384,
+	"TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305":    tls.TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305,
+	"TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305":  tls.TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305,
+}
+
+// GetCipherSuite returns the corresponding cipher suite,
+// and boolean value if it is supported.
+func GetCipherSuite(s string) (uint16, bool) {
+	v, ok := cipherSuites[s]
+	return v, ok
+}

vendor/github.com/coreos/etcd/pkg/transport/listener.go

@@ -69,6 +69,11 @@ type TLSInfo struct {
 	// connection will be closed immediately afterwards.
 	HandshakeFailure func(*tls.Conn, error)
 
+	// CipherSuites is a list of supported cipher suites.
+	// If empty, Go auto-populates it by default.
+	// Note that cipher suites are prioritized in the given order.
+	CipherSuites []uint16
+
 	selfCert bool
 
 	// parseFunc exists to simplify testing. Typically, parseFunc
@@ -162,16 +167,20 @@ func (info TLSInfo) baseConfig() (*tls.Config, error) {
 		return nil, fmt.Errorf("KeyFile and CertFile must both be present[key: %v, cert: %v]", info.KeyFile, info.CertFile)
 	}
 
-	tlsCert, err := tlsutil.NewCert(info.CertFile, info.KeyFile, info.parseFunc)
+	_, err := tlsutil.NewCert(info.CertFile, info.KeyFile, info.parseFunc)
 	if err != nil {
 		return nil, err
 	}
 
 	cfg := &tls.Config{
-		Certificates: []tls.Certificate{*tlsCert},
-		MinVersion:   tls.VersionTLS12,
-		ServerName:   info.ServerName,
+		MinVersion: tls.VersionTLS12,
+		ServerName: info.ServerName,
+	}
+
+	if len(info.CipherSuites) > 0 {
+		cfg.CipherSuites = info.CipherSuites
 	}
+
 	// this only reloads certs when there's a client request
 	// TODO: support server-side refresh (e.g. inotify, SIGHUP), caching
 	cfg.GetCertificate = func(clientHello *tls.ClientHelloInfo) (*tls.Certificate, error) {

vendor/github.com/coreos/etcd/rafthttp/peer.go

@@ -225,6 +225,7 @@ func (p *peer) send(m raftpb.Message) {
 			plog.MergeWarningf("dropped internal raft message to %s since %s's sending buffer is full (bad/overloaded network)", p.id, name)
 		}
 		plog.Debugf("dropped %s to %s since %s's sending buffer is full", m.Type, p.id, name)
+		sentFailures.WithLabelValues(types.ID(m.To).String()).Inc()
 	}
 }
 

vendor/github.com/coreos/etcd/rafthttp/remote.go

@@ -53,6 +53,7 @@ func (g *remote) send(m raftpb.Message) {
 			plog.MergeWarningf("dropped internal raft message to %s since sending buffer is full (bad/overloaded network)", g.id)
 		}
 		plog.Debugf("dropped %s to %s since sending buffer is full", m.Type, g.id)
+		sentFailures.WithLabelValues(types.ID(m.To).String()).Inc()
 	}
 }
 

vendor/github.com/coreos/etcd/rafthttp/transport.go

@@ -83,6 +83,8 @@ type Transporter interface {
 	// If the connection is active since peer was added, it returns the adding time.
 	// If the connection is currently inactive, it returns zero time.
 	ActiveSince(id types.ID) time.Time
+	// ActivePeers returns the number of active peers.
+	ActivePeers() int
 	// Stop closes the connections and stops the transporter.
 	Stop()
 }
@@ -362,6 +364,20 @@ func (t *Transport) Resume() {
 	}
 }
 
+// ActivePeers returns a channel that closes when an initial
+// peer connection has been established. Use this to wait until the
+// first peer connection becomes active.
+func (t *Transport) ActivePeers() (cnt int) {
+	t.mu.RLock()
+	defer t.mu.RUnlock()
+	for _, p := range t.peers {
+		if !p.activeSince().IsZero() {
+			cnt++
+		}
+	}
+	return cnt
+}
+
 type nopTransporter struct{}
 
 func NewNopTransporter() Transporter {
@@ -378,6 +394,7 @@ func (s *nopTransporter) RemovePeer(id types.ID)              {}
 func (s *nopTransporter) RemoveAllPeers()                     {}
 func (s *nopTransporter) UpdatePeer(id types.ID, us []string) {}
 func (s *nopTransporter) ActiveSince(id types.ID) time.Time   { return time.Time{} }
+func (s *nopTransporter) ActivePeers() int                    { return 0 }
 func (s *nopTransporter) Stop()                               {}
 func (s *nopTransporter) Pause()                              {}
 func (s *nopTransporter) Resume()                             {}

vendor/github.com/coreos/etcd/version/version.go

@@ -26,7 +26,7 @@ import (
 var (
 	// MinClusterVersion is the min cluster version this etcd binary is compatible with.
 	MinClusterVersion = "3.0.0"
-	Version           = "3.2.13"
+	Version           = "3.2.24"
 	APIVersion        = "unknown"
 
 	// Git SHA Value will be set during build

vendor/k8s.io/cli-runtime

-../../staging/src/k8s.io/cli-runtime/
+../../staging/src/k8s.io/cli-runtime
\ No newline at end of file

vendor/k8s.io/kube-proxy

-../../staging/src/k8s.io/kube-proxy/
+../../staging/src/k8s.io/kube-proxy
\ No newline at end of file

vendor/k8s.io/kube-scheduler

-../../staging/src/k8s.io/kube-scheduler/
+../../staging/src/k8s.io/kube-scheduler
\ No newline at end of file