977142214c
Before these fixes, it was possible to see errors on new RPCs after a connection began draining, and before establishing a new connection. There is an inherent race between choosing a SubConn and attempting to creating a stream on it. We should be able to avoid application-visible RPC errors due to this with transparent retry. However, several bugs were preventing this from working correctly: 1. Non-wait-for-ready RPCs were skipping transparent retry, though the retry design calls for retrying them. 2. The transport closed itself (and would consequently error new RPCs) before notifying the SubConn that it was draining. 3. The SubConn wasn't synchronously updating itself once it was notified about the closing or draining state. 4. The SubConn would go into the TRANSIENT_FAILURE state instantaneously, causing RPCs to fail instead of queue.
510 lines
13 KiB
Go
510 lines
13 KiB
Go
/*
|
|
*
|
|
* Copyright 2018 gRPC 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 grpc
|
|
|
|
import (
|
|
"context"
|
|
"net"
|
|
"sync"
|
|
"testing"
|
|
"time"
|
|
|
|
"golang.org/x/net/http2"
|
|
"google.golang.org/grpc/balancer"
|
|
"google.golang.org/grpc/connectivity"
|
|
"google.golang.org/grpc/internal/testutils"
|
|
"google.golang.org/grpc/resolver"
|
|
"google.golang.org/grpc/resolver/manual"
|
|
)
|
|
|
|
const stateRecordingBalancerName = "state_recoding_balancer"
|
|
|
|
var testBalancerBuilder = newStateRecordingBalancerBuilder()
|
|
|
|
func init() {
|
|
balancer.Register(testBalancerBuilder)
|
|
}
|
|
|
|
// These tests use a pipeListener. This listener is similar to net.Listener
|
|
// except that it is unbuffered, so each read and write will wait for the other
|
|
// side's corresponding write or read.
|
|
func (s) TestStateTransitions_SingleAddress(t *testing.T) {
|
|
for _, test := range []struct {
|
|
desc string
|
|
want []connectivity.State
|
|
server func(net.Listener) net.Conn
|
|
}{
|
|
{
|
|
desc: "When the server returns server preface, the client enters READY.",
|
|
want: []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.Ready,
|
|
},
|
|
server: func(lis net.Listener) net.Conn {
|
|
conn, err := lis.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return nil
|
|
}
|
|
|
|
go keepReading(conn)
|
|
|
|
framer := http2.NewFramer(conn, conn)
|
|
if err := framer.WriteSettings(http2.Setting{}); err != nil {
|
|
t.Errorf("Error while writing settings frame. %v", err)
|
|
return nil
|
|
}
|
|
|
|
return conn
|
|
},
|
|
},
|
|
{
|
|
desc: "When the connection is closed, the client enters TRANSIENT FAILURE.",
|
|
want: []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.TransientFailure,
|
|
},
|
|
server: func(lis net.Listener) net.Conn {
|
|
conn, err := lis.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return nil
|
|
}
|
|
|
|
conn.Close()
|
|
return nil
|
|
},
|
|
},
|
|
{
|
|
desc: `When the server sends its connection preface, but the connection dies before the client can write its
|
|
connection preface, the client enters TRANSIENT FAILURE.`,
|
|
want: []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.TransientFailure,
|
|
},
|
|
server: func(lis net.Listener) net.Conn {
|
|
conn, err := lis.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return nil
|
|
}
|
|
|
|
framer := http2.NewFramer(conn, conn)
|
|
if err := framer.WriteSettings(http2.Setting{}); err != nil {
|
|
t.Errorf("Error while writing settings frame. %v", err)
|
|
return nil
|
|
}
|
|
|
|
conn.Close()
|
|
return nil
|
|
},
|
|
},
|
|
{
|
|
desc: `When the server reads the client connection preface but does not send its connection preface, the
|
|
client enters TRANSIENT FAILURE.`,
|
|
want: []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.TransientFailure,
|
|
},
|
|
server: func(lis net.Listener) net.Conn {
|
|
conn, err := lis.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return nil
|
|
}
|
|
|
|
go keepReading(conn)
|
|
|
|
return conn
|
|
},
|
|
},
|
|
} {
|
|
t.Log(test.desc)
|
|
testStateTransitionSingleAddress(t, test.want, test.server)
|
|
}
|
|
}
|
|
|
|
func testStateTransitionSingleAddress(t *testing.T, want []connectivity.State, server func(net.Listener) net.Conn) {
|
|
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
|
|
defer cancel()
|
|
|
|
pl := testutils.NewPipeListener()
|
|
defer pl.Close()
|
|
|
|
// Launch the server.
|
|
var conn net.Conn
|
|
var connMu sync.Mutex
|
|
go func() {
|
|
connMu.Lock()
|
|
conn = server(pl)
|
|
connMu.Unlock()
|
|
}()
|
|
|
|
client, err := DialContext(ctx,
|
|
"",
|
|
WithInsecure(),
|
|
WithBalancerName(stateRecordingBalancerName),
|
|
WithDialer(pl.Dialer()),
|
|
withBackoff(noBackoff{}),
|
|
withMinConnectDeadline(func() time.Duration { return time.Millisecond * 100 }))
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
defer client.Close()
|
|
|
|
stateNotifications := testBalancerBuilder.nextStateNotifier()
|
|
|
|
timeout := time.After(5 * time.Second)
|
|
|
|
for i := 0; i < len(want); i++ {
|
|
select {
|
|
case <-timeout:
|
|
t.Fatalf("timed out waiting for state %d (%v) in flow %v", i, want[i], want)
|
|
case seen := <-stateNotifications:
|
|
if seen != want[i] {
|
|
t.Fatalf("expected to see %v at position %d in flow %v, got %v", want[i], i, want, seen)
|
|
}
|
|
}
|
|
}
|
|
|
|
connMu.Lock()
|
|
defer connMu.Unlock()
|
|
if conn != nil {
|
|
err = conn.Close()
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
}
|
|
}
|
|
|
|
// When a READY connection is closed, the client enters CONNECTING.
|
|
func (s) TestStateTransitions_ReadyToConnecting(t *testing.T) {
|
|
want := []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.Ready,
|
|
connectivity.Connecting,
|
|
}
|
|
|
|
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
|
|
defer cancel()
|
|
|
|
lis, err := net.Listen("tcp", "localhost:0")
|
|
if err != nil {
|
|
t.Fatalf("Error while listening. Err: %v", err)
|
|
}
|
|
defer lis.Close()
|
|
|
|
sawReady := make(chan struct{})
|
|
|
|
// Launch the server.
|
|
go func() {
|
|
conn, err := lis.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return
|
|
}
|
|
|
|
go keepReading(conn)
|
|
|
|
framer := http2.NewFramer(conn, conn)
|
|
if err := framer.WriteSettings(http2.Setting{}); err != nil {
|
|
t.Errorf("Error while writing settings frame. %v", err)
|
|
return
|
|
}
|
|
|
|
// Prevents race between onPrefaceReceipt and onClose.
|
|
<-sawReady
|
|
|
|
conn.Close()
|
|
}()
|
|
|
|
client, err := DialContext(ctx, lis.Addr().String(), WithInsecure(), WithBalancerName(stateRecordingBalancerName))
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
defer client.Close()
|
|
|
|
stateNotifications := testBalancerBuilder.nextStateNotifier()
|
|
|
|
timeout := time.After(5 * time.Second)
|
|
|
|
for i := 0; i < len(want); i++ {
|
|
select {
|
|
case <-timeout:
|
|
t.Fatalf("timed out waiting for state %d (%v) in flow %v", i, want[i], want)
|
|
case seen := <-stateNotifications:
|
|
if seen == connectivity.Ready {
|
|
close(sawReady)
|
|
}
|
|
if seen != want[i] {
|
|
t.Fatalf("expected to see %v at position %d in flow %v, got %v", want[i], i, want, seen)
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
// When the first connection is closed, the client stays in CONNECTING until it
|
|
// tries the second address (which succeeds, and then it enters READY).
|
|
func (s) TestStateTransitions_TriesAllAddrsBeforeTransientFailure(t *testing.T) {
|
|
want := []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.Ready,
|
|
}
|
|
|
|
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
|
|
defer cancel()
|
|
|
|
lis1, err := net.Listen("tcp", "localhost:0")
|
|
if err != nil {
|
|
t.Fatalf("Error while listening. Err: %v", err)
|
|
}
|
|
defer lis1.Close()
|
|
|
|
lis2, err := net.Listen("tcp", "localhost:0")
|
|
if err != nil {
|
|
t.Fatalf("Error while listening. Err: %v", err)
|
|
}
|
|
defer lis2.Close()
|
|
|
|
server1Done := make(chan struct{})
|
|
server2Done := make(chan struct{})
|
|
|
|
// Launch server 1.
|
|
go func() {
|
|
conn, err := lis1.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return
|
|
}
|
|
|
|
conn.Close()
|
|
close(server1Done)
|
|
}()
|
|
// Launch server 2.
|
|
go func() {
|
|
conn, err := lis2.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return
|
|
}
|
|
|
|
go keepReading(conn)
|
|
|
|
framer := http2.NewFramer(conn, conn)
|
|
if err := framer.WriteSettings(http2.Setting{}); err != nil {
|
|
t.Errorf("Error while writing settings frame. %v", err)
|
|
return
|
|
}
|
|
|
|
close(server2Done)
|
|
}()
|
|
|
|
rb := manual.NewBuilderWithScheme("whatever")
|
|
rb.InitialState(resolver.State{Addresses: []resolver.Address{
|
|
{Addr: lis1.Addr().String()},
|
|
{Addr: lis2.Addr().String()},
|
|
}})
|
|
client, err := DialContext(ctx, "this-gets-overwritten", WithInsecure(), WithBalancerName(stateRecordingBalancerName), withResolverBuilder(rb))
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
defer client.Close()
|
|
|
|
stateNotifications := testBalancerBuilder.nextStateNotifier()
|
|
|
|
timeout := time.After(5 * time.Second)
|
|
|
|
for i := 0; i < len(want); i++ {
|
|
select {
|
|
case <-timeout:
|
|
t.Fatalf("timed out waiting for state %d (%v) in flow %v", i, want[i], want)
|
|
case seen := <-stateNotifications:
|
|
if seen != want[i] {
|
|
t.Fatalf("expected to see %v at position %d in flow %v, got %v", want[i], i, want, seen)
|
|
}
|
|
}
|
|
}
|
|
select {
|
|
case <-timeout:
|
|
t.Fatal("saw the correct state transitions, but timed out waiting for client to finish interactions with server 1")
|
|
case <-server1Done:
|
|
}
|
|
select {
|
|
case <-timeout:
|
|
t.Fatal("saw the correct state transitions, but timed out waiting for client to finish interactions with server 2")
|
|
case <-server2Done:
|
|
}
|
|
}
|
|
|
|
// When there are multiple addresses, and we enter READY on one of them, a
|
|
// later closure should cause the client to enter CONNECTING
|
|
func (s) TestStateTransitions_MultipleAddrsEntersReady(t *testing.T) {
|
|
want := []connectivity.State{
|
|
connectivity.Connecting,
|
|
connectivity.Ready,
|
|
connectivity.Connecting,
|
|
}
|
|
|
|
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
|
|
defer cancel()
|
|
|
|
lis1, err := net.Listen("tcp", "localhost:0")
|
|
if err != nil {
|
|
t.Fatalf("Error while listening. Err: %v", err)
|
|
}
|
|
defer lis1.Close()
|
|
|
|
// Never actually gets used; we just want it to be alive so that the resolver has two addresses to target.
|
|
lis2, err := net.Listen("tcp", "localhost:0")
|
|
if err != nil {
|
|
t.Fatalf("Error while listening. Err: %v", err)
|
|
}
|
|
defer lis2.Close()
|
|
|
|
server1Done := make(chan struct{})
|
|
sawReady := make(chan struct{})
|
|
|
|
// Launch server 1.
|
|
go func() {
|
|
conn, err := lis1.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return
|
|
}
|
|
|
|
go keepReading(conn)
|
|
|
|
framer := http2.NewFramer(conn, conn)
|
|
if err := framer.WriteSettings(http2.Setting{}); err != nil {
|
|
t.Errorf("Error while writing settings frame. %v", err)
|
|
return
|
|
}
|
|
|
|
<-sawReady
|
|
|
|
conn.Close()
|
|
|
|
_, err = lis1.Accept()
|
|
if err != nil {
|
|
t.Error(err)
|
|
return
|
|
}
|
|
|
|
close(server1Done)
|
|
}()
|
|
|
|
rb := manual.NewBuilderWithScheme("whatever")
|
|
rb.InitialState(resolver.State{Addresses: []resolver.Address{
|
|
{Addr: lis1.Addr().String()},
|
|
{Addr: lis2.Addr().String()},
|
|
}})
|
|
client, err := DialContext(ctx, "this-gets-overwritten", WithInsecure(), WithBalancerName(stateRecordingBalancerName), withResolverBuilder(rb))
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
defer client.Close()
|
|
|
|
stateNotifications := testBalancerBuilder.nextStateNotifier()
|
|
|
|
timeout := time.After(2 * time.Second)
|
|
|
|
for i := 0; i < len(want); i++ {
|
|
select {
|
|
case <-timeout:
|
|
t.Fatalf("timed out waiting for state %d (%v) in flow %v", i, want[i], want)
|
|
case seen := <-stateNotifications:
|
|
if seen == connectivity.Ready {
|
|
close(sawReady)
|
|
}
|
|
if seen != want[i] {
|
|
t.Fatalf("expected to see %v at position %d in flow %v, got %v", want[i], i, want, seen)
|
|
}
|
|
}
|
|
}
|
|
select {
|
|
case <-timeout:
|
|
t.Fatal("saw the correct state transitions, but timed out waiting for client to finish interactions with server 1")
|
|
case <-server1Done:
|
|
}
|
|
}
|
|
|
|
type stateRecordingBalancer struct {
|
|
notifier chan<- connectivity.State
|
|
balancer.Balancer
|
|
}
|
|
|
|
func (b *stateRecordingBalancer) HandleSubConnStateChange(sc balancer.SubConn, s connectivity.State) {
|
|
b.notifier <- s
|
|
b.Balancer.HandleSubConnStateChange(sc, s)
|
|
}
|
|
|
|
func (b *stateRecordingBalancer) ResetNotifier(r chan<- connectivity.State) {
|
|
b.notifier = r
|
|
}
|
|
|
|
func (b *stateRecordingBalancer) Close() {
|
|
b.Balancer.Close()
|
|
}
|
|
|
|
type stateRecordingBalancerBuilder struct {
|
|
mu sync.Mutex
|
|
notifier chan connectivity.State // The notifier used in the last Balancer.
|
|
}
|
|
|
|
func newStateRecordingBalancerBuilder() *stateRecordingBalancerBuilder {
|
|
return &stateRecordingBalancerBuilder{}
|
|
}
|
|
|
|
func (b *stateRecordingBalancerBuilder) Name() string {
|
|
return stateRecordingBalancerName
|
|
}
|
|
|
|
func (b *stateRecordingBalancerBuilder) Build(cc balancer.ClientConn, opts balancer.BuildOptions) balancer.Balancer {
|
|
stateNotifications := make(chan connectivity.State, 10)
|
|
b.mu.Lock()
|
|
b.notifier = stateNotifications
|
|
b.mu.Unlock()
|
|
return &stateRecordingBalancer{
|
|
notifier: stateNotifications,
|
|
Balancer: balancer.Get(PickFirstBalancerName).Build(cc, opts),
|
|
}
|
|
}
|
|
|
|
func (b *stateRecordingBalancerBuilder) nextStateNotifier() <-chan connectivity.State {
|
|
b.mu.Lock()
|
|
defer b.mu.Unlock()
|
|
ret := b.notifier
|
|
b.notifier = nil
|
|
return ret
|
|
}
|
|
|
|
type noBackoff struct{}
|
|
|
|
func (b noBackoff) Backoff(int) time.Duration { return time.Duration(0) }
|
|
|
|
// Keep reading until something causes the connection to die (EOF, server
|
|
// closed, etc). Useful as a tool for mindlessly keeping the connection
|
|
// healthy, since the client will error if things like client prefaces are not
|
|
// accepted in a timely fashion.
|
|
func keepReading(conn net.Conn) {
|
|
buf := make([]byte, 1024)
|
|
for _, err := conn.Read(buf); err == nil; _, err = conn.Read(buf) {
|
|
}
|
|
}
|