kotovalexarian-likes-github/moby--moby
1
0
Fork 0
mirror of https://github.com/moby/moby.git synced 2022-11-09 12:21:53 -05:00
Code Releases Activity

Merge pull request #43147 from PettitWesley/backport-fluentd-fix

Browse source 
[20.10 backport] backport fluentd log driver async connect fix
This commit is contained in:
Sebastiaan van Stijn 2022-01-20 09:23:15 +01:00 committed by GitHub
commit 22ff2ed34b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 411 additions and 88 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
daemon/logger/fluentd/fluentd.go
View file

@ -242,6 +242,7 @@ func parseConfig(cfg map[string]string) (fluent.Config, error) {
AsyncConnect: asyncConnect,
SubSecondPrecision: subSecondPrecision,
RequestAck: requestAck,
ForceStopAsyncSend: async || asyncConnect,
}
return config, nil

2
vendor.conf
View file

@ -106,7 +106,7 @@ github.com/godbus/dbus/v5 37bf87eef99d69c4f1d3528bd66e
github.com/Graylog2/go-gelf 1550ee647df0510058c9d67a45c56f18911d80b8 # v2 branch
# fluent-logger-golang deps
github.com/fluent/fluent-logger-golang 7a6c9dcd7f14c2ed5d8c55c11b894e5455ee311b # v1.4.0
github.com/fluent/fluent-logger-golang 0b652e850a9140d0b1db6390d8925d0601e952db # v1.8.0
github.com/philhofer/fwd bb6d471dc95d4fe11e432687f8b70ff496cf3136 # v1.0.0
github.com/tinylib/msgp af6442a0fcf6e2a1b824f70dd0c734f01e817751 # v1.1.0

112
vendor/github.com/fluent/fluent-logger-golang/README.md generated vendored
View file

@ -1,7 +1,8 @@
fluent-logger-golang
====
[![Build Status](https://travis-ci.org/fluent/fluent-logger-golang.png?branch=master)](https://travis-ci.org/fluent/fluent-logger-golang)
[![Build Status](https://github.com/fluent/fluent-logger-golang/actions/workflows/ci.yaml/badge.svg?branch=master)](https://github.com/fluent/fluent-logger-golang/actions)
[![GoDoc](https://godoc.org/github.com/fluent/fluent-logger-golang/fluent?status.svg)](https://godoc.org/github.com/fluent/fluent-logger-golang/fluent)
## A structured event logger for Fluentd (Golang)
@ -19,8 +20,6 @@ Install the package with `go get` and use `import` to include it in your project
import "github.com/fluent/fluent-logger-golang/fluent"
```
GoDoc: http://godoc.org/github.com/fluent/fluent-logger-golang/fluent
## Example
```go
@ -29,7 +28,7 @@ package main
import (
"github.com/fluent/fluent-logger-golang/fluent"
"fmt"
"time"
//"time"
)
func main() {
@ -59,22 +58,125 @@ func main() {
f := fluent.New(fluent.Config{FluentPort: 80, FluentHost: "example.com"})
```
### FluentNetwork
Specify the network protocol. The supported values are:
* "tcp" (use `FluentHost` and `FluentPort`)
* "tls" (use`FluentHost` and `FluentPort`)
* "unix" (use `FluentSocketPath`)
The default is "tcp".
### FluentHost
Specify a hostname or IP address as a string for the destination of the "tcp" protocol.
The default is "127.0.0.1".
### FluentPort
Specify the TCP port of the destination. The default is 24224.
### FluentSocketPath
Specify the unix socket path when `FluentNetwork` is "unix".
### Timeout
Set the timeout value of `time.Duration` to connect to the destination.
The default is 3 seconds.
### WriteTimeout
Sets the timeout for Write call of logger.Post.
Sets the timeout value of `time.Duration` for Write call of `logger.Post`.
Since the default is zero value, Write will not time out.
### BufferLimit
Sets the number of events buffered on the memory. Records will be stored in memory up to this number. If the buffer is full, the call to record logs will fail.
The default is 8192.
### RetryWait
Set the duration of the initial wait for the first retry, in milliseconds. The actual retry wait will be `r * 1.5^(N-1)` (r: this value, N: the number of retries).
The default is 500.
### MaxRetry
Sets the maximum number of retries. If the number of retries become larger than this value, the write/send operation will fail.
The default is 13.
### MaxRetryWait
The maximum duration of wait between retries, in milliseconds. If the calculated retry wait is larger than this value, the actual retry wait will be this value.
The default is 60,000 (60 seconds).
### TagPrefix
Sets the prefix string of the tag. Prefix will be appended with a dot `.`, like `ppp.tag` (ppp: the value of this parameter, tag: the tag string specified in a call).
The default is blank.
### Async
Enable asynchronous I/O (connect and write) for sending events to Fluentd.
The default is false.
### ForceStopAsyncSend
When Async is enabled, immediately discard the event queue on close() and return (instead of trying MaxRetry times for each event in the queue before returning)
The default is false.
### AsyncResultCallback
When Async is enabled, if this is callback is provided, it will be called on every write to Fluentd. The callback function
takes two arguments - a `[]byte` of the message that was to be sent and an `error`. If the `error` is not nil this means the
delivery of the message was unsuccessful.
### SubSecondPrecision
Enable time encoding as EventTime, which contains sub-second precision values. The messages encoded with this option can be received only by Fluentd v0.14 or later.
The default is false.
### MarshalAsJson
Enable Json data marshaling to send messages using Json format (instead of the standard MessagePack). It is supported by Fluentd `in_forward` plugin.
The default is false.
### RequestAck
Sets whether to request acknowledgment from Fluentd to increase the reliability
of the connection. The default is false.
### TlsInsecureSkipVerify
Skip verifying the server certificate. Useful for development and testing. The default is false.
## FAQ
### Does this logger support the features of Fluentd Forward Protocol v1?
"the features" includes heartbeat messages (for TCP keepalive), TLS transport and shared key authentication.
This logger doesn't support those features. Patches are welcome!
### Is it allowed to call `Fluent.Post()` after connection close?
Before v1.8.0, the Fluent logger silently reopened connections whenever
`Fluent.Post()` was called.
```go
logger, _ := fluent.New(fluent.Config{})
logger.Post(tag, data)
logger.Close()
logger.Post(tag, data) /* reopen connection */
```
However, this behavior was confusing, in particular when multiple goroutines
were involved. Starting v1.8.0, the logger no longer accepts `Fluent.Post()`
after `Fluent.Close()`, and instead returns a "Logger already closed" error.
## Tests
```
go test
```

368
vendor/github.com/fluent/fluent-logger-golang/fluent/fluent.go generated vendored
View file

@ -1,10 +1,13 @@
package fluent
import (
"context"
"crypto/tls"
"encoding/json"
"errors"
"fmt"
"math"
"math/rand"
"net"
"os"
"reflect"
@ -15,8 +18,8 @@ import (
"bytes"
"encoding/base64"
"encoding/binary"
"github.com/tinylib/msgp/msgp"
"math/rand"
)
const (
@ -34,21 +37,30 @@ const (
// Default sub-second precision value to false since it is only compatible
// with fluentd versions v0.14 and above.
defaultSubSecondPrecision = false
// Default value whether to skip checking insecure certs on TLS connections.
defaultTlsInsecureSkipVerify = false
)
// randomGenerator is used by getUniqueId to generate ack hashes. Its value is replaced
// during tests with a deterministic function.
var randomGenerator = rand.Uint64
type Config struct {
FluentPort int `json:"fluent_port"`
FluentHost string `json:"fluent_host"`
FluentNetwork string `json:"fluent_network"`
FluentSocketPath string `json:"fluent_socket_path"`
Timeout time.Duration `json:"timeout"`
WriteTimeout time.Duration `json:"write_timeout"`
BufferLimit int `json:"buffer_limit"`
RetryWait int `json:"retry_wait"`
MaxRetry int `json:"max_retry"`
MaxRetryWait int `json:"max_retry_wait"`
TagPrefix string `json:"tag_prefix"`
Async bool `json:"async"`
FluentPort int `json:"fluent_port"`
FluentHost string `json:"fluent_host"`
FluentNetwork string `json:"fluent_network"`
FluentSocketPath string `json:"fluent_socket_path"`
Timeout time.Duration `json:"timeout"`
WriteTimeout time.Duration `json:"write_timeout"`
BufferLimit int `json:"buffer_limit"`
RetryWait int `json:"retry_wait"`
MaxRetry int `json:"max_retry"`
MaxRetryWait int `json:"max_retry_wait"`
TagPrefix string `json:"tag_prefix"`
Async bool `json:"async"`
ForceStopAsyncSend bool `json:"force_stop_async_send"`
AsyncResultCallback func(data []byte, err error)
// Deprecated: Use Async instead
AsyncConnect bool `json:"async_connect"`
MarshalAsJSON bool `json:"marshal_as_json"`
@ -61,6 +73,21 @@ type Config struct {
// respond with an acknowledgement. This option improves the reliability
// of the message transmission.
RequestAck bool `json:"request_ack"`
// Flag to skip verifying insecure certs on TLS connections
TlsInsecureSkipVerify bool `json: "tls_insecure_skip_verify"`
}
type ErrUnknownNetwork struct {
network string
}
func (e *ErrUnknownNetwork) Error() string {
return "unknown network " + e.network
}
func NewErrUnknownNetwork(network string) error {
return &ErrUnknownNetwork{network}
}
type msgToSend struct {
@ -71,15 +98,35 @@ type msgToSend struct {
type Fluent struct {
Config
pending chan *msgToSend
wg sync.WaitGroup
dialer dialer
// stopRunning is used in async mode to signal to run() it should abort.
stopRunning chan struct{}
// cancelDialings is used by Close() to stop any in-progress dialing.
cancelDialings context.CancelFunc
pending chan *msgToSend
pendingMutex sync.RWMutex
closed bool
wg sync.WaitGroup
muconn sync.Mutex
muconn sync.RWMutex
conn net.Conn
}
type dialer interface {
DialContext(ctx context.Context, network, address string) (net.Conn, error)
}
// New creates a new Logger.
func New(config Config) (f *Fluent, err error) {
func New(config Config) (*Fluent, error) {
if config.Timeout == 0 {
config.Timeout = defaultTimeout
}
return newWithDialer(config, &net.Dialer{
Timeout: config.Timeout,
})
}
func newWithDialer(config Config, d dialer) (f *Fluent, err error) {
if config.FluentNetwork == "" {
config.FluentNetwork = defaultNetwork
}
@ -92,9 +139,6 @@ func New(config Config) (f *Fluent, err error) {
if config.FluentSocketPath == "" {
config.FluentSocketPath = defaultSocketPath
}
if config.Timeout == 0 {
config.Timeout = defaultTimeout
}
if config.WriteTimeout == 0 {
config.WriteTimeout = defaultWriteTimeout
}
@ -110,20 +154,36 @@ func New(config Config) (f *Fluent, err error) {
if config.MaxRetryWait == 0 {
config.MaxRetryWait = defaultMaxRetryWait
}
if !config.TlsInsecureSkipVerify {
config.TlsInsecureSkipVerify = defaultTlsInsecureSkipVerify
}
if config.AsyncConnect {
fmt.Fprintf(os.Stderr, "fluent#New: AsyncConnect is now deprecated, please use Async instead")
config.Async = config.Async || config.AsyncConnect
}
if config.Async {
ctx, cancel := context.WithCancel(context.Background())
f = &Fluent{
Config: config,
pending: make(chan *msgToSend, config.BufferLimit),
Config: config,
dialer: d,
stopRunning: make(chan struct{}),
cancelDialings: cancel,
pending: make(chan *msgToSend, config.BufferLimit),
pendingMutex: sync.RWMutex{},
muconn: sync.RWMutex{},
}
f.wg.Add(1)
go f.run()
go f.run(ctx)
} else {
f = &Fluent{Config: config}
err = f.connect()
f = &Fluent{
Config: config,
dialer: d,
muconn: sync.RWMutex{},
}
err = f.connect(context.Background())
}
return
}
@ -176,13 +236,18 @@ func (f *Fluent) PostWithTime(tag string, tm time.Time, message interface{}) err
fields := msgtype.NumField()
for i := 0; i < fields; i++ {
field := msgtype.Field(i)
value := msg.FieldByIndex(field.Index)
// ignore unexported fields
if !value.CanInterface() {
continue
}
name := field.Name
if n1 := field.Tag.Get("msg"); n1 != "" {
name = n1
} else if n2 := field.Tag.Get("codec"); n2 != "" {
name = n2
}
kv[name] = msg.FieldByIndex(field.Index).Interface()
kv[name] = value.Interface()
}
return f.EncodeAndPostData(tag, tm, kv)
}
@ -219,8 +284,12 @@ func (f *Fluent) postRawData(msg *msgToSend) error {
if f.Config.Async {
return f.appendBuffer(msg)
}
// Synchronous write
return f.write(msg)
if f.closed {
return fmt.Errorf("fluent#postRawData: Logger already closed")
}
return f.writeWithRetry(context.Background(), msg)
}
// For sending forward protocol adopted JSON
@ -254,7 +323,7 @@ func getUniqueID(timeUnix int64) (string, error) {
enc.Close()
return "", err
}
if err := binary.Write(enc, binary.LittleEndian, rand.Uint64()); err != nil {
if err := binary.Write(enc, binary.LittleEndian, randomGenerator()); err != nil {
enc.Close()
return "", err
}
@ -290,18 +359,55 @@ func (f *Fluent) EncodeData(tag string, tm time.Time, message interface{}) (msg
return
}
// Close closes the connection, waiting for pending logs to be sent
// Close closes the connection, waiting for pending logs to be sent. If the client is
// running in async mode, the run() goroutine exits before Close() returns.
func (f *Fluent) Close() (err error) {
if f.Config.Async {
f.pendingMutex.Lock()
if f.closed {
f.pendingMutex.Unlock()
return nil
}
f.closed = true
f.pendingMutex.Unlock()
if f.Config.ForceStopAsyncSend {
close(f.stopRunning)
f.cancelDialings()
}
close(f.pending)
// If ForceStopAsyncSend is false, all logs in the channel have to be sent
// before closing the connection. At this point closed is true so no more
// logs are written to the channel and f.pending has been closed, so run()
// goroutine will exit as soon as all logs in the channel are sent.
if !f.Config.ForceStopAsyncSend {
f.wg.Wait()
}
}
f.muconn.Lock()
f.close()
f.closed = true
f.muconn.Unlock()
// If ForceStopAsyncSend is true, we shall close the connection before waiting for
// run() goroutine to exit to be sure we aren't waiting on ack message that might
// never come (eg. because fluentd server is down). However we want to be sure the
// run() goroutine stops before returning from Close().
if f.Config.ForceStopAsyncSend {
f.wg.Wait()
}
f.close()
return
}
// appendBuffer appends data to buffer with lock.
func (f *Fluent) appendBuffer(msg *msgToSend) error {
f.pendingMutex.RLock()
defer f.pendingMutex.RUnlock()
if f.closed {
return fmt.Errorf("fluent#appendBuffer: Logger already closed")
}
select {
case f.pending <- msg:
default:
@ -310,42 +416,114 @@ func (f *Fluent) appendBuffer(msg *msgToSend) error {
return nil
}
// close closes the connection.
// close closes the connection. Callers should take care of locking muconn first.
func (f *Fluent) close() {
f.muconn.Lock()
if f.conn != nil {
f.conn.Close()
f.conn = nil
}
f.muconn.Unlock()
}
// connect establishes a new connection using the specified transport.
func (f *Fluent) connect() (err error) {
// connect establishes a new connection using the specified transport. Caller should
// take care of locking muconn first.
func (f *Fluent) connect(ctx context.Context) (err error) {
switch f.Config.FluentNetwork {
case "tcp":
f.conn, err = net.DialTimeout(f.Config.FluentNetwork, f.Config.FluentHost+":"+strconv.Itoa(f.Config.FluentPort), f.Config.Timeout)
f.conn, err = f.dialer.DialContext(ctx,
f.Config.FluentNetwork,
f.Config.FluentHost+":"+strconv.Itoa(f.Config.FluentPort))
case "tls":
tlsConfig := &tls.Config{InsecureSkipVerify: f.Config.TlsInsecureSkipVerify}
f.conn, err = tls.DialWithDialer(
&net.Dialer{Timeout: f.Config.Timeout},
"tcp",
f.Config.FluentHost+":"+strconv.Itoa(f.Config.FluentPort), tlsConfig,
)
case "unix":
f.conn, err = net.DialTimeout(f.Config.FluentNetwork, f.Config.FluentSocketPath, f.Config.Timeout)
f.conn, err = f.dialer.DialContext(ctx,
f.Config.FluentNetwork,
f.Config.FluentSocketPath)
default:
err = net.UnknownNetworkError(f.Config.FluentNetwork)
err = NewErrUnknownNetwork(f.Config.FluentNetwork)
}
return err
}
func (f *Fluent) run() {
var errIsClosing = errors.New("fluent logger is closing")
// Caller should take care of locking muconn first.
func (f *Fluent) connectWithRetry(ctx context.Context) error {
// A Time channel is used instead of time.Sleep() to avoid blocking this
// goroutine during way too much time (because of the exponential back-off
// retry).
// time.NewTimer() is used instead of time.After() to avoid leaking the
// timer channel (cf. https://pkg.go.dev/time#After).
timeout := time.NewTimer(time.Duration(0))
defer func() {
// timeout.Stop() is called in a function literal instead of being
// defered directly as it's re-assigned below when the retry loop spins.
timeout.Stop()
}()
for i := 0; i < f.Config.MaxRetry; i++ {
select {
case <-timeout.C:
err := f.connect(ctx)
if err == nil {
return nil
}
if _, ok := err.(*ErrUnknownNetwork); ok {
return err
}
if err == context.Canceled {
return errIsClosing
}
waitTime := f.Config.RetryWait * e(defaultReconnectWaitIncreRate, float64(i-1))
if waitTime > f.Config.MaxRetryWait {
waitTime = f.Config.MaxRetryWait
}
timeout = time.NewTimer(time.Duration(waitTime) * time.Millisecond)
case <-ctx.Done():
return errIsClosing
}
}
return fmt.Errorf("could not connect to fluentd after %d retries", f.Config.MaxRetry)
}
// run is the goroutine used to unqueue and write logs in async mode. That
// goroutine is meant to run during the whole life of the Fluent logger.
func (f *Fluent) run(ctx context.Context) {
for {
select {
case entry, ok := <-f.pending:
// f.stopRunning is closed before f.pending only when ForceStopAsyncSend
// is enabled. Otherwise, f.pending is closed when Close() is called.
if !ok {
f.wg.Done()
return
}
err := f.write(entry)
if err != nil {
err := f.writeWithRetry(ctx, entry)
if err != nil && err != errIsClosing {
fmt.Fprintf(os.Stderr, "[%s] Unable to send logs to fluentd, reconnecting...\n", time.Now().Format(time.RFC3339))
}
if f.AsyncResultCallback != nil {
var data []byte
if entry != nil {
data = entry.data
}
f.AsyncResultCallback(data, err)
}
case <-f.stopRunning:
fmt.Fprintf(os.Stderr, "[%s] Discarding queued events...\n", time.Now().Format(time.RFC3339))
f.wg.Done()
return
}
}
}
@ -354,55 +532,85 @@ func e(x, y float64) int {
return int(math.Pow(x, y))
}
func (f *Fluent) write(msg *msgToSend) error {
func (f *Fluent) writeWithRetry(ctx context.Context, msg *msgToSend) error {
for i := 0; i < f.Config.MaxRetry; i++ {
// Connect if needed
f.muconn.Lock()
if f.conn == nil {
err := f.connect()
if err != nil {
f.muconn.Unlock()
waitTime := f.Config.RetryWait * e(defaultReconnectWaitIncreRate, float64(i-1))
if waitTime > f.Config.MaxRetryWait {
waitTime = f.Config.MaxRetryWait
}
time.Sleep(time.Duration(waitTime) * time.Millisecond)
continue
}
if retry, err := f.write(ctx, msg); !retry {
return err
}
}
return fmt.Errorf("fluent#write: failed to write after %d attempts", f.Config.MaxRetry)
}
// write writes the provided msg to fluentd server. Its first return values is
// a bool indicating whether the write should be retried.
// This method relies on function literals to execute muconn.Unlock or
// muconn.RUnlock in deferred calls to ensure the mutex is unlocked even in
// the case of panic recovering.
func (f *Fluent) write(ctx context.Context, msg *msgToSend) (bool, error) {
closer := func() {
f.muconn.Lock()
defer f.muconn.Unlock()
f.close()
}
if err := func() (err error) {
f.muconn.Lock()
defer f.muconn.Unlock()
if f.conn == nil {
err = f.connectWithRetry(ctx)
}
return err
}(); err != nil {
// Here, we don't want to retry the write since connectWithRetry already
// retries Config.MaxRetry times to connect.
return false, fmt.Errorf("fluent#write: %v", err)
}
if err := func() (err error) {
f.muconn.RLock()
defer f.muconn.RUnlock()
if f.conn == nil {
return fmt.Errorf("connection has been closed before writing to it")
}
f.muconn.Unlock()
// We're connected, write msg
t := f.Config.WriteTimeout
if time.Duration(0) < t {
f.conn.SetWriteDeadline(time.Now().Add(t))
} else {
f.conn.SetWriteDeadline(time.Time{})
}
_, err := f.conn.Write(msg.data)
if err != nil {
f.close()
_, err = f.conn.Write(msg.data)
return err
}(); err != nil {
closer()
return true, fmt.Errorf("fluent#write: %v", err)
}
// Acknowledgment check
if msg.ack != "" {
resp := &AckResp{}
var err error
if f.Config.MarshalAsJSON {
dec := json.NewDecoder(f.conn)
err = dec.Decode(resp)
} else {
// Acknowledgment check
if msg.ack != "" {
resp := &AckResp{}
if f.Config.MarshalAsJSON {
dec := json.NewDecoder(f.conn)
err = dec.Decode(resp)
} else {
r := msgp.NewReader(f.conn)
err = resp.DecodeMsg(r)
}
if err != nil || resp.Ack != msg.ack {
f.close()
continue
}
}
return err
r := msgp.NewReader(f.conn)
err = resp.DecodeMsg(r)
}
if err != nil || resp.Ack != msg.ack {
fmt.Fprintf(os.Stderr, "fluent#write: message ack (%s) doesn't match expected one (%s). Closing connection...", resp.Ack, msg.ack)
closer()
return true, err
}
}
return fmt.Errorf("fluent#write: failed to reconnect, max retry: %v", f.Config.MaxRetry)
return false, nil
}

16
vendor/github.com/fluent/fluent-logger-golang/fluent/proto.go generated vendored
View file

@ -3,6 +3,7 @@
package fluent
import (
"fmt"
"time"
"github.com/tinylib/msgp/msgp"
@ -93,8 +94,19 @@ func (t *EventTime) MarshalBinaryTo(b []byte) error {
return nil
}
// UnmarshalBinary is not implemented since decoding messages is not supported
// by this library.
// Although decoding messages is not officially supported by this library,
// UnmarshalBinary is implemented for testing and general completeness.
func (t *EventTime) UnmarshalBinary(b []byte) error {
if len(b) != length {
return fmt.Errorf("Invalid EventTime byte length: %d", len(b))
}
sec := (int32(b[0]) << 24) | (int32(b[1]) << 16)
sec = sec | (int32(b[2]) << 8) | int32(b[3])
nsec := (int32(b[4]) << 24) | (int32(b[5]) << 16)
nsec = nsec | (int32(b[6]) << 8) | int32(b[7])
*t = EventTime(time.Unix(int64(sec), int64(nsec)))
return nil
}