oversser在go中创建可监控、重启、自升级二进制文件的包。分为master模式和slave模式;master为父进程,创建监听套接字、设置isSlave标记等一些初始化工作,之后调用forkExec创建slave子进程来执行bin类型的二进制可执行文件;热升级时发送SIGUSR2信号给master进程,master会转给slave使其停止,之后master再创建slave子进程来执行升级后的bin文件;虽然子进程重新fork,但是此时父进程还是可以接受新的连接请求,保证业务不中断。
不论是Run还是RunErr,底层调用都是runErr,热重启流程分析如下,见注释
func runErr(c *Config) error { //os not supported if !supported { return fmt.Errorf("os (%s) not supported", runtime.GOOS) } //配置检查 if err := validate(c); err != nil { return err } //环境变量合法性检查 if sanityCheck() { return nil } //run either in master or slave mode if os.Getenv(envIsSlave) == "1" { currentProcess = &slave{Config: c} } else { currentProcess = &master{Config: c}//第一次启动会执行此分支 } //master/slave 分别实现了currentProcess此接口 return currentProcess.run() }
master的run
func (mp *master) run() error { mp.debugf("run") if err := mp.checkBinary(); err != nil { return err } if mp.Config.Fetcher != nil { if err := mp.Config.Fetcher.Init(); err != nil { mp.warnf("fetcher init failed (%s). fetcher disabled.", err) mp.Config.Fetcher = nil } } //处理信号,开启协程,对应的信号调用其对应的处理函数 mp.setupSignalling() //处理监听套接字,监听请求到来 if err := mp.retreiveFileDescriptors(); err != nil { return err } if mp.Config.Fetcher != nil { mp.printCheckUpdate = true mp.fetch() go mp.fetchLoop() } return mp.forkLoop() } //not a real fork func (mp *master) forkLoop() error { //loop, restart command for { if err := mp.fork(); err != nil { return err } } } func (mp *master) fork() error { mp.debugf("starting %s", mp.binPath) cmd := exec.Command(mp.binPath) //mark this new process as the "active" slave process. //this process is assumed to be holding the socket files. mp.slaveCmd = cmd mp.slaveID++ //provide the slave process with some state e := os.Environ() e = append(e, envBinID+"="+hex.EncodeToString(mp.binHash)) e = append(e, envBinPath+"="+mp.binPath) e = append(e, envSlaveID+"="+strconv.Itoa(mp.slaveID)) e = append(e, envIsSlave+"=1")//标记为子进程 e = append(e, envNumFDs+"="+strconv.Itoa(len(mp.slaveExtraFiles))) cmd.Env = e //inherit master args/stdfiles cmd.Args = os.Args cmd.Stdin = os.Stdin cmd.Stdout = os.Stdout cmd.Stderr = os.Stderr //include socket files cmd.ExtraFiles = mp.slaveExtraFiles if err := cmd.Start(); err != nil {//fork子进程,执行bin文件 return fmt.Errorf("Failed to start slave process: %s", err) } //was scheduled to restart, notify success if mp.restarting { mp.restartedAt = time.Now() mp.restarting = false mp.restarted <- true } //convert wait into channel cmdwait := make(chan error) go func() { cmdwait <- cmd.Wait() }() //wait.... select { case err := <-cmdwait: //program exited before releasing descriptors //proxy exit code out to master code := 0 if err != nil { code = 1 if exiterr, ok := err.(*exec.ExitError); ok { if status, ok := exiterr.Sys().(syscall.WaitStatus); ok { code = status.ExitStatus() } } } mp.debugf("prog exited with %d", code) //if a restarts are disabled or if it was an //unexpected crash, proxy this exit straight //through to the main process if mp.NoRestart || !mp.restarting { os.Exit(code) } case <-mp.descriptorsReleased://子进程收到SIGUSR2腿出后,会给master发信号,master收到信号会置此变量,然后返回nil,重新fork //if descriptors are released, the program //has yielded control of its sockets and //a parallel instance of the program can be //started safely. it should serve state.Listeners //to ensure downtime is kept at <1sec. The previous //cmd.Wait() will still be consumed though the //result will be discarded. } return nil }
收到重启信号的处理
func (mp *master) setupSignalling() { //updater-forker comms mp.restarted = make(chan bool) mp.descriptorsReleased = make(chan bool) //read all master process signals signals := make(chan os.Signal) signal.Notify(signals)//master接受所有信号 go func() { for s := range signals { mp.handleSignal(s)//执行对应信号处理程序 } }() } func (mp *master) handleSignal(s os.Signal) { if s == mp.RestartSignal { //user initiated manual restart go mp.triggerRestart()//此信号为自定义信号,在调用处可指定,一般用SIGUSR2 } else if s.String() == "child exited" { // will occur on every restart, ignore it } else //**during a restart** a SIGUSR1 signals //to the master process that, the file //descriptors have been released if mp.awaitingUSR1 && s == SIGUSR1 {//slave进程退出时发送SIGUSR1给master,master收到后设置descriptorsReleased标记后,master的fork中阻塞的select返回nil,重新fork mp.debugf("signaled, sockets ready") mp.awaitingUSR1 = false mp.descriptorsReleased <- true } else //while the slave process is running, proxy //all signals through if mp.slaveCmd != nil && mp.slaveCmd.Process != nil { mp.debugf("proxy signal (%s)", s) mp.sendSignal(s) } else //otherwise if not running, kill on CTRL+c if s == os.Interrupt { mp.debugf("interupt with no slave") os.Exit(1) } else { mp.debugf("signal discarded (%s), no slave process", s) } } func (mp *master) triggerRestart() { if mp.restarting { mp.debugf("already graceful restarting") return //skip } else if mp.slaveCmd == nil || mp.restarting { mp.debugf("no slave process") return //skip } mp.debugf("graceful restart triggered") mp.restarting = true mp.awaitingUSR1 = true mp.signalledAt = time.Now() mp.sendSignal(mp.Config.RestartSignal) //ask nicely to terminate//向slave发送停止信号 select { case <-mp.restarted: //success mp.debugf("restart success") case <-time.After(mp.TerminateTimeout): //times up mr. process, we did ask nicely! mp.debugf("graceful timeout, forcing exit") mp.sendSignal(os.Kill) } }
总的来说:master fork出slave执行bin文件后,在另外开启的协程中循环处理信号
- 如果是
RestartSignal
信号,那么就开启一个协程调用triggerRestart
方法,即执行重启逻辑。triggerRestart
方法向slave
发送一个RestartSignal
信号 - 如果是其它信号,并且
overseer
的slave
进程已在运行,那么就发送这个信号给slave
进程 - 如果
slave
进程没有开启,并且是Ctrl+c
操作,那么就退出overseer
- 如果
slave
进程没有打开,并且是非Crtl+c
操作发出的信号,那么就打印日志
func (sp *slave) run() error { sp.id = os.Getenv(envSlaveID) sp.debugf("run") sp.state.Enabled = true sp.state.ID = os.Getenv(envBinID) sp.state.StartedAt = time.Now() sp.state.Address = sp.Config.Address sp.state.Addresses = sp.Config.Addresses sp.state.GracefulShutdown = make(chan bool, 1) sp.state.BinPath = os.Getenv(envBinPath) //由master pid找到master信息,以便在slava终止后向master发送SIGUSR1 if err := sp.watchParent(); err != nil { return err } //初始化文件描述符 if err := sp.initFileDescriptors(); err != nil { return err } //处理信号 sp.watchSignal() //run program with state sp.debugf("start program") sp.Config.Program(sp.state)//执行prog return nil } func (sp *slave) watchSignal() { signals := make(chan os.Signal) signal.Notify(signals, sp.Config.RestartSignal) go func() { <-signals signal.Stop(signals) sp.debugf("graceful shutdown requested") //master wants to restart, close(sp.state.GracefulShutdown) //release any sockets and notify master if len(sp.listeners) > 0 { //perform graceful shutdown for _, l := range sp.listeners { l.release(sp.Config.TerminateTimeout) } //signal release of held sockets, allows master to start //a new process before this child has actually exited. //early restarts not supported with restarts disabled. if !sp.NoRestart { sp.masterProc.Signal(SIGUSR1)//释放完资源后向master发送SIGUSR1 } //listeners should be waiting on connections to close... } //start death-timer go func() { time.Sleep(sp.Config.TerminateTimeout) sp.debugf("timeout. forceful shutdown") os.Exit(1) }() }() }
总的来说:slave收到终止信号释放完资源后要向master发送SIGUSR1,master收到SIGUSR1后,设置mp.descriptorsReleased <- true,因为master的fork阻塞在select descriptorsReleased数据写入,master写入其数据后,fork中select返回nil,继续执行循环fork。
实践
//create another main() to run the overseer process //and then convert your old main() into a ‘prog(state)‘ overseer.Run(overseer.Config{ Program: prog,//slave进程运行prog Addresses: []string{":5005", ":5006"}, //Fetcher: &fetcher.File{Path: "my_app_next"}, Debug: true,
RestartSignal: syscall.SIGUSR2,
})// splay log of overseer actions //prog(state) runs in a child process //convert your ‘main()‘ into a ‘prog(state)‘ //‘prog()‘ is run in a child process func prog(state overseer.State) { fmt.Printf("app#%s (%s) listening...\n", BuildID, state.ID) http.Handle("/", http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) { duration, err := time.ParseDuration(r.FormValue("duration")) if err != nil { http.Error(w, err.Error(), 400) return } time.Sleep(duration) w.Write([]byte("Hello World")) fmt.Fprintf(w, "app#%s (%s) says hello\n", BuildID, state.ID) })) http.Serve(state.Listener, nil) fmt.Printf("app#%s (%s) exiting...\n", BuildID, state.ID) }
日志如下,可对照上述流程梳理下逻辑
/main &
[1] 63416
2021/08/10 14:34:59 [overseer master] run
2021/08/10 14:34:59 [overseer master] starting /Users/user/test_go/main
2021/08/10 14:34:59 [overseer slave#1] run
2021/08/10 14:34:59 [overseer slave#1] start program
app#1 (b12eb45d8f6dfa86a8529a6aa25f59569d74f16b) listening...
//获取旧的api数据
test_go % curl "http://127.0.0.1:5005/sleep?duration=60s" &
[2] 63423
//替换bin文件,热升级
test_go % rm -rf main
test_go % go build -ldflags ‘-X main.BuildID=2‘ main.go
test_go % kill -USR2 63416
2021/08/10 14:36:21 [overseer master] graceful restart triggered
2021/08/10 14:36:21 [overseer slave#1] graceful shutdown requested
2021/08/10 14:36:21 [overseer master] signaled, sockets ready
2021/08/10 14:36:21 [overseer master] starting /Users/user/test_go/main
2021/08/10 14:36:21 [overseer master] restart success
2021/08/10 14:36:21 [overseer master] signal discarded (urgent I/O condition), no slave process
//旧的api数据输出
Hello Worldapp#1 (b12eb45d8f6dfa86a8529a6aa25f59569d74f16b) says array
app#1 (b12eb45d8f6dfa86a8529a6aa25f59569d74f16b) exiting...
[2] + done curl "http://127.0.0.1:5005/sleep?duration=60s"
2021/08/10 14:36:22 [overseer master] proxy signal (urgent I/O condition)
2021/08/10 14:36:22 [overseer master] proxy signal (urgent I/O condition)
2021/08/10 14:36:22 [overseer master] proxy signal (urgent I/O condition)
2021/08/10 14:36:22 [overseer master] proxy signal (urgent I/O condition)
2021/08/10 14:36:23 [overseer slave#2] run
2021/08/10 14:36:23 [overseer slave#2] start program
app#2 (b12eb45d8f6dfa86a8529a6aa25f59569d74f16b) listening...
//获取新的api数据
test_go % curl "http://127.0.0.1:5005/sleep?duration=1s"
Hello Worldapp#2 (b12eb45d8f6dfa86a8529a6aa25f59569d74f16b) says hello
oversser github https://github.com/jpillora/overseer