摘要：守護(hù)進(jìn)程是在一類脫離終端在后臺(tái)執(zhí)行的程序通常以結(jié)尾隨系統(tǒng)啟動(dòng)其父進(jìn)程通常是進(jìn)程一般要讓當(dāng)前程序以守護(hù)進(jìn)程形式運(yùn)行在命令后加并重定向輸出即可或者使用也可以這是直接運(yùn)行程序的方式如果是用具體語言代碼的形式來實(shí)現(xiàn)呢首先看一下守護(hù)進(jìn)程的實(shí)現(xiàn)方式創(chuàng)建

守護(hù)進(jìn)程是在一類脫離終端在后臺(tái)執(zhí)行的程序, 通常以d結(jié)尾, 隨系統(tǒng)啟動(dòng), 其父進(jìn)程(ppid)通常是init進(jìn)程

一般要讓當(dāng)前程序以守護(hù)進(jìn)程形式運(yùn)行, 在命令后加&并重定向輸出即可

$ python someprogram.py > /dev/null 2>&1 &

或者使用nohup也可以
這是直接運(yùn)行程序的方式, 如果是用具體語言代碼的形式來實(shí)現(xiàn)呢, 首先看一下守護(hù)進(jìn)程的實(shí)現(xiàn)方式

創(chuàng)建子進(jìn)程, 父進(jìn)程退出
父進(jìn)程先于子進(jìn)程退出會(huì)造成子進(jìn)程成為孤兒進(jìn)程，而每當(dāng)系統(tǒng)發(fā)現(xiàn)一個(gè)孤兒進(jìn)程時(shí)，就會(huì)自動(dòng)由1號(hào)進(jìn)程(init)收養(yǎng)它，這樣，原先的子進(jìn)程就會(huì)變成init進(jìn)程的子進(jìn)程

在子進(jìn)程中創(chuàng)建新會(huì)話

更改工作目錄到/, 以便umount一個(gè)文件系統(tǒng)

重設(shè)文件權(quán)限掩碼, 以便擁有完全的寫的權(quán)限, 即重設(shè)繼承來的默認(rèn)文件權(quán)限值

調(diào)用setuid, 讓當(dāng)前進(jìn)程成為新的會(huì)話組長(zhǎng)和進(jìn)程組長(zhǎng)

執(zhí)行第二次fork
關(guān)閉文件描述符, 一般是輸入/輸出和錯(cuò)誤輸出, 重定向到/dev/null

py代碼
https://gist.github.com/jamiesun/3097215

上面守護(hù)進(jìn)程的生成步驟中涉及到了孤兒進(jìn)程
任何孤兒進(jìn)程產(chǎn)生時(shí)都會(huì)立即為系統(tǒng)進(jìn)程init自動(dòng)接收為子進(jìn)程，這一過程也被稱為“收養(yǎng)”. 但由于創(chuàng)建該進(jìn)程的進(jìn)程已不存在，所以仍應(yīng)稱之為“孤兒進(jìn)程”

與之相關(guān)的一個(gè)概念就是 僵尸進(jìn)程了. 當(dāng)子進(jìn)程退出時(shí), 父進(jìn)程需要wait/waitpid系統(tǒng)調(diào)用來讀取子進(jìn)程的exit status, 然后子進(jìn)程被系統(tǒng)回收. 如果父進(jìn)程沒有wait的話, 子進(jìn)程將變成一個(gè)"僵尸進(jìn)程", 內(nèi)核會(huì)釋放這個(gè)子進(jìn)程所有的資源，包括打開的文件占用的內(nèi)存等, 但在進(jìn)程表中仍然有一個(gè)PCB, 記錄進(jìn)程號(hào)和退出狀態(tài)等信息, 并導(dǎo)致進(jìn)程號(hào)一直被占用, 而系統(tǒng)能使用的進(jìn)程號(hào)數(shù)量是有限的(可以用ulimit查看相關(guān)限制), 如果產(chǎn)生大量僵尸進(jìn)程的話, 將因?yàn)闆]有可用的進(jìn)程號(hào)而導(dǎo)致系統(tǒng)不能產(chǎn)生新的進(jìn)程

因此很多自帶重啟功能的服務(wù)實(shí)現(xiàn)就是用wait/waitpid實(shí)現(xiàn)的.
waitpid()會(huì)暫時(shí)停止目前進(jìn)程的執(zhí)行，直到有信號(hào)來到或子進(jìn)程結(jié)束
比如tornado中fork多進(jìn)程就是這樣, 監(jiān)控子進(jìn)程的運(yùn)行狀態(tài), 當(dāng)其意外退出時(shí)自動(dòng)重啟子進(jìn)程

def fork_processes(num_processes, max_restarts=100):
    """Starts multiple worker processes.

    If ``num_processes`` is None or <= 0, we detect the number of cores
    available on this machine and fork that number of child
    processes. If ``num_processes`` is given and > 0, we fork that
    specific number of sub-processes.

    Since we use processes and not threads, there is no shared memory
    between any server code.

    Note that multiple processes are not compatible with the autoreload
    module (or the ``autoreload=True`` option to `tornado.web.Application`
    which defaults to True when ``debug=True``).
    When using multiple processes, no IOLoops can be created or
    referenced until after the call to ``fork_processes``.

    In each child process, ``fork_processes`` returns its *task id*, a
    number between 0 and ``num_processes``.  Processes that exit
    abnormally (due to a signal or non-zero exit status) are restarted
    with the same id (up to ``max_restarts`` times).  In the parent
    process, ``fork_processes`` returns None if all child processes
    have exited normally, but will otherwise only exit by throwing an
    exception.
    """
    global _task_id
    assert _task_id is None
    if num_processes is None or num_processes <= 0:
        num_processes = cpu_count()
    if ioloop.IOLoop.initialized():
        raise RuntimeError("Cannot run in multiple processes: IOLoop instance "
                           "has already been initialized. You cannot call "
                           "IOLoop.instance() before calling start_processes()")
    gen_log.info("Starting %d processes", num_processes)
    children = {}

    def start_child(i):
        pid = os.fork()
        if pid == 0:
            # child process
            _reseed_random()
            global _task_id
            _task_id = i
            return i
        else:
            children[pid] = i
            return None
    for i in range(num_processes):
        id = start_child(i)
        if id is not None:
            return id
    num_restarts = 0
    while children:
        try:
            pid, status = os.wait()
        except OSError as e:
            if errno_from_exception(e) == errno.EINTR:
                continue
            raise
        if pid not in children:
            continue
        id = children.pop(pid)
        if os.WIFSIGNALED(status):
            gen_log.warning("child %d (pid %d) killed by signal %d, restarting",
                            id, pid, os.WTERMSIG(status))
        elif os.WEXITSTATUS(status) != 0:
            gen_log.warning("child %d (pid %d) exited with status %d, restarting",
                            id, pid, os.WEXITSTATUS(status))
        else:
            gen_log.info("child %d (pid %d) exited normally", id, pid)
            continue
        num_restarts += 1
        if num_restarts > max_restarts:
            raise RuntimeError("Too many child restarts, giving up")
        new_id = start_child(id)
        if new_id is not None:
            return new_id
    # All child processes exited cleanly, so exit the master process
    # instead of just returning to right after the call to
    # fork_processes (which will probably just start up another IOLoop
    # unless the caller checks the return value).
    sys.exit(0)

參考: http://bbs.chinaunix.net/thread-4071026-1-1.html