問題

最近寫了涉及線程池及線程的 python 腳本，運行過程中發(fā)現(xiàn)一個有趣的現(xiàn)象，線程池中的工作線程出現(xiàn)問題，引發(fā)了異常，但是主線程沒有捕獲異常，還在發(fā)現(xiàn) BUG 之前一度以為線程池代碼正常返回。

先說重點

這里主要想介紹 python concurrent.futuresthread.ThreadPoolExecutor 線程池中的 worker 引發(fā)異常的時候，并不會直接向上拋起異常，而是需要主線程通過調(diào)用concurrent.futures.Future.exception(timeout=None) 方法主動獲取 worker 的異常。

問題重現(xiàn)及解決

引子

問題主要由這樣一段代碼引起的：

def thread_executor(): logger.info('I am slave. I am working. I am going to sleep 3s') sleep(3) logger.info('Exit thread executor')def main(): thread_obj = threading.Thread(target=thread_executor) while True: logger.info('Master starts thread worker') try: # 工作線程由于某種異常而結(jié)束并退出了，想重啟工作線程的工作，但又不想重復(fù)創(chuàng)建線程 thread_obj.start() # 這一行會報錯，同一線程不能重復(fù)啟動 except Exception as e: logger.error('Master start thread error', exc_info=True) raise e logger.info('Master is going to sleep 5s') sleep(5)

上面這段代碼的功能如注釋中解釋的，主要要實現(xiàn)類似生產(chǎn)者消費者的功能，工作線程一直去生產(chǎn)資源，主線程去消費工作線程生產(chǎn)的資源。但是工作線程由于異常推出了，想重新啟動生產(chǎn)工作。顯然，這個代碼會報錯。

運行結(jié)果:

thread: MainThread [INFO] Master starts thread workerthread: Thread-1 [INFO] I am slave. I am working. I am going to sleep 3sthread: MainThread [INFO] Master is going to sleep 5sthread: Thread-1 [INFO] Exit thread executor because of some exceptionthread: MainThread [INFO] Master starts thread workerthread: MainThread [ERROR] Master start thread errorTraceback (most recent call last):File 'xxx.py', line 47, in main thread_obj.start()File 'E:anacondalibthreading.py', line 843, in start raise RuntimeError('threads can only be started once')RuntimeError: threads can only be started onceTraceback (most recent call last):File 'xxx.py', line 56, in <module> main()File 'xxx.py', line 50, in main raise eFile 'xxx.py', line 47, in main thread_obj.start()File 'E:anacondalibthreading.py', line 843, in start raise RuntimeError('threads can only be started once')RuntimeError: threads can only be started once

切入正題

然而腳本還有其他業(yè)務(wù)代碼要運行，所以需要把上面的資源生產(chǎn)和消費的代碼放到一個線程里完成，所以引入線程池來執(zhí)行這段代碼：

def thread_executor(): while True: logger.info('I am slave. I am working. I am going to sleep 3s') sleep(3) logger.info('Exit thread executor because of some exception') breakdef main(): thread_obj = threading.Thread(target=thread_executor) while True: logger.info('Master starts thread worker') # 工作線程由于某種異常而結(jié)束并退出了，想重啟工作線程的工作，但又不想重復(fù)創(chuàng)建線程 # 沒有想到這里會有異常 thread_obj.start() # 這一行會報錯，同一線程不能重復(fù)啟動 logger.info('Master is going to sleep 5s') sleep(5)def thread_pool_main(): thread_obj = ThreadPoolExecutor(max_workers=1, thread_name_prefix='WorkExecutor') logger.info('Master ThreadPool Executor starts thread worker') thread_obj.submit(main) while True: logger.info('Master ThreadPool Executor is going to sleep 5s') sleep(5)if __name__ == ’__main__’: thread_pool_main()

代碼運行結(jié)果如下:

INFO [thread: MainThread] Master ThreadPool Executor starts thread workerINFO [thread: WorkExecutor_0] Master starts thread workerINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: Thread-1] I am slave. I am working. I am going to sleep 3sINFO [thread: WorkExecutor_0] Master is going to sleep 5sINFO [thread: Thread-1] Exit thread executor because of some exceptionINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: WorkExecutor_0] Master starts thread workerINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5s... ...

顯然，由上面的結(jié)果，在線程池 worker 執(zhí)行到 INFO [thread: WorkExecutor_0] Master starts thread worker 的時候，是會有異常產(chǎn)生的，但是整個代碼并沒有拋棄任何異常。

解決方法

發(fā)現(xiàn)上面的 bug 后，想在線程池 worker 出錯的時候，把異常記錄到日志。查閱資料，要獲取線程池的異常信息，需要調(diào)用 concurrent.futures.Future.exception(timeout=None) 方法，為了記錄日志，這里加了線程池執(zhí)行結(jié)束的回調(diào)函數(shù)。同時，日志中記錄異常信息，用了 logging.exception() 方法。

def thread_executor(): while True: logger.info('I am slave. I am working. I am going to sleep 3s') sleep(3) logger.info('Exit thread executor because of some exception') breakdef main(): thread_obj = threading.Thread(target=thread_executor) while True: logger.info('Master starts thread worker') # 工作線程由于某種異常而結(jié)束并退出了，想重啟工作線程的工作，但又不想重復(fù)創(chuàng)建線程 # 沒有想到這里會有異常 thread_obj.start() # 這一行會報錯，同一線程不能重復(fù)啟動 logger.info('Master is going to sleep 5s') sleep(5)def thread_pool_callback(worker): logger.info('called thread pool executor callback function') worker_exception = worker.exception() if worker_exception: logger.exception('Worker return exception: {}'.format(worker_exception))def thread_pool_main(): thread_obj = ThreadPoolExecutor(max_workers=1, thread_name_prefix='WorkExecutor') logger.info('Master ThreadPool Executor starts thread worker') thread_pool_exc = thread_obj.submit(main) thread_pool_exc.add_done_callback(thread_pool_callback) # logger.info('thread pool exception: {}'.format(thread_pool_exc.exception())) while True: logger.info('Master ThreadPool Executor is going to sleep 5s') sleep(5)if __name__ == ’__main__’: thread_pool_main()

代碼運行結(jié)果:

INFO [thread: MainThread] Master ThreadPool Executor starts thread workerINFO [thread: WorkExecutor_0] Master starts thread workerINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: Thread-1] I am slave. I am working. I am going to sleep 3sINFO [thread: WorkExecutor_0] Master is going to sleep 5sINFO [thread: Thread-1] Exit thread executor because of some exceptionINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: WorkExecutor_0] Master starts thread workerINFO [thread: WorkExecutor_0] called thread pool executor callback functionERROR [thread: WorkExecutor_0] Worker return exception: threads can only be started onceTraceback (most recent call last):File 'E:anacondalibconcurrentfuturesthread.py', line 57, in run result = self.fn(*self.args, **self.kwargs)File 'xxxx.py', line 46, in main thread_obj.start() # 這一行會報錯，同一線程不能重復(fù)啟動File 'E:anacondalibthreading.py', line 843, in start raise RuntimeError('threads can only be started once')RuntimeError: threads can only be started onceINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5sINFO [thread: MainThread] Master ThreadPool Executor is going to sleep 5s... ...

最終的寫法

其實，上面寫法中，想重復(fù)利用一個線程去實現(xiàn)生產(chǎn)者線程的實現(xiàn)方法是有問題的，在此處，一般情況下，線程執(zhí)行結(jié)束后，線程資源會被會被操作系統(tǒng)，所以線程不能被重復(fù)調(diào)用 start() 。

一種可行的實現(xiàn)方式就是，用線程池替代。當(dāng)然，這樣做得注意上面提到的線程池執(zhí)行體的異常捕獲問題。

def thread_executor(): while True: logger.info('I am slave. I am working. I am going to sleep 3s') sleep(3) logger.info('Exit thread executor because of some exception') breakdef executor_callback(worker): logger.info('called worker callback function') worker_exception = worker.exception() if worker_exception: logger.exception('Worker return exception: {}'.format(worker_exception)) # raise worker_exceptiondef main(): slave_thread_pool = ThreadPoolExecutor(max_workers=1, thread_name_prefix='SlaveExecutor') restart_flag = False while True: logger.info('Master starts thread worker') if not restart_flag: restart_flag = not restart_flag logger.info('Restart Slave work') slave_thread_pool.submit(thread_executor).add_done_callback(executor_callback) logger.info('Master is going to sleep 5s') sleep(5)

總結(jié)

這個問題主要還是因為對 Python 的 concurrent.futuresthread.ThreadPoolExecutor 不夠了解導(dǎo)致的，接觸這個包是在書本上，但是書本沒完全介紹包的全部 API 及用法，所以代碼產(chǎn)生異常情況后，DEBUG 了許久在真正找到問題所在。查閱 python docs 后才對其完整用法有所認(rèn)識，所以，以后學(xué)習(xí)新的 python 包的時候還是可以查一查官方文檔的。

參考資料

英文版: docs of python concurrent.futures

中文版: python docs concurrent.futures — 啟動并行任務(wù)

exception(timeout=None)

返回由調(diào)用引發(fā)的異常。如果調(diào)用還沒完成那么這個方法將等待 timeout 秒。如果在 timeout 秒內(nèi)沒有執(zhí)行完成，concurrent.futures.TimeoutError 將會被觸發(fā)。timeout 可以是整數(shù)或浮點數(shù)。如果 timeout 沒有指定或為 None，那么等待時間就沒有限制。

如果 futrue 在完成前被取消則 CancelledError 將被觸發(fā)。

如果調(diào)用正常完成那么返回 None。

add_done_callback(fn)

附加可調(diào)用 fn 到期程。當(dāng)期程被取消或完成運行時，將會調(diào)用 fn，而這個期程將作為它唯一的參數(shù)。

加入的可調(diào)用對象總被屬于添加它們的進程中的線程按加入的順序調(diào)用。如果可調(diào)用對象引發(fā)一個 Exception 子類，它會被記錄下來并被忽略掉。如果可調(diào)用對象引發(fā)一個 BaseException 子類，這個行為沒有定義。

如果期程已經(jīng)完成或已取消，fn 會被立即調(diào)用。

以上這篇解決python ThreadPoolExecutor 線程池中的異常捕獲問題就是小編分享給大家的全部內(nèi)容了，希望能給大家一個參考，也希望大家多多支持好吧啦網(wǎng)。