Python multiprocessing RuntimeError 错误排查与解决方案

木白

发布时间：2026-10-15| 4人浏览过

原创

本文详细解析 Python 中 multiprocessing 模块常见的 RuntimeError 错误。主要涵盖两类问题：一是 Windows 系统下未使用 if __name__ == '__main__' 导致的启动引导错误，需将逻辑封装并添加保护语句；二是 Python 与 numpy 版本不兼容引发的运行时检查失败。文章提供了通过 PyCharm 降级 numpy 版本的具体操作步骤，帮助开发者快速修复代码。

以下是该错误的详细解释说明

RuntimeError: \n An attempt has been made to start a new process before the\n current process has finished its bootstrapping phase.\n \n This probably means that you are not using fork to start your\n child processes and you have forgotten to use the proper idiom\n in the main module:\n \n if __name__ == '__main__':\n freeze_support()\n ...\n \n The "freeze_support()" line can be omitted if the program\n is not going to be frozen to produce an executable.

以下是触发该错误的原始代码片段

import multiprocessing as mp
import time
from urllib.request import urlopen,urljoin
from bs4 import BeautifulSoup
import re
 
base_url = "https://morvanzhou.github.io/"
 
#crawl爬取网页
def crawl(url):
    response = urlopen(url)
    time.sleep(0.1)
    return response.read().decode()
 
#parse解析网页
def parse(html):
    soup = BeautifulSoup(html,'html.parser')
    urls = soup.find_all('a',{"href":re.compile('^/.+?/$')})
    title = soup.find('h1').get_text().strip()
    page_urls = set([urljoin(base_url,url['href'])for url in urls])
    url = soup.find('meta',{'property':"og:url"})['content']
    return title,page_urls,url
 
unseen = set([base_url])
seen = set()
restricted_crawl = True
 
pool = mp.Pool(4)
count, t1 = 1, time.time()
while len(unseen) != 0:                 # still get some url to visit
    if restricted_crawl and len(seen) > 20:
            break
    print('\nDistributed Crawling...')
    crawl_jobs = [pool.apply_async(crawl, args=(url,)) for url in unseen]
    htmls = [j.get() for j in crawl_jobs]      # request connection
 
    print('\nDistributed Parsing...')
    parse_jobs = [pool.apply_async(parse, args=(html,)) for html in htmls]
    results = [j.get() for j in parse_jobs]    # parse html
 
    print('\nAnalysing...')
    seen.update(unseen)         # seen the crawled
    unseen.clear()              # nothing unseen
 
    for title, page_urls, url in results:
        print(count, title, url)
        count += 1
        unseen.update(page_urls - seen)     # get new url to crawl
print('Total time: %.1f s' % (time.time()-t1))    # 16 s !!!

以下是修正后的正确代码示例

import multiprocessing as mp
import time
from urllib.request import urlopen,urljoin
from bs4 import BeautifulSoup
import re
 
base_url = "https://morvanzhou.github.io/"
 
#crawl爬取网页
def crawl(url):
    response = urlopen(url)
    time.sleep(0.1)
    return response.read().decode()
 
#parse解析网页
def parse(html):
    soup = BeautifulSoup(html,'html.parser')
    urls = soup.find_all('a',{"href":re.compile('^/.+?/$')})
    title = soup.find('h1').get_text().strip()
    page_urls = set([urljoin(base_url,url['href'])for url in urls])
    url = soup.find('meta',{'property':"og:url"})['content']
    return title,page_urls,url
 
def main():
    unseen = set([base_url])
    seen = set()
    restricted_crawl = True
 
    pool = mp.Pool(4)
    count, t1 = 1, time.time()
    while len(unseen) != 0:                 # still get some url to visit
        if restricted_crawl and len(seen) > 20:
                break
        print('\nDistributed Crawling...')
        crawl_jobs = [pool.apply_async(crawl, args=(url,)) for url in unseen]
        htmls = [j.get() for j in crawl_jobs]      # request connection
 
        print('\nDistributed Parsing...')
        parse_jobs = [pool.apply_async(parse, args=(html,)) for html in htmls]
        results = [j.get() for j in parse_jobs]    # parse html
 
        print('\nAnalysing...')
        seen.update(unseen)         # seen the crawled
        unseen.clear()              # nothing unseen
 
        for title, page_urls, url in results:
            print(count, title, url)
            count += 1
            unseen.update(page_urls - seen)     # get new url to crawl
    print('Total time: %.1f s' % (time.time()-t1))    # 16 s !!!
 
 
if __name__ == '__main__':
    main()

综上所述，解决此类问题的核心方法是将运行逻辑整合到一个函数中，并在主模块入口处加入