先定位到卡住的进程。
ps -auxf
trace -p PID
发现 write(2, .......)
应该是在 写 文件描述符2 阻塞了
cat /proc/PID/fd
ls -l
查看 文件描述符链接到哪个文件 发现是一个管道 pipe
pipe 的buffer 是 4k
发现 源代码
def execute_cmd(cmd, cwd=''):
cur_cwd = os.getcwd()
if cwd != '':
os.chdir(cwd)
ps = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True, universal_newlines=True)
with ps.stdout:
for line in iter(ps.stdout.readline, b''):
if len(line) > 0:
logger.info(line)
# stdout, stderr = ps.communicate()
status = ps.wait()
引用：
Warning This will deadlock when using stdout=PIPE and/or stderr=PIPE and the child process generates enough output to a pipe such that it blocks waiting for the OS pipe buffer to accept more data. Use communicate() to avoid that.
原因是 子进程往pipe 写数据 ，而另一边没有及时读取
复现 问题：
保存如下两个文件， 执行
python a.py 会发现 a.py 阻塞
# a.py
import subprocess
cmd = "python exe.py"
log_f = open("test_err.log", 'w')
def cmd_exe(cmd):
ps = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True, universal_newlines=True)
with ps.stdout:
for line in iter(ps.stdout.readline, b''):
if len(line) > 0:
print(line)
status = ps.wait()
if __name__ == '__main__':
cmd_exe(cmd)
log_f.close()
print("finish")
# exe.py
import sys
import time
a = ''
for i in range(10000):
a += "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
sys.stderr.write(a)
参考；
