piping output of subprocess.Popen to files

Question

I need to launch a number of long-running processes with subprocess.Popen, and would like to have the stdout and stderr from each automatically piped to separate log files. Each process will run simultaneously for several minutes, and I want two log files (stdout and stderr) per process to be written to as the processes run.

Do I need to continually call p.communicate() on each process in a loop in order to update each log file, or is there some way to invoke the original Popen command so that stdout and stderr are automatically streamed to open file handles?

Answer 1

Per the docs,

stdin, stdout and stderr specify the executed programs’ standard input, standard output and standard error file handles, respectively. Valid values are PIPE, an existing file descriptor (a positive integer), an existing file object, and None.

So just pass the open-for-writing file objects as named arguments stdout= and stderr= and you should be fine!

Answer 2

I am simultaneously running two subprocesses, and saving the output from both into a single log file. I have also built in a timeout to handle hung subprocesses. When the output gets too big, the timeout always triggers, and none of the stdout from either subprocess gets saved to the log file. The answer posed by Alex above does not solve it.

# Currently open log file.
log = None

# If we send stdout to subprocess.PIPE, the tests with lots of output fill up the pipe and
# make the script hang. So, write the subprocess's stdout directly to the log file.
def run(cmd, logfile):
   #print os.getcwd()
   #print ("Running test: %s" % cmd)
   global log
   p = subprocess.Popen(cmd, shell=True, universal_newlines = True, stderr=subprocess.STDOUT, stdout=logfile)
   log = logfile
   return p


# To make a subprocess capable of timing out
class Alarm(Exception):
   pass

def alarm_handler(signum, frame):
   log.flush()
   raise Alarm


####
## This function runs a given command with the given flags, and records the
## results in a log file. 
####
def runTest(cmd_path, flags, name):

  log = open(name, 'w')

  print >> log, "header"
  log.flush()

  cmd1_ret = run(cmd_path + "command1 " + flags, log)
  log.flush()
  cmd2_ret = run(cmd_path + "command2", log)
  #log.flush()
  sys.stdout.flush()

  start_timer = time.time()  # time how long this took to finish

  signal.signal(signal.SIGALRM, alarm_handler)
  signal.alarm(5)  #seconds

  try:
    cmd1_ret.communicate()

  except Alarm:
    print "myScript.py: Oops, taking too long!"
    kill_string = ("kill -9 %d" % cmd1_ret.pid)
    os.system(kill_string)
    kill_string = ("kill -9 %d" % cmd2_ret.pid)
    os.system(kill_string)
    #sys.exit()

  end_timer = time.time()
  print >> log, "closing message"

  log.close()

Answer 3

You can pass stdout and stderr as parameters to Popen()

subprocess.Popen(self, args, bufsize=0, executable=None, stdin=None, stdout=None,
                 stderr=None, preexec_fn=None, close_fds=False, shell=False,
                 cwd=None, env=None, universal_newlines=False, startupinfo=None, 
                 creationflags=0)

For example

>>> import subprocess
>>> with open("stdout.txt","wb") as out, open("stderr.txt","wb") as err:
...    subprocess.Popen("ls",stdout=out,stderr=err)
... 
<subprocess.Popen object at 0xa3519ec>
>>>