Getting pid of the program just started in Python

Question

I need to create startup and stop scripts for a python program. My temporary solution was to have the startup.sh:

python -m root.scripts.run 
python server.py 'localhost' 42345

This starts my program and a server that is needed to do some computations. Now for stop.sh I just do:

killall -m Python
killall python
killall Python

This works and stops my program. However I need to find a less "brutal" solution, since this obviously kills all python related processes. A solution I'm thinking of (not sure if it is possible) is in the startup.sh to get the PID of the two started processes and store them in a file somewhere. Then for the stop.sh I would just get those pids and kill those processes. Now I have a few questions:

Is my proposed solution viable? If so how can I get the pid of the recently started processes ? In terms of storing them and then getting them from a file, is this doable and if so how?
Do you have any other proposed solutions for my problem?
In terms of cross-platform, how hard would it be to mimic something like this on windows?

Ferdinand Beyer · Accepted Answer · 2011-08-04 07:52:36Z

4

Use the subprocess module to start the program from Python (this is the recommended way anyway). The object returned from subprocess.Popen has terminate() and kill() methods as well as a pid attribute that you can save and use for stopping the process again.

On Unix systems, you might consider saving the PID in a file in /var/run/<yourscript>.pid (see the Filesystem Hierarchy Standard), then read this file to retrieve the PID when requested to stop.

edited Aug 4, 2011 at 7:52

answered Aug 4, 2011 at 6:56

Ferdinand Beyer

67.6k18 gold badges160 silver badges147 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Bogdan Over a year ago

Thanks for the input. But I'm not sure if I understand your suggestion. So you are saying I should replace my startup.sh shell script with a python script. Then in this python script use subprocess.Popen(root.scripts.run) subprocess.Popen(server.py). From this i get my PID's and store them somewhere. Then replace the stop.sh shell with a python script also and get those pids and kill the processes. Is this correct ?

Ferdinand Beyer Over a year ago

Maybe I misunderstood your question. Why don't you just write an init shell script based on some existing one? For example, Debian systems come with a skeleton file /etc/init.d/skeleton that you can easily adapt to start your Python program.

alnacle · Accepted Answer · 2011-08-04 09:14:48Z

1

I would replace your startup.sh for a Python script which runs both scripts. Use subprocess.Popen() for running new processes and terminate() for killing them.

Python solves your third question. Otherwise you have to use cygwin/mingw under Windows.

answered Aug 4, 2011 at 9:14

alnacle

3621 gold badge3 silver badges9 bronze badges

2 Comments

Bogdan Over a year ago

But I need to be able to call start and stop separately. I mean the user should be able to call let's say python start_program.py to start the application. Then if something goes wrong or he just wants to stop he would python stop_program.py. Your suggestion would be to create only one Python script which would be started like python app.py "start" respectively python app.py "stop" ?

alnacle Over a year ago

Exactly. That's the idea: Use app.py [start|stop] and parse the argument from the command line. You could even add more flags or arguments by using argparse module. This approach works both Linux and Windows.

Collectives™ on Stack Overflow

Getting pid of the program just started in Python

2 Answers 2

2 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related