2

On my Raspbian machine, I want a successful OpenVPN connection to my VPN provider to initiate a Python script, that persistently should run in the background as long as the VPN connection is active. If/when the VPN connection goes offline, I would like to terminate the Python script.

My thinking is this: the up and down flags in OpenVPN's configuration file will execute shell scripts that start and stop a systemd service. That service Requires the openvpn.service and cannot (should not) run without it.

What's complicating things is that I want to run the Python script inside a Virtualenv which in bash is accessible by alias coding="cd /home/p1/Coding/Python/Virtual/py279/bin; source activate; cd /home/p1/Coding/Python/Projects". The location of the Python file is in the latter folder.

However, it is my understanding that systemd cannot run command lines in the way as the alias coding above would.

So I guess a solution could be for systemd to execute a .service that executes a .sh script that in turn executes the .py Python script.

Schematic:

OpenVPN upovpn_up.shsudo systemctl start vpn_up.serviceExecStart=/path/to/start_pythonscript.shcoding; python my_script.py

OpenVPN downovpn_down.shsudo systemctl stop vpn_up.service

My assumption is that there are a few caveats and/or problems with this, especially regarding how to properly terminate the Python script.

Questions:

  • Is this feasible?
  • What sections and service commands in the service files are needed for this to work?
  • Are there any special considerations that must be made for this overall arrangement, e.g. proper clean-up after terminating the Python script, what happens at system shutdown/reboot, etc?
3

You should first simplify your alias, there is no need to do the activate stuff. What is important is the first, shebang, line of the python script. It should read:

 #!/path/to/your/virtualenv/bin/python

If the script is executable ( chmod +x script_name ) then the above line make sure that the python installed in virtualenv is called, and that ensures that packages installed in the virtualenv are taken.

You can make an alias to the script, or put the script in some directory that is in your PATH.


In the systemd service file you should use the full path to the script. Such a script could look like:

[Unit]
Description=your program description
Requires=any services that need to be installed
After=run after this service

[Service]
Restart=always
ExecStart=/root/bin/application start
ExecStop=/root/bin/application stop

[Install]
WantedBy=multi-user.target

The above assumes your python script is in /root/bin/application and that it starts if passed in the start commandline argument, and that when it is called with stop, somehow the second invocation knows how to notify the started application and gracefully shut down. IIRC if you don't have a ExecStop systemd will try to stop/kill your app, which you could catch with a signal handler

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.