django change default runserver port

Question

I would like to make the default port that manage.py runserver listens on specifiable in an extraneous config.ini. Is there an easier fix than parsing sys.argv inside manage.py and inserting the configured port?

The goal is to run ./manage.py runserver without having to specify address and port every time but having it take the arguments from the config.ini.

Solution 1

create a bash script with the following:

#!/bin/bash
exec ./manage.py runserver 0.0.0.0:<your_port>

save it as runserver in the same dir as manage.py

chmod +x runserver

and run it as

./runserver

Solution 2

Actually the easiest way to change (only) port in development Django server is just like:

python manage.py runserver 7000

that should run development server on http://127.0.0.1:7000/

Solution 3

As of Django 1.9, the simplest solution I have found (based on Quentin Stafford-Fraser's solution) is to add a few lines to manage.py which dynamically modify the default port number before invoking the runserver command:

if __name__ == "__main__":
    os.environ.setdefault("DJANGO_SETTINGS_MODULE", "project.settings.dev")

    import django
    django.setup()

    # Override default port for `runserver` command
    from django.core.management.commands.runserver import Command as runserver
    runserver.default_port = "8080"

    from django.core.management import execute_from_command_line

    execute_from_command_line(sys.argv)

Solution 4

All of the following commands are possible to change the port while running django:

python manage.py runserver 127.0.0.1:7000

python manage.py runserver 7000

python manage.py runserver 0:7000

Solution 5

Create a subclass of django.core.management.commands.runserver.Command and overwrite the default_port member. Save the file as a management command of your own, e.g. under <app-name>/management/commands/runserver.py:

from django.conf import settings
from django.core.management.commands import runserver

class Command(runserver.Command):
    default_port = settings.RUNSERVER_PORT

I'm loading the default port form settings here (which in turn reads other configuration files), but you could just as well read it from some other file directly.

View more solutions

Comments

[Jonathan Hartley]

If you run the command from your command line history, then it's even fewer keystrokes. E.g. for me it's "<ctrl-r>runs<enter>", and it automatically runs the same command as last time, so the port etc is included.

[jonny]

Either that, or I am thinking about adding a custom management command.

[fixmycode]

You can't run the development server programmatically, so a custom command won't work unless it calls the shell using something like call. If this solution works for you please make sure to mark it as solved.

[jonny]

I actually started using supervisor for this now, that makes it even easier to manage. :-) But yours is probably the cleanest solution other than writing a dedicated runserver script.

[fixmycode]

Supervisor is an good solution for I wouldn't recommend it to run the development environment. You lose the advantage of having the server output on the terminal, among other things. If you really want to use supervisor my advice would be to use it with a fully featured WSGI server like Gunicorn. Please don't run the development server as your production server...

[PurpleDiane]

Although it doesn't answer the original question exactly, and the indenting is messed up (code should be indented from "import django" on down), I prefer this answer because it is entirely self-contained and does not require changing what is entered on the command line.

[Flimm]

@PurpleDiane The indentation is fixed.

[arogachev]

Modifying the sources of dependent packages is considered bad practice. Changes can be lost on the update / reinstalling.

[Feuermurmel]

This is the worst suggestion of all here. IMHO, editing a file from a distribution is never a good idea and leads to confusion because the change is not tracked by a VCS and gets easily overwritten.

[Amrendra]

That's right @arogachev, It is just an option by which you can make default port and host. But it's really not a good practice to modify dependent packages. you can set the IP and port when you run your server in command prompt as well.

[Admin]

This is useful but not greate when juggling multiple projects at a time - I would have accepted the answer below which specifies the port to be used for each distinct project. Just my opinion tho.

[NaturalBornCamper]

My manage.py doesn't really look like that, I guess Django changed since

[Vargas]

Currently (as of 2.0.3) you can just add: from django.core.management.commands.runserver import Command as runserver; runserver.default_port = "8080 to your manage.py. You can also change the listening address with: runserver.default_addr

[physicalattraction]

This answer is about changing the port, not changing the default port.

[physicalattraction]

This seems like the best solution, however, my Django 1.8.14 doesn't recognize my self-created file runserver.py. Should I register it somewhere?

[Feuermurmel]

@physicalattraction Your file is probably not in the right location. docs.djangoproject.com/en/2.0/howto/custom-management-comman‌​ds starts with a description of where to put the python module for the runserver command. Also maybe you have another app also registering a command named runserver. Try renaming your command and see whether it is recognized.

[physicalattraction]

I put it in <app_name>/management/commands/runserver.py, but then Django's original runserver is used. When I rename it to run_server.py, it is recognized. I don't see anything special about runserver on the page you link to.

[x-yuri]

The thing is staticfiles does exactly what you suggest. So following your instructions breaks serving static files in development. It's best to import from django.contrib.staticfiles.management.commands.

[Aderemi Dayo]

The 0.0.0.0 is important especially if you are running python inside docker container

[JuSTMOnIcAjUSTmONiCAJusTMoNICa]

"The goal is to run ./manage.py runserver without having to specify address and port every time"

[Sathindu]

It's not recommended to edit the Django module at the site packages. It can be updated by new version. Also it affect to all Django apps.

[E.Serra]

This should be the correct answer, accepted one also implies the IP Address, and executing manage.py for some reason, instead of calling it via python

[basquiatraphaeu]

Also, make sure to add your port to the ALLOWED_HOSTS variable.