I'm trying to run a Pyramid app with Heroku's Foreman (via the these
instructions from the Pyramid docs<http://docs.pylonsproject.org/projects/pyramid_cookbook/en/latest/dep...>).
I've already used these lines to try creating run:

python setup.py develop
python runapp.py

But it keeps giving me this error when I try to foreman start:

zak$ sudo foreman start
12:27:59 web.1     | started with pid 1308
12:27:59 web.1     | /usr/local/foreman/bin/runner: line 36:
/Users/MyProject/run: No such file or directory
12:27:59 web.1     | process terminated
12:27:59 system    | sending SIGTERM to all processes

this is the contents of my procfile:

web: ./run

I've already created a runapp.py. Why is it not able to located run?

I use the following template for running pyramid apps on heroku.
https://github.com/bcajes/pyramid-heroku-template

I'm pretty sure my runapp.py is correct...but for some reason, foreman is
not detecting the run command. Do you know what determines the run command?
What is that line from the procfile pointing to exactly?

Also a quick correction:

12:27:59 web.1     | /usr/local/foreman/bin/runner: line 36:
/Users/zak/MyProject/run: No such file or directory

You should be calling your runapp.py.  run isn't supposed to be command.
eg. my proc file:
web: ./run.py

Hey Zak,

I wrote the most updated version of those Heroku docs. Are you trying to
run your stack locally or have it deploy successfully to Heroku? Generally
I just use the standard "paster serve development.ini" or "pserve
development.ini", etc. when testing locally.

Doug

Doug -

I'm trying to run the stack locally with foreman. I'm not using "paster
serve development.ini" because I'm using Gunicorn and it takes arguments
that can't be set exclusively in an INI file. In the Pyramid tutorial for
Heroku, the example says to create a procfile with the line "web: ./run" This
is the source of my confusion. What is "run" exactly? How am I suppose to
generate that command? web: python runapp.py seems to work because I do
actually have a runapp.py in the app directory, but that is different than
the example in the tutorial.

correction: "I'm not using "paster serve development.ini" because I'm using
Gunicorn and it takes arguments that can't be set exclusively in an INI
file *to my knowledge*"

The "run" file is just an executable shell file (it's created in the step
after creating the Procfile). You can create it by doing something like:
"vim run" and then add the following lines to that file:

#!/bin/bash
python setup.py develop
python runapp.py

 After saving that file, run "chmod +x run" to make that file executable. I
hope that helps!

P.S. I'm not an expert with gunicorn, so I'll yield to other's expertise on
that.

Hi Zak,

Like you, I also tried to use a run shell script when deploying my Pyramid
aplication locally on my own server (not Heroku).   Maybe I can shed some
light on the run script.  

The run script is just a shell script (in my case a BASH script) to run
your pyramid app.  I also use Gunicorn.  In my project root directory I had
a config.py file to store the gunicorn initialization variables.  It's
something like:

import multiprocessing

workers = multiprocessing.cpu_count() * 2 +1
timeout = 60
daemon = True
pidfile = "/srv/web/run/test/test.pid"
errorlog = "/srv/web/log/test/test.log"

Then in my run script I have:

#!/bin/bash
gunicorn_paster --config=<path_to_config.py_file>
/srv/web/prod/test/production.ini

I'm pretty sure this will be similar when used under Heroku.  

I hope this helps your use case.  

Regards,
Mark Huang