multitenancy in supervisor with fab deployment

[David Novakovic]

Hey,

I found a bit of a quirk today when trying to deploy two  mezzanine sites to the same server using the vanilla fabric deployment.

Once I had to instance of gunicorn, each in their own program group and each running on different ports I found that the gunicorn in the other process group would start trying to listen on the same port as gunicorn in the first group.

When I renamed the program to "gunicorn2" on the second site and told the appropriate program group to use it, the issue went away. Below I've shown the working combination.

Now, I know this is most likely an issue with supervisor, but i wanted to raise it here to see if anyone else has had this issue or maybe I'm missing something obvious?

[group:site_one]

programs=gunicorn

[program:gunicorn]

command=..

[group:site_two]

programs=gunicorn2

[program:gunicorn2]

command=...

Cheers!

D

[Stephen McDonald]

Thanks David - that's a bug and your solution is correct. 

It was recently fixed here:

https://github.com/stephenmcd/mezzanine/commit/66041bcbb464dd08c3d2e84a971cc937d53273d2

--
Stephen McDonald
http://jupo.org

[David Novakovic]

OK, I'll move to using mezzanine from github rather than the 1.4.1 defined in the requirements file.

I'm assuming for stuff like this I'll have to update the relevant files in the project template I'm using from the older version.

Are there a list of changes where I need to update my current project somewhere?

D

[Stephen McDonald]

I haven't had the time to put a list like that together unfortunately.

Sorry about that - it should be part of the release process but it hasn't been.