[Switching 839...@bugs.debian.org → 839...@bugs.debian.org]
Hi Neil,
Thanks for your patience and understanding; didn't mean to break your
installation and I can understand that the initscript (etc.) sticking
around would have been rather confusing. That (ie. #839183) is now fixed.
> OK - should the process just start automatically via apache or similar?
No; it really needs to run as a regular standalone process rather than
being managed by a web browser. Apache might (?) support such a thing but
its not only a bit old-fashioned, nginx — the far more common web server
for gunicorn IME! — does not support such things.
> What is the new upstream method for starting gunicorn?
They basically don't have one and simply point you towards some kind of
process monitor such as runit, gaffer, etc. This is mostly because most of
"their" users are using some combination of:
- virtualenv
- python2 or python3
- weird install paths
- user/groups
> My test system uses systemd
Lets get that bit working first. Can you try a .service file like:
== lavaserver.service ================================================
[Unit]
Description=lavaserver
After=network.target
[Install]
WantedBy=multi-user.target
[Service]
User=lavaserver
Group=lavaserver
Restart=always
ExecStart=/usr/bin/gunicorn --workers=4 --log-level=debug lava_server.wsgi
ExecReload=/bin/kill -s HUP $MAINPID
ExecStop=/bin/kill -s TERM $MAINPID
PrivateTmp=true
====================================================================
Regards,
--
,''`.
: :' : Chris Lamb
`. `'` la...@debian.org / chris-lamb.co.uk
`-
