Hello all, We successfully run a number of FCGI apps written in Python under supervisor using plain [program:xxxx] entries, behind Lighttpd as the HTTP proxy. Lighttpd runs on a physically separate host, directing incoming requests to preconfigured TCP sockets on 'worker' servers for each site. Each app is configured to use flup to handle the FastCGI protocol.
I've taken a look at the [fcgi-program:xxxx] support in supervisor, to understand whether this better supports our deployment scenario. I setup a test-app with a 'socket=tcp:0.0.0.0:9302' entry, and using 'lsof -i' can see that supervisor starts listening on this socket. The documentation seems a little unclear on how supervisor now forwards incoming requests to my app workers: 1. Does each app still need to include flup to decompose the FastCGI protocol, or is supervisor translating the requests? 2. If requests are forwarded untranslated, can supervisor proxy the incoming TCP socket request to a Unix domain socket locally? 3. If requests are forwarded on stdin, how do you keep the apps long-running? In my quick test, flup treats such requests on stdin as oneshot CGI and exits the app after handling the request. I'd appreciate a small worked example in the documentation to make this easier to understand. Cheers, Sigurd
_______________________________________________ Supervisor-users mailing list Supervisor-users@lists.supervisord.org http://lists.supervisord.org/mailman/listinfo/supervisor-users
