Supervising FastCGI Listeners
bougyman <[email protected]>
| Newsgroups | gmane.comp.web.fastcgi.devel |
|---|---|
| Message-ID | <[email protected]> |
I've get a setup where i have many fastcgi listeners on a private
backend network that are server by a Linux Virtual Server Director to
a lighttpd webserver. This is done so (ruby on rails dispatch.fcgi)
listeners can be added on the fly at the LVS directory without having
to restart lighttpd to point it to a new fastcgi listeners. I've been
trying to manage the listeners themselves through a supervisor (i've
tried daemontools and runit) in order to do monitoring on them and
restart runaway processes. My problem is this: when the supervisor
starts the fastcgi listener, currently with a runscript:
---- Runscript Start
#!/bin/sh -e
railnumber=$(basename $PWD)
exec 2>&1 \
fghack cgi-fcgi -start -connect
$myip:197${railnumber}/path/to/railsapp/public/dispatch.fcgi \
<&-
---- Runscript End
The cgi-fcgi process STILL detaches and does not accept signals from
the supervisor.
Running a debian packaged libfcgi0
libfcgi0 2.4.0-5 Shared library of FastCGI
And the latest gem ruby-fastcgi library:
ruby-fcgi-0.8.5
Is there a way to reliably get this listener _not_ to detach so it
will cleanly go up and down via the supervisor?
--
bougyman
___________________________________
fastcgi-developers mailing list
http://fastcgi.com/fastcgi-developers/