Bug #534
closedlighttpd does not restart dead fastcgi backends
Description
If lighttpd is set up to start up a FastCGI backend and use it to serve requests, it works fine until the backend crashes or is killed, at which point it starts issuing errors like:
2006-02-20 20:29:05: (mod_fastcgi.c.1561) connect failed: 249 Connection refused 61 0 /tmp/python-infogami-lighttpd.socket-0
on every request. Instead of issuing errors, it should restart the backend.
For example, I have a simple FastCGI backend that prints "hello world" called foo.py:
server.modules = ("mod_fastcgi") server.document-root = "/home/aaronsw/funnybug/" server.port = 8081 fastcgi.server = ( "/foo.py" => ( "main" => ( "socket" => "/tmp/bugfinder.socket", "bin-path" => "/home/aaronsw/funnybug/foo.py", "max-procs" => 1 ) ))
When I start up the server with this config, it starts up a FastCGI backend. If I kill the FastCGI backend, my entire site dies. I'm pretty sure this didn't happen before 1.4.9.
-- me
Updated by Anonymous over 18 years ago
I just wanted to confirm this. I have been using lighttpd+factcgi on my development server, and started having database problems. Come to find out, my /var partition had been filled up with a 2.4GB error.log
The errors start on 2006-02-23, which I assume is when I upgraded lighttpd.
-- mmarshall
Updated by stbuehler almost 17 years ago
- Status changed from New to Fixed
- Resolution set to fixed
Works for me in current version, so i think this got fixed.
Also available in: Atom