On Sat, Jun 20, 2009 at 9:11 AM, Aristotle Pagaltzis<[email protected]> wrote: > Hi Brad, > > * Brad Bowman <[email protected]> [2009-06-09 10:05]: >> Aristotle Pagaltzis wrote: >> >I like to use ::Engine::HTTP::Prefork coupled with whatever >> >reverse proxy server strikes one’s fancy (whether it be Squid, >> >Apache mod_proxy, Varnish, lighttpd, whatever). Additionally I >> >like to use ::Plugin::Static::Simple, sending proper Expires >> >headers so that the reverse proxy will keep those cached files >> >around forever. >> > >> >That takes decoupling to its logical conclusion: the >> >application server is standalone and works completely >> >independently from the internet-facing server. You can fire >> >requests at it like you would at any webserver. You can use >> >the same engine during development and in production. There >> >are more advantages, but I forget. >> > >> >It’s all very, very nice. >> >> I'd like to know more about this. >> >> (It almost sounds to good to be true...) > > what questions do you have? All I can think to say right now is > look at Catalyst::Engine::HTTP::Prefork and work from there…
I'm curious if anyone's implemented a zero downtime restart system (the likes of which FastCGI gives you for free) or if it already exists somehow. Currently we just ^C and restart, which I guess is a bit lame. The restart_graceful and pidfile options would go most of the way, presumably? Paul _______________________________________________ List: [email protected] Listinfo: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/[email protected]/ Dev site: http://dev.catalyst.perl.org/
