> Hrmm... maybe it was Apache that was croaking under the load. I probably > need to tweak Apache further. I'll be doing the usual research (Google > and the Apache website, primarily), but would anyone have concrete tips > to help me along? I'm using Apache 2.0.54 using mpm-prefork. >
Prefork is the apache 1.3 way of doing things, which is a memory hog. If your are using 2.6 or/with nptl you can reap bigtime by using the Worker mpm. Checkout FastCGI and SCGI too. > The first optimization that comes to mind is putting a caching HTTP > accelerator (eg: Squid) in front of Apache. This is probably desired, > since behind Apache I also have Zope, which if subjected to the Slashdot > effect would surely die. > And then there's lighttpd[1] that, perhaps providentially, I stumbled on > last night, an hour or so before the Slashdot effect took effect (by > which time I was in bed). I'm already very comfortable with Apache, > though, so maybe tuning that might be a better route. Still, I'd love to > hear from anyone using lighttpd in production. > > Anyone? Thanks in advance! > > [1] http://www.lighttpd.net/ Lighttpd trounces apache and it supports FCGI and SCGI, what a treat. If you want to serve em static, there's gatling, fnord and thttpd to the rescue. Gatling uses OS-specific performance and scalability APIs, (FastCGI support pending) while fnord(forking!) and thttpd(select()) are quite similar with mmap and sendfile utilization. --ed
_________________________________________________ Philippine Linux Users' Group (PLUG) Mailing List [email protected] (#PLUG @ irc.free.net.ph) Read the Guidelines: http://linux.org.ph/lists Searchable Archives: http://archives.free.net.ph
