Mailing List Archive
tlug.jp Mailing List tlug archive tlug Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][tlug] Re: [RFC] Outline of the fast HTTP talk
- Date: Mon, 3 Nov 2008 23:18:07 +0900
- From: Curt Sampson <cjs@example.com>
- Subject: [tlug] Re: [RFC] Outline of the fast HTTP talk
- References: <87fxm9tfx7.fsf-genuine-vii@john.fremlin.org> <20081103093743.GB14296@lucky.cynic.net> <87zlkhrreb.fsf-genuine-vii@john.fremlin.org>
- User-agent: Mutt/1.5.17+20080114 (2008-01-14)
On 2008-11-03 11:16 +0000 (Mon), John Fremlin wrote: > The benchmark consists of running apachebench with concurrency 100 > against a page that should return an HTTP body of "<h1>Hello $name</h1>". The > "name" parameter is given as part of the GET request, and should be > correctly escaped (if it contains < characters, etc.). > > The objective of the test is to show how fast a very simple dynamic page > can be, without worrying about databases. That sounds like an excellent simple benchmark for that sort of thing. Can I just send you a tar file of a qam project that you can unpack? Ruby 1.8.whatever will be required, but you're already set up for that. Do you want me to include lighttpd source, or are you ok with installing your lighttpd 1.4.18 or 1.4.19 package from your package system? As well: we gave up on ab quite a while ago, and use httperf[1] for benchmarking. We've just started extracting a generic test project for load testing websites; I'll see if I can put together something to try with this. Httperf does have one unfortunate, and fairly major, issue: it uses select(). Given that it's oriented around the idea of maintaining a certain number of requests per second, regardless of how long it takes to receive a response, this usually results in it running out of file descriptors long before it's generated the kind of load a modern machine is capable of generating. It really needs to be using kqueue, but of course my issue here is that I have a large pile of NetBSD boxen, but (almost) all my clients have Linux boxen, which support only epoll. Perhaps for this app I could hack in some #ifdefs that would do the job, or maybe I just fall back to poll() and live with the performance hit. Who knows. But anything has to be better than select(), so if anybody's interested in working on this.... [1] http://www.hpl.hp.com/research/linux/httperf/ cjs -- Curt Sampson <cjs@example.com> +81 90 7737 2974 Mobile sites and software consulting: http://www.starling-software.com
- Follow-Ups:
- [tlug] Re: [RFC] Outline of the fast HTTP talk
- From: John Fremlin
- Re: [tlug] Re: [RFC] Outline of the fast HTTP talk
- From: Edward Middleton
- References:
- [tlug] [RFC] Outline of the fast HTTP talk
- From: John Fremlin
- Re: [tlug] [RFC] Outline of the fast HTTP talk
- From: Curt Sampson
- [tlug] Re: [RFC] Outline of the fast HTTP talk
- From: John Fremlin
Home | Main Index | Thread Index
- Prev by Date: [tlug] Re: [RFC] Outline of the fast HTTP talk
- Next by Date: [tlug] Re: [RFC] Outline of the fast HTTP talk
- Previous by thread: [tlug] Re: [RFC] Outline of the fast HTTP talk
- Next by thread: [tlug] Re: [RFC] Outline of the fast HTTP talk
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links