On Tue, Sep 15, 2009 at 10:43:26AM +0200, Mic J wrote: > So since benchmarking is out, how do we then find out where > potential problems are. > What does OpenBSD developers do, since surely they don't benchmark :) > > Maybe we should profile instead ? > > I'm not very experienced with webservers, but here > how i would approach it. > > 1. i have a problem, i think about it where/what the problem could be > 2 i check the logs - test my equipment > 3. I create 1 or a few profiling tests / micro benchmarks to test my > assumptions > or make certain i haven't misinterpreted my problems.
I have *never* seen a micro benchmark yield anything of value. Ever! While one is writing some code one tends to avoid stupid constructs. > 4. Step back and interpret results > 5. think of other tests / micro benchmarks that could further enlighten me > and confirm/unconfirm(?) my "findings" You find me a test and I'll find you a bunch of corner cases. Been there, done that, got the t-shirt. > What i wouldnt do, is "design" a mother of a benchmark that covers > all the bases. > It's to hard to get right. It would take to much time. Does not exist. This is a bunch of malarkey. This is the same as measuring the position _AND_ momentum. If you don't get this then you need to write more OS code... that runs On 16 or so architectures... > How would OpenBSD dev's approach a issue. > How are issues generally searched for/ found out? > > I imagine something like > OpenBSD dev works on the httpd daemon - asks for testing. > I find a problem, ex: it'd slow like heck - check configuration - For various degrees of heck? Large quantities of 1? > interfaces - logs > What now - i write back to dev. > dev asks me to do what? > What does the dev do behind the scenes? Profiling under a "reasonable" load and that is where all benchmarks fall apart. My definition of a "reasonable" load is different from everyone else's.