Hello, It is important to consider: * the processor * the modules you have loaded in apache * the mpm you are using (hopefully mpm_event). * the type of load. (static, or dynamic, small files, big files, proxying, is there cache, etc?). With a more or less actual server processor, perhaps 12 cores for 300 paralell requests is overkill and maybe even counterproductive. I would start by setting 3 processes, 100 threads each and benchmark it. Once you get a figure you could try the same number of processes more threads each and the opposite, more processes less threads, and benchmark your findings, always with the same scenarios to try getting a clear picture. El vie, 12 mar 2021 a las 14:36, Nick Gearls (<nickgearls@xxxxxxxxx>) escribió: > > Hi everybody. > I have a question about the distribution of workers between processes > and threads. > > Let's say that I have a 12 core machine and that I want to server up to > 300 parallel requests. > I have the choice between, for instance, 10 processes with 30 threads or > 30 processes with 10 threads. > Am I right with the following logic: > I keep 2 cores for the OS (machine dedicated to httpd), so I have 10 > cores for httpd. > Optimal use is one process per core, so 10 processes. > I need to use all CPUs, and multiple processes per CPU will consume more > resource, correct? > > Thanks > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx > For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx > -- Daniel Ferradal HTTPD Project #httpd help at Freenode --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|