My understanding is that Apache (and hence Zend Server) should automatically send requests to all available cores. However, on our client's virtualized CentOS server, only one core is fielding requests.
Our web app gets spikes of usage during certain seasonal periods, and we saw 90-100% usage of the first CPU and only tiny blips on a few others which I believe were not httpd's.
Running cat /proc/cpuinfo shows 8 cores with similar info. Running htop during the spike showed cpu0 at 90-100%. MySQL has innodb_thread_concurrency at 32.
I'm not a server guy, but I've done a bunch of research and I can't seem to find any clues as to how to make the server use all 8 cores.
Happy to provide more info upon request.