I've had this dedicated for about 2 months now with hardly any problem. All of the sudden, yesterday morning the server is 'clogged'. It isn't down, its pingable...and I can login via ssh, but it takes about 15 minutes just to get to a prompt. The server has basically eaten all of it's own resources and has slowed to a halt.
If I reboot, the problem is gone, but it begins again, with the process count slowly sliding up until around 1000 processes, i really cant do anything else. This only takes about 10 mins. I've seen the process count go as high as 2200+. If instead after the reboot, I run in and shut down apache, then restart, i dont have this problem. Everything comes back fine and settles in around 150-170 processes (normal). I'll go have a meal and all of the suddent a phone call 10 hrs later or something and it has happened again. Its happened 3 times in the last 48 hrs. Most of the processes are httpd and mysqld, and almost all of them are sleeping.
I had maxspareservers at 100, so I dont understand how there can be 500 sleeping httpd processes (i think that is what i am seeing). I also had mysql max servers at 500, but I cut both of these numbers in half last time this happened, and set a max total apache servers at 800. No new software has been installed, it is a single web site that I run, heavy traffic and heavily dynamic content though. Apparently rackshack is looking at it, but Id like to see if anyone here has any theories?