It looks like the new server setup is settling in rather well... All systems are go, and it's handling the increased traffic without a glitch...
I forced some overloads on the server yesterday, and my FQguardian script stepped in and handled the situation rather quickly... I had intentionally driven our load from the average (1.50) to load 150 in under 3 seconds... 2 weeks ago - this would have brought the server to a screeching halt, but this time the FQguardian locked on and communicates to some special routines I wrote into the Linux kernel and it clamped everything to 165 load, systematically and gracefully began to shutdown the overloaded Apache daemon's... Though this takes every domain offline, until the server falls back down to load 4, it avoids having to reboot the server and usually has everything back up in 4 minutes or less...
The FQguardian and Linux kernel hacks, will handle about 85% of all the overload problems we were having... There are some types that just can't be trapped or handled gracefully... It's a tradeoff to providing a very capable and powerful server by not being too restrictive in it's setup... It's a difficult decision as there are different camps - one wanting stability/uptime and others wanting power, flexibility, and capability... I try to satisfy both, but will always lean towards the power/flexibility/capability side... It's what FutureQuest was designed for from the beginning, and capability is what we are all about...
--
Andrew Gillespie
--Doing the best I can, with what I have--
FutureQuest.net
Not that I fully understand them, but I know less hops=better and less milliseconds=better 
Linear Mode
