[Drunken F00l]

I agree that server stability has been an issue. I still haven't been able to conclude whether it's a problem with my plugin or srcds in general. The core dumps almost always point to things internal to srcds/tf2 such as engineer objects, physics, or model/bone manipulation. The remainder of the crashes end up having odd stack traces which could indicate stack corruption and a mistake on my part.

So, it remains unclear if it's something wrong with my plugin/configuration or if it's a problem with TF2.

Valve maps #2 seems to have much better stability this week. Proof:
http://sourceop.hlstatsx.com/?mode=serve...=4&game=tf

Goldrush is better, but still has issues:
http://sourceop.hlstatsx.com/?mode=serve...=5&game=tf

I'm thinking that since goldrush has 1 day map cycles, that could be an issue. None of the other servers, including the 24/7 2fort server waits a day before refreshing the map. If reducing the rotation to 2-3hours improves stability, it's probably safe to say that this is a main factor in the stability issues.

Valve maps #2 stability increased with the automatic ban for fast class switching. I think exploits/malicious players may have been a factor in its previous instability, but I'm uncertain.

24/7 payload is something I'm willing to try out.