yeah like, i'm fine with dishing out more money, and the complaints about ram are only a minor problem bc i can afford it. also wine has that weird problem with scaling at a certain playercount, has happened on new years and any other linux server. i dont wanna risk rubber banding again
plus i'd probably host something else after, so w/e
how well is well though? i mean we're pushing blockland to it's absolute limit here and if Linux with wine is not up to par with windows (which I've heard multiple times) then it doesn't seem like we should be using it
what weird problem with scaling? the Wine project is getting better, day by day..
and, extremely well. stable, under high load, and, if you need ultimate stability, then run a stability test. get 100+ players (simulated, or real. doesn’t matter.). attach a debugger to it. wait for a crash, then, send the memory address the exception happened at, as well as the backtrace, and the full dump of the memory at the time of the crash. that would help you determine a counter measure for the crash, develop against it, and have better stability. converting the player movement code to being less server authoritative for one would help decrease overhead at the expense of hackability- but, with only ~5 people in the community who actually know how to exploit the move stuff to their desires, that isn’t much of a concern. updating the net code to allow for higher bandwidth streams- such as in cases like downloading files from the server would also help. offloading all the computationally tough things and optimizing it all- by either having it all written in C++ or something, also could help. whatever your problems are- they mostly lie in the fact that torquescript is slow.
Poll