MegaComet testing part 2
I've been busy testing megacomet on an EC2 micro instance. Keep in mind this testing is pretty crummy, given the instance only has 600M memory and i'm running the fake-client-flooding app on the same box as the server. Anyway, i was able to get up to 50k connections open before things started getting slow. I'm not sure if i hit a configuration limit, or a memory limit, or what. Testing continues...
Anyway, here's the 'top' results, showing that 50k connections roughly means 16M memory usage:
top - 08:52:17 up 58 min, 2 users, load average: 0.13, 0.16, 0.22 Tasks: 68 total, 1 running, 67 sleeping, 0 stopped, 0 zombie Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 611212k total, 581608k used, 29604k free, 3760k buffers Swap: 0k total, 0k used, 0k free, 25356k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 22542 root 20 0 10392 2020 116 S 0.0 0.3 0:01.47 megacomet 22548 root 20 0 10400 2016 116 S 0.0 0.3 0:01.48 megacomet 22544 root 20 0 10452 2004 108 S 0.0 0.3 0:01.49 megacomet 22546 root 20 0 10392 1968 108 S 0.0 0.3 0:01.49 megacomet 22543 root 20 0 10388 1908 60 S 0.0 0.3 0:01.48 megacomet 22549 root 20 0 10396 1860 20 S 0.0 0.3 0:01.47 megacomet 22545 root 20 0 10396 1840 0 S 0.0 0.3 0:01.47 megacomet 22547 root 20 0 10392 1840 0 S 0.0 0.3 0:01.50 megacomet
And here's where i got the kernel tuning tips from:
http://www.metabrew.com/article/a-million-user-comet-application-with-mochiwe...
http://www.cs.wisc.edu/condor/condorg/linux_scalability.html
