Seeing some serious weirdness with the L4D2 SRCDS in forked mode. I'm running several hundred L4D2 forks across multiple CentOS 5 machines. Several times now I've noticed all forks on a server dying, and I think I've figured out why.
In L4D, you couldn't issue any commands on the forked server console - you only saw output. You can telnet to the base port and broadcast commands to all servers though. In L4D2 can you actually type commands on the server console. You don't get character echo, but if you type status <enter> it works just like you were telnetting in. I just noticed all the locked up servers appeared to have some invisible text waiting on the console that hadn't been <enter>'d . It looks like the server processes all stop until any input on the console is completely entered. In my case, I'd left an 'a' on the console when switching screens. All forks stopped, I went back to the console and pressed enter a couple of times . It told me 'a' was an unknown command then all the forks resumed executing. I've tested this multiple times on a very high traffic server. Processing stops the instant I type any text and resumes as soon as I press enter. If you hold off pressing enter long enough the forks will die and players will drop off. #28:SLEÃGE: ridin with the tank #13:*Harbost: omg i hate being capped <- Here is where I typed asdasdasdasd on console -> [S_API FAIL] SteamAPI_Init() failed; unable to update local steamclient. Continuing with current version anyway. Adding master server 72.165.61.151:27012 Adding master server 68.142.72.250:27011 [S_API FAIL] SteamAPI_Init() failed; unable to update local steamclient. Continuing with current version anyway. Adding master server 72.165.61.151:27013 Adding master server 72.165.61.136:27014 [S_API FAIL] SteamAPI_Init() failed; unable to update local steamclient. Continuing with current version anyway. [S_API FAIL] SteamAPI_Init() failed; unable to update local steamclient. Continuing with current version anyway. Adding master server 69.28.140.246:27011 Adding master server 72.165.61.136:27011 Adding master server 72.165.61.136:27014 Adding master server 72.165.61.136:27013 <here is where I pressed enter nearly a minute later> Bad command asdasdasdasd args Child 7 exited with status 0 Child 18 exited with status 0 Child 27 exited with status 0 Child 14 exited with status 0 Those 4 forks died and restarted. Console output of players chatting indicated threads which were still up had glitched and had issues too. Oh and this works if you just press backspace on any empty console input buffer too. Thread execution will pause until you press enter. Anyone else seen this happening? All our OS installs are nearly identical - only minor differences in kernel version, so it may be CentOS related. -wob _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
