LinuxGSM autorestarts failed servers when the periodic monitor detects that they're down. Fine and dandy, but potentially too slow. Many players who attend a failed server will not remain unless it restarts quickly. Also, for some games and in some particular situations, if the server restarts quickly enough, the players may not even notice that it went down at all.
The LinuxGSM wrapper script which starts the dedicated server process should be able to instantly detect when that process fails, and proceed to restart it immediately. In order to prevent the situation when the script attempts to restart a completely failed server ad infinitum, a counter should be in place which stops more than X "instant restart" attempts. The monitor portion of LinuxGSM takes it from there, for future restart attempts.
LinuxGSM autorestarts failed servers when the periodic monitor detects that they're down. Fine and dandy, but potentially too slow. Many players who attend a failed server will not remain unless it restarts quickly. Also, for some games and in some particular situations, if the server restarts quickly enough, the players may not even notice that it went down at all.
The LinuxGSM wrapper script which starts the dedicated server process should be able to instantly detect when that process fails, and proceed to restart it immediately. In order to prevent the situation when the script attempts to restart a completely failed server ad infinitum, a counter should be in place which stops more than X "instant restart" attempts. The monitor portion of LinuxGSM takes it from there, for future restart attempts.