Open games keep dropping

Seems to be just open games. ideas?

There is a bug in the server. I havn't located it yet. Seems to be related to starting a strict game on minbot. Sometimes, it drops randomly, and it doesn't always drop when a new strict is started. So, there seems to be some kind of bug.

Min

Has anything changed recently?

Not really. Other than rat and others starting chew toy 3 every 5 minutes or so. This problem has been an issue since I created minbot, so, before 1.17 even came out over a year ago. I believe its due to a memory leak in the server.

Min

Min wrote:

Not really. Other than rat and others starting chew toy 3 every 5 minutes or so.
Min

that shit is fuckn annoying...specialy when your haveing fun in a open game.. stupid fuckn morons...

annoying yes, stupid fuckin morons for mapping a strict game? no.

could it be due to the ports being reused while one is still active?

is the code for the game spawner open source?

Strict > Open

underdog wrote:

could it be due to the ports being reused while one is still active?

is the code for the game spawner open source?

........ seriously? no its not due to the ports being reused while one is still active. The open game spawner uses a port that the strict games don't even go into.

The game spawner is just a php script that starts a game. Its not complicated and certainly not the source of this issue. The issue is the server crashes. It doesn't crash every time, just sometimes, usually when an open game has lasted long enough. This leads me to believe that the winbolo server is accessing memory that hasn't been properly allocated for it. And when the new game fires up, it starts using that memory. Once that occurs, the server hits some 'bad' memory and crashes. The solution is simple, figure out the bug and fix it. Feel free to attempt to figure it out UD.

I noticed this issue a very very long time ago, I tried to figure it out, I was unable to. You guys just didn't notice until now I guess.

Here is the spawner code UD, as you can see, it is extremely unlikely to be the cause of this issue.

[code:1:2e12abb3c0]while(true) { //start the spawning
$map = get_rand_map();
$loggingfile = "\"".$map. "-". time() . "-" . $port . "-debuglog.txt\"";
echo $map."\n";
echo $loggingfile."\n";
$exec = $path ."linbolods -map \"" . $path . "spawnermaps/" . $map . "\" -port $port -gametype $gametype -mines $mines -ai $ai -tracker $tracker -autoclose -log -logfile $loggingfile";
//echo $exec;
exec($exec);
}
[/code:1:2e12abb3c0]

Min

If the spawner code did not change, it can be removed as a a possible cause.

Is there a way to verify or at least more likely a memory leak? E.g. restart the server and try reproducing the problem. If it takes a reasoanble time for the problem to start occurring again could it be concluded that it is probably a memory leak?

ErYan wrote:

If the spawner code did not change, it can be removed as a a possible cause.

What part of 'nothing has changed, its been dropping for over a year you guys just never noticed' is difficult to understand exactly? Basically, its been crashing this way since I started running a open game spawner. You can't 'remove' the spawner because the code did not change. I noticed this crash when I was playing an open game, and someone started a strict. I spent some time replicating the issue, I know when it crashes. I already explained that. But I guess my post must have been too long and you didn't bother reading it.

ErYan wrote:

Is there a way to verify or at least more likely a memory leak? E.g. restart the server and try reproducing the problem. If it takes a reasoanble time for the problem to start occurring again could it be concluded that it is probably a memory leak?

heh, it doesn't work like that, how long the server has been turned on is irrelevant. You guys are trying to be 'helpful' but are mostly just annoying honestly. Open the code, have a look yourself, quit trying to give advice on how to fix it. You don't even know the correct kind of advice to give. I have a few suspicions on what the problem could be, but they aren't things I care to investigate much due to how difficult it is to properly troubleshoot this stuff without the RSA keys.

Min