[Gigabit Nick]

We run two servers (one zombiemod and one cssdm) on the same machine under the same account, and both under ServerChecker.

The zombiemod server generates mdmps when it falls over, and the cssdm server doesn't when we encounter this problem.

The only differences in command line options are they're on different ports (using +port) that the cssdm one is a 75 tick server running de_dust2 as the first map, and the zombiemod one is a 33 tick server running zm_ruined_town_v3.

We don't have any specific configuration options set in the config files for debugging or logging, so I'm not sure if there is anything we can tune to get a mdmp file.

[TSCDan]

Are you guys who are having crash issues running mani + bots?

http://www.mani-admin-plugin.com/ind...6&topic=3659.0

I'm having the crashing issue as well, although I did not go in depth troubleshooting it; I just have it staying on dust2 for now...

[BAILOPAN]

Here is another test build. If this fails, please set "body_stay_time" to 0 in your cssdm.cfg and see if that fixes it. If it does, please set it back and continue to use these debug builds, as I want to actually fix the problem.

[Gigabit Nick]

The level switch failed, but the server did not appear to crash. The second level started but none of the players were entered into the player list (i.e. pressing tab showed no names, including mine when I was in as a player).

We're switching the config flag you mentioned now.

[BAILOPAN]

If you switch the config flag, try it with the original 0.96c binary.

[kunk]

Originally Posted by BAILOPAN If you switch the config flag, try it with the original 0.96c binary.

I'm back on .96c with body_stay_time set to 0. So far so good. I've gone through 3 maps so far with no crash. Will play a few more maps to see how it goes.

*update* Played through about 6 maps now and no crash yet so its looking like the issue might be with the body_stay_time

[BAILOPAN]

Yep. Basically what happens is, the timing of ragdoll deletion gets mixed up and on mapchange the server crashes. If the ragdolls are deleted right away instead of timed, there's no problem.

Sadly, I haven't quite nailed the fix yet since I can't reproduce it. Did 0.96d crash for you, kunk?

EDIT: Gigabit Nick: Any way I could have access to your server to test fixes? Unfortunately, I can't reproduce this myself, and testing one patch per day isn't going to be effective. I need RCON+FTP to changelevel and upload.

[darkdata]

i can give you ftp access to mine and rcon its stays full if you want it pm me your ip because we have our rcon and ftp ip restricted...

Originally Posted by BAILOPAN Yep. Basically what happens is, the timing of ragdoll deletion gets mixed up and on mapchange the server crashes. If the ragdolls are deleted right away instead of timed, there's no problem. Sadly, I haven't quite nailed the fix yet since I can't reproduce it. Did 0.96d crash for you, kunk? EDIT: Gigabit Nick: Any way I could have access to your server to test fixes? Unfortunately, I can't reproduce this myself, and testing one patch per day isn't going to be effective. I need RCON+FTP to changelevel and upload.

[BAILOPAN]

darkdata: does your server still crash with the latest binary I posted? (0.96d, 6 posts above)

[darkdata]

it hasnt yet looking good so far

Originally Posted by BAILOPAN darkdata: does your server still crash with the latest binary I posted? (0.96d, 6 posts above)