Exploding turrets crashing the server

was a ball

Now that the new version (2.17) is up, does anyone know if this is fixed now?

Tonight there were some new players that were telling others to buy eject before getting into the turrets.
It sounded like they thought the issue was that they would loose all of their experience.

Now that Dru and Shan are back, maybe we can get some guidance on how to treat this issue. Can we consider this fixed until the server actually crashes? Or will there be a time set aside for testing?

I think Szlat said that he has made some changes in an attempt to reduce the number of crashes by reducing damage in a mini turret and also to eject the player at any time even if they dont have eject, but its not a positive guarantee that this will fix it completely, I would still not allow players without eject to get into a turret just to be on the safe side.

If we let people try it, then Druid will be able to find what crash message is actually logged, if it still crashes, and we might be able to fix it properly.

Server crashed 16jan2010 at 00:55 central time.
Minigun turret exploded.
It was re-done's turret, but a low level player (non-engineer) was in it.
I think the player name was wrathy

yep. thats what happened, however i have come with a clever explanation...

we were in the map parish my turret was below some ark, so when the turret explode it did turn on the eject button, however wrathyhit the top of the ark and the turret explotion killed him... then server crashed..
well could be that, or could be that he hasnt eject botton... still, can the minigun tower be coded so no one without eject botton come in?
btw it was a pitty that road has lost 2.5k exp points..

Szlat, in 217 it seemed the bug was fixed, but players appeared to be sitting on the floor instead of the turret.
Now the sitting issue is fixed, but we have the original problem again.
What code caused the changes in the player model location?

Of course, it is possible that the server crash issue was never fixed, but I am pretty sure we tested it extensively.

greg11 wrote:

Szlat, in 217 it seemed the bug was fixed, but players appeared to be sitting on the floor instead of the turret.
Now the sitting issue is fixed, but we have the original problem again.
What code caused the changes in the player model location?

For 217 I set the turret to be a remote control turret. Because the system assumed you were not in the turret, the player never suffered damage and always survived at the end of it. And it didn't seem to crash much - although I can't for definite say it didn't.

However, because it was remote controlled, it didn't support the drawing of the player in the turret very well.

Szlat wrote:

greg11 wrote:

Szlat, in 217 it seemed the bug was fixed, but players appeared to be sitting on the floor instead of the turret.
Now the sitting issue is fixed, but we have the original problem again.
What code caused the changes in the player model location?
For 217 I set the turret to be a remote control turret. Because the system assumed you were not in the turret, the player never suffered damage and always survived at the end of it. And it didn't seem to crash much - although I can't for definite say it didn't.

However, because it was remote controlled, it didn't support the drawing of the player in the turret very well.

Have you ever been able to replicate the crash on your server?
I have a suspicion that UT servers can crash if the logfile gets too big.
I never bothered to verify this. Instead when I ran my server, I set it to restart any time there was a period of time with no player activity. I rarely had the server crash with players on it.
I assume you only run your test server when needed, and my not have the run time to recreate the crash. Can you see if any log entries are generated when the turret explodes.
I haven't updated my server to support engineers yet. I have been tied up with other projects.

Also is there a way to get bots to use turrets?

greg11 wrote:

Have you ever been able to replicate the crash on your server?
........
I assume you only run your test server when needed, and my not have the run time to recreate the crash. Can you see if any log entries are generated when the turret explodes.

I haven't managed to duplicate. However, as you say, the server is by no means as extensively used as DC - only me and bots on it.

greg11 wrote:

Also is there a way to get bots to use turrets?

Not that I know of.

Szlat wrote:

greg11 wrote:

Also is there a way to get bots to use turrets?
Not that I know of.

there is : get in a turret and then order one of them to "hold this Position" they don't shoot much in a ball turret but function fairly well in a mini

greg11 wrote:

Server crashed 16jan2010 at 00:55 central time.
Minigun turret exploded.
It was re-done's turret, but a low level player (non-engineer) was in it.
I think the player name was wrathy

Was the map DM-Parish?

yes dm-parish

Druid posted me the logs. The crash at 23:53 Mountain time was due to the DCMonsterController getting a runaway loop on FindNewEnemy.
So, I don't think it is anything to do with the Minigun turret.

Szlat wrote:

Druid posted me the logs. The crash at 23:53 Mountain time was due to the DCMonsterController getting a runaway loop on FindNewEnemy.
So, I don't think it is anything to do with the Minigun turret.

Yay! Looks like you got some good info from those logs.
I wonder if Druid would be up for making future logs available on the site.
The sensitive information could be stripped out, but instant access to errors and warnings would assist with taking care of these bugs.

What if the minigun issue was originally caused by the same thing.