Jump to content

Home

Server Crash prob


rickard64

Recommended Posts

Posted

it seems like always one of my two servers will crash when im not at my computer, couz everytime i leave my computer and come back one of my two dedicated servers has crashed??!?

 

 

im running win xp on a amd 1200 w 384mb ram

 

:mob:

Posted

Same problem here, either it crashes or shutsdown with "opStack corrupted in compiled code" or something.

Posted

Had the same problem, solved it by adding :

seta sv_master1 "masterjk2.ravensoft.com"

at jk2mpconfig.cfg file. Give it a try.

Posted
Originally posted by Leo

Had the same problem, solved it by adding :

seta sv_master1 "masterjk2.ravensoft.com"

at jk2mpconfig.cfg file. Give it a try.

 

I've had that in my jk2mpconfig.cfg file all along. My server tends to just crash without warning or without message. The application just quits unpredictably.

 

I'm running Win2k AMD1.3 512MB.

Posted
Originally posted by The Austrian

 

I've had that in my jk2mpconfig.cfg file all along. My server tends to just crash without warning or without message. The application just quits unpredictably.

 

I'm running Win2k AMD1.3 512MB.

 

I configured/keep up with the Cheater's Wayside server... and the damn thing will sometimes last all day and sometimes itll last two minutes before the server proggy dies.

 

RTCW runs for weeks without a problem :|.

Posted

Wow glad I'm not the only person who is having this problem. Just think it ran all day today with no problems now I can get it to run for more than 5 minutes.

 

System specs

 

p4 1.7

1GB of ram

Win2k

 

Hopefully someone comes up with a solution or Raven fixes a bug.

 

Awol

Posted

I ran a linux server all weekend. And ever restarted the server once today without a problem. Now I seg fault ever 2 mins of the game starting. Seg fault Seg fault Seg fault. WHy today? nothing has changed. I even completely reinstalled. Anybody have any Idea?

Posted

Same here. I posted in a previous thread that it may be some sort of transport layer problem. Seems like mine would always segfault after the master server heartbeat.

 

Other times, it would core right after it started the maps and was waiting for a client to connect.

 

I'm running tcpdump to see if there are any weird packets associated with the segfaulting, but has anyone else done this (I don't want to reinvent the wheel if so.. )

Posted
Originally posted by tard99

I ran a linux server all weekend. And ever restarted the server once today without a problem. Now I seg fault ever 2 mins of the game starting. Seg fault Seg fault Seg fault. WHy today? nothing has changed. I even completely reinstalled. Anybody have any Idea?

 

my linux jk2 servers started doing this today too... they've been running fine for the past couple days... just about an hour or so ago (4pm PST), the jk2ded process would segfault a few seconds after starting the server and it continued for a few minutes... they seem to be running fine now...

 

 

/stem

Posted

When it's happening, you might want to verify that you can ping the master server. I suspect that either there's problems with the master server (or at least your route to the master server) and the app isn't setup to handle such a problem "gracefully". I do remember seeing a post elsewhere that there had been problems with the master server and they had to restart it.

 

Hope that helps explain it, even if it doesn't fix it. :(

Posted

Yup happend to CoolNet as well... Ran 4 days flawlessly (accept eratic ping rates and eating up all the bandwith) then all of a sudden tonight crashed about every 30 sec. The only way i got it back up was to start it from within the game. So what your telling us is if the master server goes down then it crashes our games?? never had a game do that before they normally just give an error till they reconnect to the master. If what you say is true all admins are going to have to babbysit the servers to keep them alive and thats not good.

 

Originally posted by MatrixCPA

When it's happening, you might want to verify that you can ping the master server. I suspect that either there's problems with the master server (or at least your route to the master server) and the app isn't setup to handle such a problem "gracefully". I do remember seeing a post elsewhere that there had been problems with the master server and they had to restart it.

 

Hope that helps explain it, even if it doesn't fix it. :(

Posted

I can't say that is the cause with certainty. However, it can be validated if you can catch it while it is happening and test for a response from the master server. I admit that the server software should be able to handle the "exception" that occurs when the master server doesn't respond...but it could just be a bug. It could be something else, but the fact it occurs during periods of time and for a number of people right when the heartbeat is being sent makes me suspect the master server.

Posted

(All times here are Eastern US Standard Time, synched with NTP)

 

I have been running my monitor-and-restart script every minute, and the server has been quite stable, until 6:55PM, at which time the server began crashing:

 

Restarting server Mon Apr 1 18:56:00 EST 2002

Restarting server Mon Apr 1 18:58:00 EST 2002

Restarting server Mon Apr 1 19:01:00 EST 2002

Restarting server Mon Apr 1 19:01:59 EST 2002

Restarting server Mon Apr 1 19:03:00 EST 2002

Restarting server Mon Apr 1 19:04:00 EST 2002

Restarting server Mon Apr 1 19:06:00 EST 2002

Restarting server Mon Apr 1 19:08:00 EST 2002

Restarting server Mon Apr 1 19:09:00 EST 2002

Restarting server Mon Apr 1 19:11:00 EST 2002

Restarting server Mon Apr 1 19:12:00 EST 2002

Restarting server Mon Apr 1 19:13:00 EST 2002

Restarting server Mon Apr 1 19:14:00 EST 2002

Restarting server Mon Apr 1 19:15:01 EST 2002

Restarting server Mon Apr 1 19:15:59 EST 2002

Restarting server Mon Apr 1 19:18:00 EST 2002

Restarting server Mon Apr 1 19:20:00 EST 2002

Restarting server Mon Apr 1 19:21:01 EST 2002

Restarting server Mon Apr 1 19:22:00 EST 2002

Restarting server Mon Apr 1 19:24:01 EST 2002

Restarting server Mon Apr 1 19:28:00 EST 2002

Restarting server Mon Apr 1 19:29:00 EST 2002

Restarting server Mon Apr 1 19:31:00 EST 2002

Restarting server Mon Apr 1 19:32:00 EST 2002

Restarting server Mon Apr 1 19:33:00 EST 2002

Restarting server Mon Apr 1 19:34:00 EST 2002

Restarting server Mon Apr 1 19:35:00 EST 2002

Restarting server Mon Apr 1 19:36:01 EST 2002

Restarting server Mon Apr 1 19:38:00 EST 2002

Restarting server Mon Apr 1 19:40:01 EST 2002

 

From 6:56PM until 7:40PM, the server crashed a LOT, but then suddenly began running smoothly again.

 

 

Can anyone say if the master server was having problems during that time? Because I am starting to suspect perhaps a malicious user has a way to segfault the server.

 

44 minutes is the amount of total time I had trouble in 24-hours, and that's about the same as the average attention span of a normal 15-year-old script kiddie, in my experience.

 

Before reading this new information about the master server, I was nearly convinced that someone is shutting down the server by using a malicious exploit. So, at least the server keeps rebounding. Must suck for the users who keep getting bounced.

 

I don't have IP logging in place yet, I'd love to see if the same IP logged in just prior to each crash.

 

r00d

Posted

There is an exploit for the Dedicated server that dates back to Quake 3 Point Release 1.29g (and was later fixed in 1.3x)... The servers can be taken down at will.

 

People at ID software are/were aware of it at the time and fixed it. These servers can be very stable, if this exploit was fixed.

 

If anyone needs more info (that anyone being someone who can help fix the problem), drop me a line and I will give you everything I have related to it.

 

Thanks,

tubesock

http://www.quake****.com

Posted

Yep, I just came across that same exe myself and tested it out...crashes the server every darn time. Quite a shame too. I'm not putting my server back up until a patch comes out....sorry folks. You'd think that it wouldn't take much sense to update the point release being used as the basis for the game to a version where this exploit didn't exist. *sigh*

 

I'll keep helpin' out. But this is not good news.

 

It's a shame that JO doesn't have fixed identification similar to guid's (based on cd-keys). At least then we could track down and perm ban the slime. They could get around it, but over time it would make it a lot more difficult than just sitting around waiting for someone with a bad attitude to see how many servers they can bring down in 20 minutes.

Posted
Originally posted by tubesock

There is an exploit for the Dedicated server that dates back to Quake 3 Point Release 1.29g (and was later fixed in 1.3x)... The servers can be taken down at will.

 

People at ID software are/were aware of it at the time and fixed it. These servers can be very stable, if this exploit was fixed.

 

If anyone needs more info (that anyone being someone who can help fix the problem), drop me a line and I will give you everything I have related to it.

 

Thanks,

tubesock

http://www.quake****.com

prove that this is causing the problems. otherwise its just hot air. sure maybe someone can crash one server. but EVERYONe is having problems. It also doesnt explain why the server is sometimes stable for hours and then crashes every few minutes and is then totally stable again.
Posted

Also, I've been running this engine since q3 test in early 1999, so I am quite familiar with ALL the in's and out's of the server itself. Also have been involved in quite a few MOD's and helped alot of admins.

 

I'm just trying to help...

 

-Tube

http://www.planetqs.com

Posted

This was a pretty well known problem with point releases 1.29f/g. Here is a portion of the readme that comes with the app:

 

-----------

:: Summary

-----------

 

There exists a very large hole in Quake 3

Arena, version 1.29f and 1.29g (the latest,

1.29g which got released just under a week

ago).

 

The hole is not fixable in any way by

the user, and most of the servers that

are up (thousands of them) are vulnerable.

To have this hole fixed, a PR (point

release) will have to be given to the

public by iD Software.

 

----------

:: Details

----------

 

As a result of a previous Q30wnerz-discovered

vulnerability, iD Software had to redesign the

protocol, closing up the previous vulnerability.

 

However, we have discovered a new one which

segment faults the servers cleanly (it gives back

the memory it had taken before, which is a lot

since Quake 3 is a memory hog). If the server

is logging, it will segment fault before it has

a chance to append it to the log file.

 

This is 100% real and I fully expect (or rather hope) that we'll be seeing a patch SOON.

 

Here is what you'll see in the console after it happens:

 

startup.sh: line 2: 4557 Segmentation fault ./jk2ded +set dedicated 2 +set vm_game 2 +set ttycon 0 +set com_hunkMegs 128 +set com_zoneMegs 24 +exec server.cfg

 

Now, startup.sh is my start-up script and everything after the seg fault is my command line (in the script). It took me all of about 4 seconds to do this. If you'd like, just get hold of me via icq (74850239), aim (MatrixCPA), or MSN Messenger (matrixcpa@msn.com) and I'll "test" your server while you watch the console.

Posted

My server used to crash all the time too...

 

So then I did a virus scan and found the nimda virus on the box. :eek:

 

I cleaned it all up, and it works fine now.

Posted

hmm if there is a proggie out there that makes the server crash....

 

is there any point in having an server then????

 

bah i hater that kind of ppl.... just out there to **** with us

Posted

i call bs. people dont just go around crashing servers without a reason.

 

now if ou kicked someone off your server or something happened like that. ok ill buy that. but randomly crashing jk2 servers "just because" doesnt make sense.

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...