Server Dedicated Server Crush bug
- Thread starter Z-MAN
- Start date
Seconding this.
Please revert the relevant code in MonsterController, ZombieBoss and whatever else that came in with the Heavy Metal update related to this issue.
> This issue alone is having a significant negative effect on our player numbers and others here have commented so. <
Pretty worrying that this issue didn't come up in your testing (you _did_ test it thoroughly right?
) since the crashes are very frequent for many people who have commented here and these kind of regressive bugs are what thorough testing is very effective at guarding against...
Last edited:
Upvote
0
Quite, it's been nearly 2 weeks which is far too long for your loyal server admins to go without feedback especially for an issue like this...
I'm afraid to say the handling of this issue and others give a very bleak impression of TWI's attitude and long term commitment to its volunteer server admins who provide the gaming experience full stop.
For instance:
Our community is about to roll out
I'm afraid to say the handling of this issue and others give a very bleak impression of TWI's attitude and long term commitment to its volunteer server admins who provide the gaming experience full stop.
For instance:
Our community is about to roll out
Last edited:
Upvote
0
Upvote
0
Well, just for the record, I haven't had a crash in 10 days since the beta patch was released. I tried the Patriarch crash tests by staying above ground and in difficult to see/reach places but couldn't crash the server.
Upvote
0
![[TW]Ramm-Jaeger](/data/avatars/s/0/2.jpg?1568638684){kind=avatar}
We'll get the Linux version of this fix up ASAP. This slipped under the radar due to many devs being out of the office for a couple of weeks.
Regarding if we test this stuff, we do test this stuff pretty extensively before we release it. We never encountered this crash in any of our testing before release, and we still have not been able to replicate it since release. We've played at least 75 hours on our official public server since the release as we've been judging 57 contest map entries. We played most of these maps through to the Patriarch and never experienced this crash (this server actually wasn't updated with the crash fix). I'm not saying it doesn't exist, as the crash logs clearly show that it does. What I am saying is that even with a couple of weeks of bashing on a server post release, it still never happened for us. So the point is we're not a bunch of knuckleheads rolling stuff out the door willy/nilly.
We test things as best as we can but things will always be found that we didn't experience when you increase your sample size from a few test servers, up to 1000+ live servers on all ranges of hardware/OSes. Our goal in these situations is to roll out fixes as quickly as possible when problems like this crop up. As I mentioned, the Linux version of this fix slipped under the radar, and we'll work to make sure this doesn't happen again.
Regarding if we test this stuff, we do test this stuff pretty extensively before we release it. We never encountered this crash in any of our testing before release, and we still have not been able to replicate it since release. We've played at least 75 hours on our official public server since the release as we've been judging 57 contest map entries. We played most of these maps through to the Patriarch and never experienced this crash (this server actually wasn't updated with the crash fix). I'm not saying it doesn't exist, as the crash logs clearly show that it does. What I am saying is that even with a couple of weeks of bashing on a server post release, it still never happened for us. So the point is we're not a bunch of knuckleheads rolling stuff out the door willy/nilly.
We test things as best as we can but things will always be found that we didn't experience when you increase your sample size from a few test servers, up to 1000+ live servers on all ranges of hardware/OSes. Our goal in these situations is to roll out fixes as quickly as possible when problems like this crop up. As I mentioned, the Linux version of this fix slipped under the radar, and we'll work to make sure this doesn't happen again.
Upvote
0
Also, one quick note, Tripwire does monitor the HLDS mailing list regularly. Many KF server admins post issue there and its a good place to get our attention really quickly of you have an issue (since we check our e-mail more often than the forums).
Upvote
0
Alright guys, the Linux version is now available behind the same beta password:
-beta kfwinserverfix1006
Please let us know if it's working well, so we can take it out from behind the password to make it the default installation on Linux.
Thanks Guys!
-beta kfwinserverfix1006
Please let us know if it's working well, so we can take it out from behind the password to make it the default installation on Linux.
Thanks Guys!
Upvote
0
well the bug is not fixed for windows so I don't know if it will do anything for linux either. as for reproducing this crash I can provide at least 2 scenarios where it can be done 100% without a doubt. as long as you or anyone on your team are in a certain position of the map, patriarch spawning will crash the server.
KF-Farm:
- inside the closet under a flight of stairs in one of the houses, players use this spot to exploit the AI all the time. if you're in the closet they don't attack unless you come out where they can see you. I've never reproduced it here myself but other have posted it happening, and I'm sure it would crash every time based on what I know about this bug.
KF-The_Rock_beta1:
- at the top of the ladder in the elevator shaft entered from the lowest level in the map, zeds do not know how to attack if they approach you from one of the levels above, but if you are low enough for them to see you from below they will climb up to get you.
kf-meow:
- anywhere inside the trader office building with the 2 weldable doors, and one spawn in the back room. zeds have trouble attacking the building sometimes if there are no players outside. this is a common place for the team to be in the patriarch round so the final wave is unplayable as of now.
basically any place you can be where the zeds do not attack unless they have line-of-sight can cause this crash. this is common in many maps with incomplete pathing, I'm sure we can find plenty more examples for you guys to test. a single player in any of these spots can trigger it, I have tested this in games where teammates were outside of them as well.
KF-Farm:
- inside the closet under a flight of stairs in one of the houses, players use this spot to exploit the AI all the time. if you're in the closet they don't attack unless you come out where they can see you. I've never reproduced it here myself but other have posted it happening, and I'm sure it would crash every time based on what I know about this bug.
KF-The_Rock_beta1:
- at the top of the ladder in the elevator shaft entered from the lowest level in the map, zeds do not know how to attack if they approach you from one of the levels above, but if you are low enough for them to see you from below they will climb up to get you.
kf-meow:
- anywhere inside the trader office building with the 2 weldable doors, and one spawn in the back room. zeds have trouble attacking the building sometimes if there are no players outside. this is a common place for the team to be in the patriarch round so the final wave is unplayable as of now.
basically any place you can be where the zeds do not attack unless they have line-of-sight can cause this crash. this is common in many maps with incomplete pathing, I'm sure we can find plenty more examples for you guys to test. a single player in any of these spots can trigger it, I have tested this in games where teammates were outside of them as well.
Upvote
0
Installed the fix on linux yesterday evening and got three crashes in 10 hours with 3 different maps!
Upvote
0
I have been very fortunate in that I have never experienced these crashes, but a bit more info might be handy, if we are to get this fault fixed. There are many variables that I would guess could cause a bug like this, so the following might also be handy, to see if there is something in common
For example, my servers run:-
Round limit:- 7
Map that crashed:- none
Level that crashed:- Level 7
OS: Linux
Number of players:- 6
Server: Dedicated
Difficulty:- Hard
For example, my servers run:-
Round limit:- 7
Map that crashed:- none
Level that crashed:- Level 7
OS: Linux
Number of players:- 6
Server: Dedicated
Difficulty:- Hard
Upvote
0
Could you provide a bit more information. Server logs would be great, let us know what maps it happened on, etc. There may well be more than one crash, because the fix worked for people like Arramus.
Upvote
0
Here is the part from the serverlog:
NetComeGo: Close TcpipConnection 81.110.18.85:1181 Sun Sep 13 20:21:04 2009
NetComeGo: Open myLevel Sun Sep 13 20:21:04 2009 81.110.18.85:1181
NetComeGo: Close TcpipConnection 80.192.27.112:2140 Sun Sep 13 20:21:05 2009
NetComeGo: Close TcpipConnection 87.248.4.121:2244 Sun Sep 13 20:21:07 2009
NetComeGo: Close TcpipConnection 80.192.27.112:2165 Sun Sep 13 20:21:19 2009
Log: PlayAnim: Sequence 'DeathB' not found for mesh 'british_riot_police_I'
Warning: KFPlayerController KF-DegenerationMultiplayerV1.KFPlayerController (Function Engine.Controller.Dead.PawnDied:0036) KF-DegenerationMultiplayerV1.KFPlayerController Pawndied while dead
NetComeGo: Close TcpipConnection 87.248.4.121:2459 Sun Sep 13 20:21:22 2009
Log: PlayAnim: Sequence 'DeathB' not found for mesh 'British_Soldier1'
Log: RecvFrom returned SOCKET_ERROR 113
ScriptLog: ProcessServerTravel: KF-DeepThrustV1
ScriptLog: PreClientTravel
Log: Server switch level: KF-DeepThrustV1
Log: Browse: KF-DeepThrustV1?Name=KFPlayer?Class=Engine.Pawn?Character=Corporal_Lewis?team=1?Sex=M?game=KFmod.KFGameType?
NetComeGo: Close TcpipConnection 85.24.6.132:65098 Sun Sep 13 20:21:39 2009
NetComeGo: Close TcpipConnection 81.110.18.85:1181 Sun Sep 13 20:21:39 2009
Exit: Socket shut down
Log: ProcMesh 0 shared verts.
Log: xProcMesh numRigidVerts=36
Log: Collecting garbage
Log: Purging garbage
Log:
Developer Backtrace:
Log: [ 1] ./ucc-bin-real [0x869f1d7]
Log: [ 2] [0xffffe500]
Log: [ 3] ./ucc-bin-real [0x864bdd8]
Log: [ 4] ./ucc-bin-real [0x864bdef]
Log: [ 5] ./ucc-bin-real [0x864bdef]
Log: [ 6] ./ucc-bin-real [0x864bedc]
Log: [ 7] ./ucc-bin-real [0x8651380]
Log: [ 8] ./ucc-bin-real [0x8654fe4]
Log: [ 9] ./ucc-bin-real [0x81c0b4d]
Log: [10] ./ucc-bin-real [0x81a8861]
Log: [11] ./ucc-bin-real [0x81b2e99]
Log: [12] ./ucc-bin-real [0x8162d94]
Log: [13] ./ucc-bin-real [0x804da85]
Log: [14] /lib32/libc.so.6(__libc_start_main+0xd3) [0xf7ceeea3]
Log: [15] ./ucc-bin-real [0x804b0e1]
Log: Unreal Call Stack: UObject::GetPathName <- UObject::GetPathName <- UObject::GetPathName <- UObject::GetFullName <- DispatchDestroy <- DispatchDestroys <- UObject:urgeGarbage <- UObject::CollectGarbage <- Cleanup <- UGameEngine::LoadMap <- LocalMapURL <- UGameEngine::Browse <- ServerTravel <- UGameEngine::Tick <- UpdateWorld <- UServerCommandlet::Main
Exit: Exiting.
Log: FileManager: Reading 5 GByte 4219 MByte 4194777 KByte 923 Bytes from HD took 37.437458 seconds (21.155128 reading, 16.282330 seeking).
Log: FileManager: 0.000000 seconds spent with misc. duties
Uninitialized: Name subsystem shut down
Uninitialized: Allocation checking disabled
Uninitialized: Log file closed, Sun Sep 13 20:21:41 2009
NetComeGo: Close TcpipConnection 81.110.18.85:1181 Sun Sep 13 20:21:04 2009
NetComeGo: Open myLevel Sun Sep 13 20:21:04 2009 81.110.18.85:1181
NetComeGo: Close TcpipConnection 80.192.27.112:2140 Sun Sep 13 20:21:05 2009
NetComeGo: Close TcpipConnection 87.248.4.121:2244 Sun Sep 13 20:21:07 2009
NetComeGo: Close TcpipConnection 80.192.27.112:2165 Sun Sep 13 20:21:19 2009
Log: PlayAnim: Sequence 'DeathB' not found for mesh 'british_riot_police_I'
Warning: KFPlayerController KF-DegenerationMultiplayerV1.KFPlayerController (Function Engine.Controller.Dead.PawnDied:0036) KF-DegenerationMultiplayerV1.KFPlayerController Pawndied while dead
NetComeGo: Close TcpipConnection 87.248.4.121:2459 Sun Sep 13 20:21:22 2009
Log: PlayAnim: Sequence 'DeathB' not found for mesh 'British_Soldier1'
Log: RecvFrom returned SOCKET_ERROR 113
ScriptLog: ProcessServerTravel: KF-DeepThrustV1
ScriptLog: PreClientTravel
Log: Server switch level: KF-DeepThrustV1
Log: Browse: KF-DeepThrustV1?Name=KFPlayer?Class=Engine.Pawn?Character=Corporal_Lewis?team=1?Sex=M?game=KFmod.KFGameType?
NetComeGo: Close TcpipConnection 85.24.6.132:65098 Sun Sep 13 20:21:39 2009
NetComeGo: Close TcpipConnection 81.110.18.85:1181 Sun Sep 13 20:21:39 2009
Exit: Socket shut down
Log: ProcMesh 0 shared verts.
Log: xProcMesh numRigidVerts=36
Log: Collecting garbage
Log: Purging garbage
Log:
Developer Backtrace:
Log: [ 1] ./ucc-bin-real [0x869f1d7]
Log: [ 2] [0xffffe500]
Log: [ 3] ./ucc-bin-real [0x864bdd8]
Log: [ 4] ./ucc-bin-real [0x864bdef]
Log: [ 5] ./ucc-bin-real [0x864bdef]
Log: [ 6] ./ucc-bin-real [0x864bedc]
Log: [ 7] ./ucc-bin-real [0x8651380]
Log: [ 8] ./ucc-bin-real [0x8654fe4]
Log: [ 9] ./ucc-bin-real [0x81c0b4d]
Log: [10] ./ucc-bin-real [0x81a8861]
Log: [11] ./ucc-bin-real [0x81b2e99]
Log: [12] ./ucc-bin-real [0x8162d94]
Log: [13] ./ucc-bin-real [0x804da85]
Log: [14] /lib32/libc.so.6(__libc_start_main+0xd3) [0xf7ceeea3]
Log: [15] ./ucc-bin-real [0x804b0e1]
Log: Unreal Call Stack: UObject::GetPathName <- UObject::GetPathName <- UObject::GetPathName <- UObject::GetFullName <- DispatchDestroy <- DispatchDestroys <- UObject:
urgeGarbage <- UObject::CollectGarbage <- Cleanup <- UGameEngine::LoadMap <- LocalMapURL <- UGameEngine::Browse <- ServerTravel <- UGameEngine::Tick <- UpdateWorld <- UServerCommandlet::MainExit: Exiting.
Log: FileManager: Reading 5 GByte 4219 MByte 4194777 KByte 923 Bytes from HD took 37.437458 seconds (21.155128 reading, 16.282330 seeking).
Log: FileManager: 0.000000 seconds spent with misc. duties
Uninitialized: Name subsystem shut down
Uninitialized: Allocation checking disabled
Uninitialized: Log file closed, Sun Sep 13 20:21:41 2009
Upvote
0