r1ch.net forums
* Home Help Search Login Register
r1ch.net  |  r1ch.net stuff  |  R1Q2  |  Topic: UNSUPPORTED, UNOFFICIAL Q2Admin Security Update!
Pages: 1 [2] 3 4 5
Print
Author Topic: UNSUPPORTED, UNOFFICIAL Q2Admin Security Update!  (Read 75888 times)
zloy
Member

Posts: 24



« Reply #15 on: January 29, 2005, 11:59:29 am »

Maybe you want to make an UNSUPPORTED, UNOFFICIAL qtv update also? Smiley There are sources available.
Logged
iENO
Member

Posts: 7


« Reply #16 on: January 29, 2005, 01:52:13 pm »

Since we installed this version of q2admin we get a lot of "player frozen for exceeding the speed limit" but the client(s) are not speeding or using hacked clients, any ideas??
Try fiddling around with the ;Speedcheat detection vars...
Logged
R1CH
Administrator
Member

Posts: 2625



« Reply #17 on: January 29, 2005, 02:08:53 pm »

I've also had reports from other people that the speedcheat detection vars are screwy. Try just setting msec_max to 100000 or something as it doesn't seem you can disable it.
Logged
fanny
Member

Posts: 27


« Reply #18 on: January 29, 2005, 03:02:10 pm »

Thx, will give it a try, any chance of getting some descriptions of the speed cheat cvars i.e what / how do they affect the client etc.

Again thx

Logged
Snake
Member

Posts: 184


« Reply #19 on: January 29, 2005, 04:07:06 pm »

R1CH, iENO, the question of source availability hasn't been answered yet... Is it? Or maybe for serveradmins only?
Logged
trayton
Member

Posts: 3


« Reply #20 on: January 30, 2005, 11:41:34 am »

thanks for this will update asap!
Logged
DZ
Member

Posts: 26


« Reply #21 on: January 30, 2005, 09:04:22 pm »

I've also had reports from other people that the speedcheat detection vars are screwy. Try just setting msec_max to 100000 or something as it doesn't seem you can disable it.

Yeah, I see a lot of people dropped for 'irregular movement'.  I gather that is what we are talking about here?

Some are lagging when it happened, some weren't.   That is the odd part I supose.  I think they had cpu lag thoug, virus updating or something going on.  Could be the culprit there.
Logged
WORM.
Member

Posts: 71


« Reply #22 on: January 30, 2005, 09:15:35 pm »

For the love of all that is holy, do NOT release the source! Haven't we got enough cheats out there as it is? huh
Logged
_DOA_MERLYN
Member

Posts: 35



« Reply #23 on: January 30, 2005, 09:24:20 pm »

Yeah, I see a lot of people dropped for 'irregular movement'. I gather that is what we are talking about here?

I see that happen alot as well, even before I updated with this version of Q2Admin...

Quote
Sat 29 Jan 2005 20:31 modelindex 34 allocated to models/items/mega_h/tris.md2
Sat 29 Jan 2005 20:31 modelindex 35 allocated to models/items/armor/shard/tris.md2
Sat 29 Jan 2005 20:31 0 entities inhibited
Sat 29 Jan 2005 20:31 0 teams with 0 entities
Sat 29 Jan 2005 20:31 -------------------------------------
Sat 29 Jan 2005 20:31 Man_With_Plan enters the melee
Sat 29 Jan 2005 20:31 outlaws enters the melee
Sat 29 Jan 2005 20:31 Venom enters the melee
Sat 29 Jan 2005 20:31 _DOA_WOLVES enters the melee
Sat 29 Jan 2005 20:31 Stuffing client 2 (Man_With_Plan)
Sat 29 Jan 2005 20:31 Stuffing client 0 (outlaws)
Sat 29 Jan 2005 20:31 outlaws: ahh not this again
Sat 29 Jan 2005 20:31 outlaws was dropped: irregular movement
Sat 29 Jan 2005 20:31 Dropping outlaws, irregular movement.
Sat 29 Jan 2005 20:31 outlaws d1sc0nn3ct3d
Sat 29 Jan 2005 20:31 _DOA_WOLVES was dropped: irregular movement
Sat 29 Jan 2005 20:31 Dropping _DOA_WOLVES, irregular movement.
Sat 29 Jan 2005 20:31 _DOA_WOLVES d1sc0nn3ct3d
Sat 29 Jan 2005 20:31 Man_With_Plan was dropped: irregular movement
Sat 29 Jan 2005 20:31 Dropping Man_With_Plan, irregular movement.
Sat 29 Jan 2005 20:31 Man_With_Plan d1sc0nn3ct3d
Sat 29 Jan 2005 20:31 Venom was dropped: irregular movement
Sat 29 Jan 2005 20:31 Dropping Venom, irregular movement.
Sat 29 Jan 2005 20:31 Venom d1sc0nn3ct3d

why is it doing that?  can anything be done to remedy it?

Logged
iENO
Member

Posts: 7


« Reply #24 on: January 30, 2005, 09:47:33 pm »

I see that happen alot as well, even before I updated with this version of Q2Admin...

why is it doing that? can anything be done to remedy it?

If I'm not mistaken this is being done by r1q2, not q2admin.

sv_enforcetime description:
Quote
If set to values higher than 1 will kick clients that exceed this amount of "bad movement". Every bad movement detection will increase an internal counter by 1 plus however far off the movement was. When this counter exceeds sv_enforcetime, the client will be kicked. Good movements will decrease the counter over time to hopefully account for the odd lag spike causing it to be hit. Good values would be between 5-25.
Logged
R1CH
Administrator
Member

Posts: 2625



« Reply #25 on: January 30, 2005, 10:31:55 pm »

If everyone on the server is being dropped at once, it could indicate a CPU usage problem on the server. The "bad movement" detector relies heavily on the server frame time keeping up to what it's supposed to be, if the server is lagging behind then clients will start getting kicked. Also the sv_enforcetime value should be around 5 at a minimum - if you have it set to 2 then most likely everyone will get kicked upon joining as a client will almost always cause one "bad movement" due to being out of sync with the amount of ms the server is assigning.

I've not tested it but it's possible Q2Admin could be interfering with it, set sv_enforcetime 1 if this is the case.
Logged
_DOA_MERLYN
Member

Posts: 35



« Reply #26 on: January 30, 2005, 10:44:20 pm »


If I'm not mistaken this is being done by r1q2, not q2admin.

yee, poorly worded, doh...  it was happening with your 1.326d so I figured it was something to do with the r1q2 dedicated.exe...

I reset the sv_enforcetime cvar to 1 for now (had it at 25), I'm dumping Q2Admin anywayz and setting it all up through r1q2...  thx for replies... Smiley

Logged
_DOA_DEATH_JUMP
Member

Posts: 2


« Reply #27 on: January 31, 2005, 01:10:54 am »

R1CH,

I read a long, long time ago that the sv_enforcetime 1 command is not very good to use because if there is a client that has a slow connection, the whole server will slow down to enforce the same time for everyone.  Therefore, one person can slow down the entire server.

Code:
// these 2 lines sorta fix SPEED HACK cheat. Cheaters on a server will be stopped 90%
// BUT will cause excessive lag for the server.  Please update your MOD to a newer
// version that has a 100% speed-hack fix built-in.
set sv_enforcetime "1"
First of all, is this true?
Second of all, if it is, does your coding of sv_enforcetime cause the same thing to happen?

One other thing, this command checks for "movement".  Merlyn and I have been testing out a mod and I was killing myself over and over again and for the time I was in the map, I wasn't moving at all.  Can you tell me how killing yourself over and over while not moving at all would cause an irregular movement error??  Notice that I killed myself about 20 times in a row because of a glitch in the mod where you hold down the fire button to respawn and it kills you a bunch of times in a row.  But, if you kill yourself, there is no movement, and when you respawn, there is no movement, right??  So, I don't see how these 2 things can cause the error.  Unless it counts the tilting of your screen when you die to be movement.  If it does count the movement of your screen when you die, can you remove that from the coding? Got any ideas?
http://www.deathjump.com/images/irregular_movement_error.png


I have a suggestion for an extra type of speedbot detection.  I have no idea how taxing it would be on the server or how effective it would be, but I'll run it by you anyway.  When you run a speedbot, not only is your speed increased, but when you jump, you jump up and down really fast.  (at least the 2 speedbots that I've tested out with Merlyn do)  (and it happens whether you change the timescale or msec) What about if, when the server registers that a player jumps, it records the z position and time, then when the player lands, the server checks to see if the player lands at the same z position.  If so, then, the server calculates how much time it took the player to do this and then compares it to the time it should take the player to normally jump (with gravity factored in).  If the time takes less than what it should, then in theory, the player is using a speedbot.  The only thing is, did the player jump up and hit a block or another player, or get hit by something in mid air (a rocket, push/pull, etc.).  If so, then the check is dropped.  If a player laggs or if a player jumps off of something else like another player, or if the player does a matrix jump, then the time is going to be longer and it would pass.  If the player jumps to a higher or lower position, or on a slant, then the calculation gets thrown out too cause the z position is not the same.  Lagg should never cause a player to jump faster, only slower, so if the check is done correctly, all it would take is one failure to show a speedbot.  I understand that you don't always have really flat maps, and some mods have hooks, or some people don't jump very often, but all it would take is ONE jump on flat ground without being hit by something in mid air to cause the detection to pick it up.  I suppose if you used a more complex equation to describe the trajectory of someone jumping off one flat surface to another flat surface whether it is to a higher location or to a lower location and factor in gravity, you might be able to check for a lot more jumps.  As far as I know, none of the trick jumps or physics glitches will ever cause your time to jump up and down to change unless the negative z-space (crate) jump causes the time to jump to be just slightly longer.  But if you use a positive z-space jump as the shortest allowed time and subtract 3 to 5% of the time and set that as the limit, it might work.

Thanks,
DJ
Logged
R1CH
Administrator
Member

Posts: 2625



« Reply #28 on: January 31, 2005, 12:57:09 pm »

You are constantly sending clc_move packets which use up your msec regardless if you're actually moving or not. The comment about sv_enforcetime is incorrect, it will not "lag the server" or slow anyone down to the slowest speeds, it only affects single clients who exceeds their allotted msec which is why it defaults to 1 in R1Q2.

Unfortunately due to how Q2 deals with backed up or lost client packets, it's possible for a lagging client to seemingly do these 'fast jumps' too. However it's a nice idea, but given how fast some speedbots can go, some players may not even jump at all. The current setting seems to detect them just fine after a minute or so.
Logged
_DOA_DEATH_JUMP
Member

Posts: 2


« Reply #29 on: January 31, 2005, 05:06:16 pm »

Thanks.

Can you tell me what affect that has on a client that connects to the server with cl_maxfps 120?
Logged
Pages: 1 [2] 3 4 5
Print
r1ch.net  |  r1ch.net stuff  |  R1Q2  |  Topic: UNSUPPORTED, UNOFFICIAL Q2Admin Security Update!
Jump to:  

Powered by SMF 1.1.19 | SMF © 2013, Simple Machines