{HoF} Evil's Sniper Deathmatch

Image
Image
Image
Image
Image
Post Reply
Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Thu Feb 06, 2020 6:46 am

I don't know if spamming forum with screenshots is that useful but let me post what I witnessed.
Fragme and other USA player had 80-90+ ms ping. After joining a "South" lager with 300+ they went to 120+. It's not a coincidence it's a fact happened. In first stage kills weren't that many but taunt messages... a lot.

Why I witnessed that ? I checked server in browser and I could see a high skilled dude with 300+ ping. Interesting...
The dude had X names and Y IDs so far, one of them having the same LST tag (and other tags), let's recall video with "LST Ruins" and he had closer IP classes.

I think we cannot have a LST server for those 30-60 players from there for a simple reason: These X players are one or two players, that's why you cannot see 20 of them in the same place, because others don't even exist. I want to see next players playing together: Hades, Anthrax, Jossy, Dyrce, Odin, la_llorona_SF50, Mio, fco, Orale_wey, -=CoN=-Bicholator, -=CoN=-Hades, SF50, Sf50, Dupa_Biskupa, -=CoN=-Hadess, Inframundo, -]LST[-*Jossy*, PalomaPapi, ELFANTASMAX, Lautaro, mam, ninjagalforf, kingman, profe, TRAUCO, prolapsus, hitlar666, deztroy, Josselyn, blabla, angeluz, ect ect. I don't want to write 10 lines with "names".
Due to these habits I restricted some ranges as long as I'm still waiting explanations for first video from that "RUIN" discussion - having the same LST tag. I'm totally disagree to see two, or even only one jester full filing database with data and lagging server and... even cheating, as demonstrated in said thread.
Just look at NexgenPlayerLookUp and see what I mean...
Next days firewall will print their attempts and others banned by Evil... I'm not gonna react at facts in game, my reactions will go strict into server's policies, simple as that.

In other hand 1000 Mb aka 1GB network is CPU intensive, data needs a fast processing and this is affecting performance. 1GB for a 20 years old game modem ready ? I was trying to set it at 100 but that adapter is configured like for a Files Hosting server not an old Game-Server. Worth wasting CPU performance with such a network ?
I have 1GB net adapters in both player and in my server, but I have settings:
- player 10 MBits;
- server 100 Mbits.
More than suffice for UT. Their brain-farts with 1GB are useless.

As for a technical goofing of HP from that time, actually HardDisk drive installed by factory cannot write at that speed offered by 1 GB network if you start downloading files, more than that, P4 D or P4 HT CPUs cannot handle such a network at full speed - this is insanity not high tech. I read docs over Internet, I'm not talking crap here.
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

User avatar
BluEyedBitch_{HoF}
Posts: 49
Joined: Sun May 05, 2019 12:10 am
Location: in my living room LOL

Re: {HoF} Evil's Sniper Deathmatch

Post by BluEyedBitch_{HoF} » Thu Feb 06, 2020 9:08 am

Nelsona, you're a smart fuckin' man, smart on a whooole nother level (excuse my informal prose) :lol: I appreciate you, and that you're here to help.
I'll Shoot You...
...Even in the Back

Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Fri Feb 07, 2020 6:33 am

To be honest I'm not that smart, but sometimes I have eyes and... I'm able to see certain occurrences.

As for South part, I have in my house in my workstation a "happy" list with IPs ranges containing all possible ranges used by South Side. Probably excepting two countries from there, some day I will firewall the rest because of their attitude generally:
- spamming chat/taunts;
- spamming server records;
- getting over a ban from admin without coming at discussions over forum;
- cheating;
- big ping and unstable connections making a mess;
- sitting death and taking slots at random, ect.
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Fri Feb 07, 2020 5:24 pm

Another fact:
I was taking a tour to that XC_Spectator which at a moment it's crashing server for some reason...
Let's say that I completed INT file for generating configuration and added INI config file to the server - it didn't have one. I see that it has some interesting functions so I think I'll check what is about.
Aside I think XC_Spectator doesn't crash because of itself but I think SpectatorPlusv2 - stripped - does damage when is dropping player. I believe and you need to convince me that code for dropping player is legit as execution like this:

Code: Select all

P.Destroy();
//aka
Player(Owner).Destroy();
This is what in other mods was avoided. If you want to drop player I think the normal way is to deal with him to disconnect by himself as it does when he leaves game properly calling Logout or whatever chain closing connection correctly. I think the player has to be motivated with a function simulated and replicated with this content:

Code: Select all

if ( Level.NetMode != NM_DedicatedServer )
	P.ConsoleCommand("disconnect");
Just my two 0.02 $.
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Sun Feb 09, 2020 11:16 am

Also with regard to whatever lags and "slow-downs" we do have some prints:

Code: Select all

ScriptWarning: Revenge DM-[BBST]-Thailand-Rmx.Revenge0 (Function Revenge.Revenge.listPlayers:0032) Accessed array out of bounds (75/32)
ScriptWarning: Revenge DM-[BBST]-Thailand-Rmx.Revenge0 (Function Revenge.Revenge.listPlayers:0032) Accessed array out of bounds (76/32)
ScriptWarning: Revenge DM-[BBST]-Thailand-Rmx.Revenge0 (Function Revenge.Revenge.listPlayers:0032) Accessed array out of bounds (77/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ScoreKill:0101) Accessed array out of bounds (40/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ScoreKill:0124) Accessed array out of bounds (40/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:0061) Accessed array out of bounds (41/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:0061) Accessed array out of bounds (40/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:0061) Accessed array out of bounds (35/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:004D) Accessed None
XC_Engine: === Last line repeats 67 times.
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ProcessLeadChange:007E) Accessed array out of bounds (41/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ProcessLeadChange:00D8) Accessed array out of bounds (41/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ProcessLeadChange:00E9) Accessed array out of bounds (41/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ScoreKill:01F6) Accessed array out of bounds (40/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.ScoreKill:0218) Accessed array out of bounds (40/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:0061) Accessed array out of bounds (41/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:0061) Accessed array out of bounds (35/32)
ScriptWarning: NFR_FeedbackMutator DM-[BBST]-Thailand-Rmx.NFR_FeedbackMutator0 (Function NFR_Feedback.NFR_FeedbackMutator.DetermineLead:004D) Accessed None
XC_Engine: === Last line repeats 67 times.
You might want to understand that arrays might not be the fastest ever processing neither when None it's being under processing - Poor None, this None is often accessed and slapped badly by all Evil CPU units... :lol: .
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

User avatar
SC]-[WARTZ_{HoF}
Site Admin
Posts: 394
Joined: Wed May 10, 2017 7:08 am

Re: {HoF} Evil's Sniper Deathmatch

Post by SC]-[WARTZ_{HoF} » Sun Feb 09, 2020 5:02 pm

So this mod could be the culprit for the high latency. I'm not educated on fixing this type of mod. Arrays are not my specialty. I'll disable this mod till there is a fix.

Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Sun Feb 09, 2020 6:48 pm

I cannot say 100% that mod is that nasty, perhaps is missing some conditions from sanity checks, usually I don't see these but something is causing this deprecation of executing code aiming a None target. I can expect a memory boundary hit because we have arrays loaded that takes space and this error is not an usual fact:
- MapVote - 1st heart - a lot of arrays static allocated;
- Nexgen having multiple arrays;
- Mods like this one;
- UTES;
- perhaps others including default arrays, server related things.
To not forget map related assets: textures, sounds, music mapped as packages in run-time natively.

I would like to have knowledge how to calculate amount of memory consumed and replication stuff when 10 players are playing.

I believe that's why game breaks sometimes, because if Engine cannot allocate anymore memory is goes down-hill unable even to "explain" what happened. If this is happening rare, probably we can live with it and we should take in account which map was running for a few basic calculations.

Edit: I'm thinking to add a small thing into whatever current server-actor (for not adding another one) a function intended to print from time to time memory status of UT - not for OS. When game is locked/crashed I would like to see last report what is saying . Probably we can figure when what and how loads server to boundaries.

UT is capable to print this report on demand:

Code: Select all

Log: Memory available: Phys=1440492K Pagef=2987340K Virt=1545704K
Log: Memory load = 31%
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

User avatar
SC]-[WARTZ_{HoF}
Site Admin
Posts: 394
Joined: Wed May 10, 2017 7:08 am

Re: {HoF} Evil's Sniper Deathmatch

Post by SC]-[WARTZ_{HoF} » Sun Feb 09, 2020 9:29 pm

Looks like after removing the NFRFeedback mutator the pings have stabilized. Nice catch Nels.

Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Sun Feb 09, 2020 10:06 pm

Watch this (I did it a few time ago)

Code: Select all

MemoryStatus: Load = 33% at 1 Players.
ScriptLog: Match Started
MemoryStatus: Load = 34% at 1 Players.
XC_Engine: === Last line repeats 2 times.
MemoryStatus: Load = 35% at 1 Players.
XC_Engine: === Last line repeats 4 times.
MemoryStatus: Load = 36% at 1 Players.
XC_Engine: === Last line repeats 5 times.
MemoryStatus: Load = 37% at 1 Players.
MemoryStatus: Load = 38% at 1 Players.
XC_Engine: === Last line repeats 9 times.
MemoryStatus: Load = 39% at 1 Players.
XC_Engine: === Last line repeats 4 times.
Memory usage grows during match. Here is reported each 11 seconds - configurable. Reporting starts at certain number of players - also configurable, and it stops logging if players are leaving remaining under configured value.
Do you think what I noticed ? I see that memory load it's not very freed during game, it keeps being taken all this time, disregarding what other genius said that garbage is collected during game - I DON'T think so. GarbageCollector is SLOW and it lags games because is very CPU intensive, Epic definitely did not set natively this "feature". Garbage Collector works at ServerTravel - Exiting, end of discussion.

Return then:
If we speak about Sniper Server it can be completed with such a tool for figuring what was the last memory charge before some funky crash - as much as it can be read from log fragment - usually even logs are broken in such case but in a previous line we can see the old charge.
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

Nelsona
Posts: 1122
Joined: Sat Sep 30, 2017 5:03 am

Re: {HoF} Evil's Sniper Deathmatch

Post by Nelsona » Mon Feb 10, 2020 11:39 am

And :? server was out this morning for some whatever application error. I see that it WAS NOT listed in DEP - so I added server there, and because nobody joined, I rebooted machine for applying changes. I think I have to insert that memory reporter because I see that things are not that nice last time...
Not often maintained
My UT Mapping works...
Learn the rules like a pro, so you can break them like an artist.
- Pablo Picasso -

Post Reply