dedicated-servers

[frozen comet]

nitrado is shit i know from other games

[thorny dew]

Is game hosting really that garbage nowadays?

[frozen comet]

its 50/50 coinflip on nitrado some server work perfect some are just garbage and problem everyday

[deep turret]

there are lot of different providers with varying price and service quality and hardware use skils

[reef pasture]

Is there still no solution for the server that stop working after you exit the game when you run it on the same computer as you play ?

[cold estuary]

bc nobody knows that u must buy extra for better ram haha

[deep turret]

not currently apparently, appart from some weird binding workarounds

[reef pasture]

tried them all, kinda sucks nothing works..

[frozen comet]

on that server hosting who @deep turret send is writed *Plans provide uncapped RAM. Only player count is limited

[thorny dew]

Eh fair enough

Ain't that hard to just slap ptero or pelican on, but somehow I do find game hosts that struggle with that

[midnight sonnet]

so dedicated servers become worse after 1.0?

[thorny dew]

Oh yeah definitely. Experimental dedicated servers definitely did not have issues whatsoever /s

Just wait for a patch tbh

[regal marsh]

apart from desync issues (removing foundation does not remove it visually sometimes, sometimes cant pick stuff up as if it was not there, painting issues, vehicles resetting) and some config options resetting
it works great for me

[reef pasture]

do you run the server on the same computer as you play ?

[regal marsh]

no on a 7700 ryzen, 4 cores (well threads actually...2 actual cores) of it in a windows VM in a server center

[deep turret]

you must be joking right ?

are we people who helped find the workarounds for update8 a joke to you ?

it definetely was a pain in the ass

[thorny dew]

...

[deep turret]

I must not know what /s means

then

[thorny dew]

sarcasm

[deep turret]

oh thank you

1.0 is more reliable than update8 tbh, just that the options are buggy (but only the options mainly)

[midnight sonnet]

and still i have 15 tick rate in tier 7, maybe i need to change network quality or change configs such as engine.ini lol

[regal marsh]

few things that might still not be separated per client
I think I heard my colleagues say that if you use "custom" paint, it then applies it to all players for further building
didnt try myself but heard they said that
few other similar things I dont remember now
minor things, nothing game breaking or too annoying

[deep turret]

bugged options may include currently :

• ip binding (even when no special binding, fail to bind correclty unless you set workarounds)
• port setting (yes it will set the correct port for the set bind, but not for the other ipv4/6 stack)
• network setting (in server manager but does not save unless you trick it into saving it)
• .ini files (does not create by default need tripping by console quit for default creation)
• still does not know how to let client game report that it's not crashed but is simply loading, which lead to timeouts

[winter flower]

Hey, i have installed a dedicated Server on Windows with the SteamCMD, i can reach it from another PC in my Network with the ipv4 adress, but not the ipv6 adress. i need the ipv6 for my friends to reach my server. the error is: "Not possible to reach server api"
Any Ideas how to fix it?

[mellow solar]

i use 4netplayers personnaly never any issue on any game and very cheap

[midnight sonnet]

do you have settings for .ini files by any chance?

[deep turret]

unfortunately some issues make the server fail to bind correctly to both ipv4 and ipv6

you want to ask it to bind specifically to the ipv6 address

#1284813963133194341 message here some helper script (for Linux)

[winter flower]

okay, i will try this

[deep turret]

it depends on what you want to setup

there are options for unsupported player max count, unsupported network bandwitdth settings, unsupported tickrates, increasing timeouts timers, disabling world streaming

[thorny dew]

Something tells me 120tps is not gonna last me very long

[deep turret]

tps ?

tickrate ?

[thorny dew]

Yep

[deep turret]

yes

[midnight sonnet]

i have a dedicated server for 2-3 folks, but performance is not great
maybe issue in my Intel i7-8700, but maybe i can make game more comfortable by tweaking configs

[deep turret]

if you have enough ram in the first place you can try disabling server streaming

people say it increase cpu usage but it's a fake problem, it increase cpu usage because the cpu can more easily process more work, which is what you want

apart from that, no other settings will affect performance significantly

and dont expect miracles from this one either

the most impacting thing for performance, is the environment/machine in/on which you run the server

[midnight sonnet]

is 10gb enough?

[regal marsh]

wait, what server streaming?

[deep turret]

this is completly overkill
also server world streaming does not have to do with networking

[regal marsh]

they actually say 12gb minimum and 16 recommended
but I also have yet to see more than 10 used

[thorny dew]

tier 0 moment

[deep turret]

a feature to decrease RAM load on early saves, by not loading the whole world, only parts that player will access eventually
on any big save it is useless, and as the world need dynamic loading (that's the "streaming" part), it can cause performance issues

[spiral lion]

better have overhead than fully utilized.

[thorny dew]

Sorta. For Satisfactory you need every single bit of CPU you can use

As literally everything is on a single core

[deep turret]

it grows with save size

[regal marsh]

what option is that?
dont see anything in "stream" in search in the dedi/config wiki

[thorny dew]

And the problem is that most people do have that single core fully utilized by satisfactory at later stages, so w/e gets you that bit of perf back is good

[deep turret]

I would say that unless you are very interested in the impact it may have for you : if you dont know what it is, that means you dont need it

[regal marsh]

oh I guess either they changed the requirements, or I was looking at the fandom wiki back then
now its:
Minimum of 4GB (Linux), 8GB (Windows?), 16GB RAM may be recommended for larger saves or to host more than 4 players.
so yeah, sounds about right when I was checking my usage on windows

[regal marsh]

I want to know all options 😄

[deep turret]

-ini:Engine:[ConsoleVariables]:wp.Runtime.EnableServerStreaming=0

I dont know if it works as launch argument

but anyway it gives you the path

[regal marsh]

cool thx

[thorny dew]

That's 0 by default anyways

[deep turret]

not officially on 1.0 server

where do you get that from ?

[thorny dew]

UE docs

[deep turret]

that's UE default, not Coffe Stains satisfactory default

[thorny dew]

And just checked my server too

Basically fresh install and it has that setting set to 0

[deep turret]

^

[deep turret]

linux ?

[regal marsh]

yeah same here

windows

[thorny dew]

I'm on linux

[deep turret]

hm

so they didnt go with it then

you really fully fresh installed ? no setting in .ini ?

[winter flower]

now im able to reach the server with ipv6, is there a way to bind the server with -multihome to the static ipv6 instead of the dynamic?

[thorny dew]

The only thing being set in engine.ini

[/Script/FactoryGame.FGSaveSession]
mNumRotatingAutosaves=${NUM_AUTOSAVES}

[CrashReportClient]
bImplicitSend=${UPLOAD_CRASH_REPORT}

[/Script/OnlineSubsystemUtils.IpNetDriver]
InitialConnectTimeout=${INIT_CONNECT_TIMEOUT}
ConnectionTimeout=${CONNECT_TIMEOUT}
[/Script/OnlineSubsystemUtils.IpNetDriver]
NetServerMaxTickRate=${TICK_RATE}
LanServerMaxTickRate=${TICK_RATE}

[/Script/SocketSubsystemEpic.EpicNetDriver]
NetServerMaxTickRate=${TICK_RATE}
LanServerMaxTickRate=${TICK_RATE}

[/Script/Engine.Engine]
NetClientTicksPerSecond=${TICK_RATE}

Otherwise it's all the defaults that satisfactory makes

[deep turret]

I would arg it's good then forget that setting maybe

[regal marsh]

I never changed that option
and I have clean steamcmd install

no server manager or anything

[deep turret]

then forget that setting

Coffe stains said themselves they would go with upstream patched world streaming for 1.0 that's why I thought it was enabled by default

along that it use way less ram when just starting

but if they didnt enable it it's great

[regal marsh]

well they also said they will set the new interactive console as default on next update after 1.0
and they didnt, still need -newconsole
😄

[deep turret]

I was arguing for not world streaming servers back then

[past plaza]

hi can somone help me with my background service? it wont work but i have no clue why (german)

[deep turret]

wat are you specifically talking about by "background service"

[past plaza]

running it as a service on my debian server

[deep turret]

how does it "wont work" exactly ?

[past plaza]

gives me an error every time

[regal marsh]

gotta say what the error is 😄

[past plaza]

something in here ^^

[thorny dew]

Check /home/steam/Steam/logs/stderr.txt

[regal marsh]

maybe you dont run the steamcmd with the steam user?

just found something similar on reddit

[past plaza]

09/16 10:53:39 /tmp/dumps: is not owned by us - delete and recreate.
09/16 10:53:39 /tmp/dumps: could not delete, skipping.
09/16 10:53:39 minidumps folder is set to /tmp/dumps01
09/16 10:53:39 Init: Installing breakpad exception handler for appid(steam)/version(1726260406)/tid(6791)

i specificliy installes steam cmd with the steam usr

if i run it manualy it will work is really have no clue

[regal marsh]

in your .service
you have
User=steam
?

[past plaza]

ye

[deep turret]

can you show us sudo systemctl cat satisfactory

[regal marsh]

might also be no space (but if you run normally fine then I guess not)

[past plaza]

1 sec

[Unit]
Description=Satisfactory dedicated server
Wants=network-online.target
After=syslog.target network.target nss-lookup.target network-online.target

[Service]
ExecStartPre=/usr/games/steamcmd +force_install_dir "/home/steam/SatisfactoryDedicatedServer" +login anonymous +app_update 1690800 validate +quit
ExecStart=/home/steam/SatisfactoryDedicatedServer/FactoryServer.sh -multihome=0.0.0.0
User=steam
Group=steam
Restart=on-failure
RestartSec=60
KillSignal=SIGINT
WorkingDirectory=/home/steam/SatisfactoryDedicatedServer

[Install]
WantedBy=multi-user.target

[regal marsh]

guess 0x206 might be wrong user
0x202might be no space
but cant find 0x204 on google

[deep turret]

it throws local errors too which is weird ?

[regal marsh]

thats pretty normal

locale errors

should not be an issue

[past plaza]

ye but it crashes

[deep turret]

I mean I dont have any, but my server is set up to us utf8

so /shrug

[deep turret]

can you cat /etc/locale.conf

[past plaza]

no

mkdir?

[regal marsh]

can probably set it up with

sudo update-locale LANG=en_US.UTF-8```but that should not be an issue

[past plaza]

i try

nah wont do anything

[deep turret]

what does findmnt -t tmpfs /tmp output ?

[past plaza]

what linix dist are you using?

[regal marsh]

everything similar I find
either wrong permissions
wrong user
or no space

[past plaza]

30g should be enough i think

user is save steam

[deep turret]

so ?

[past plaza]

and permissions i have no clue abbout 😄

[deep turret]

it complain for when it want to write to /tmp

[regal marsh]

yeah that looks suspicious

[past plaza]

one set restarting rn

maby it cant create the dir?

outputs nothing

[deep turret]

ok then you have bad /tmp

did you update your system ?

[past plaza]

im running it in a container if this helps

ye fresh installed

debian 12

[unreal flume]

I really wish more games were designed to take advantage of multithreading.
I have a 7950x with 16 cores and 32 threads that's limited to only a fraction of its potential performance capabilities.

[deep turret]

(I'm joking, but I do sometimes write clear in discord after sending a message)

[past plaza]

ahh _D

[thorny dew]

Honestly there are some things that could be fixed by at least moving the networking to a separate thread

[stable wadi]

do dedi servers pause when no players?

[thorny dew]

It should be by default, but you can change it ingame

[stable wadi]

😮

[regal marsh]

they said its not completely single thread
but the main game stuff is on one thread

like many other games

[deep turret]

it's also limited by UnrealEngine architecture

[regal marsh]

old source games are not much different (guess the csgo one did improve it a bit, or source 2)

[deep turret]

and doing multithread things is more complicated, a freaking lot of devs do that wrong

[past plaza]

does someone have a fix or realy good tutorial for me for seting up the whole thing?

need some pot brb

[deep turret]

if you use container ask about the container I guess

I only know direct environment

maybe there are some bad things going on in container /tmp access

[past plaza]

or i just use screen ^^

[deep turret]

or maybe you are trying to bypass the container

[thorny dew]

True, but it's 2024 and I'd like to see networking related stuff like the api not be tied to the server's tick rate

[regal marsh]

would be nice

[deep turret]

either you use container or you do bare install, but you can not do both unless you setup your own container fresh

[past plaza]

nah i need my containers for cloude and stuff ^^

would be nice if there was a turnkey for satisfactory server ^^

[deep turret]

there are

[past plaza]

rly*?

[deep turret]

yes

[past plaza]

oh you have a link?

[sweet gorge]

Did anyone find out where the server stores the API-Key you generate via "server.GenerateAPIToken" ?

[deep turret]

#dedicated-servers message

[regal marsh]

#dedicated-servers message

[lunar creek]

Anyone tried writing a prometheus metrics exporter using the new server api?

[thorny dew]

I mean yeah, Satisfactory does use multiple cores every now and then. Saw some spikes up to 160% CPU so it's doing stuff

[regal marsh]

in windows 1 player usage looks like this

thats 60 ticks and increased rates

[thorny dew]

You can even see satisfactory spawn and kill subprocesses here and there

[deep turret]

if your question is too much custom specific then go https://discord.ficsit.app/

[pallid relic]

Would be interesting for the average tick rate

idk what interesting statistics we can export from the API

[deep turret]

the average tickrate is already showin in server manager and documented

[regal marsh]

not sure if windows is maybe better than linux on using other cores
I know in old source games, in linux it stricly cant use other cores
but on windows it puts networking in another
but thats old 32bit engine/srcds
UE might not have it same, I guess its more 64bit

[thorny dew]

Exporting via prometheus lets you make fancy historical graphs 🙂

[pallid relic]

Yeah, but for prometheus you can monitor the tick rate overtime

[thorny dew]

So you can look out for dips

Could even set alerts if tickrate falls below 50% or something

[deep turret]

I said its documented

some script already exist

if you want to go beyond just go on mod community

[past plaza]

thanks but i ment a pre configuted container for download ^^

[thorny dew]

They don't have to, but yeah

Prob a lot more development-oriented folks there

[deep turret]

did you read the https://satisfactory.wiki.gg/wiki/dedicated_servers

[deep turret]

I mean yes, they have, this is not a thrid party api use support channel here

[thorny dew]

Check the docker portion of the dedicated servers section in the wiki

[past plaza]

yes my config is from there but it doesnt help 😦

[deep turret]

you asked for LINKS

its in the page

[regal marsh]

you can link the item
https://satisfactory.wiki.gg/wiki/Dedicated_servers#Docker
https://satisfactory.wiki.gg/wiki/Dedicated_servers/Running_as_a_Service#Docker

[past plaza]

oh i'm sorry diddnt see this ❤️

[deep turret]

they asked for premade ones

[raw frigate]

Hey guys, I am having trouble making an external connection to my dedicated server on ubuntu.
-The server is running
-Port 7777 is forwarded and open according to yougetsignal.com, UFW allows both tcp/udp. Server is listening.
-My only start parameter is -Port=7777.

Do I need a multihome parameter?

[pine spade]

do you have a home server or a rented one with a fixed and reachable ipv4?

[raw frigate]

home server

[pine spade]

do you can reach the machine via ping?

[raw frigate]

yes

[deep turret]

how did you test external connection ?

[scenic birch]

anyone having issues on windows running multiple satisfactory servers?

[pine spade]

do you run the Server via the docker container?

[scenic birch]

technically I'm able to run multiple, but as soon as someone leaves one server, it crashes both servers with a socket error

not running in docker currently if you're talking to me

[raw frigate]

The server sits inside a separate VLAN and I'm trying to connect via the public ip and port inside the game

[pine spade]

i meant quintus, i cant't help you with windows 😦

[raw frigate]

no

[thorny dew]

Time to run it in docker o7

[pine spade]

mark his words haha

[deep turret]

what public ip then, the one of the whole house ?

[pine spade]

then it isnt really public, or am i wrong?

[raw frigate]

Yes my public ip, and connections for 7777 are forwarded to my server machine, is this incorrect?

[deep turret]

the problem is we have very little detail on your vlan setup

[sonic sable]

any trick to having the server not pause when empty?

[deep turret]

you should use local ip when connecting from inside network

[thorny dew]

Change setting ingame

[torn harness]

mate has anyone had any trouble starting a server with g portal or anyone else? none of them work atm

[deep turret]

public ip will have varying results depending on how good the networking is on the router

[pine spade]

this should do it

[deep turret]

the only real test is literally connecting from outside

[pine spade]

or you setup dyndns with a domain, worked for me

[sonic sable]

not working. Using the EGS Dedicated server if that matters

[raw frigate]

I also have a hostname with noip and a service running to update it on my machine, and I cannot connect with the hostname either

[nimble pine]

my friends can connect guys but i can't

[deep turret]

this have nothing to do with networking, and have still the same exact issues depending on what address is returned by it

[dusky wedge]

You need to open TCP 7777 as well

[pine spade]

true that lol

[deep turret]

the dns address will just return your public address, this is no change

[nimble pine]

no firewall

[raw frigate]

true

[thorny dew]

Forward the port on your router

Unless it's a local server

[nimble pine]

its a server + if i open vpn it works

[raw frigate]

I'm at a complete loss lol. By all accounts it should work

[nimble pine]

ok will try

[thorny dew]

bruh

[deep turret]

#dedicated-servers message or use local ip

[thorny dew]

Do a test by doing curl -k https:// public_ip:7777

remove the space

Cause automod doesnt allow links

whoops put local instead of public

[deep turret]

no you had it good, put local instead of public

[thorny dew]

If it reaches epic's http handler then it's somewhat working

Either way works tbh

[deep turret]

unless you are literally from outside which you have no choice

[thorny dew]

Test from LAN then test from WAN

[dusky wedge]

Or by tossing it into a web browser, which might be more accessible to most people seeking assistance

[deep turret]

nope, because joining wan address from lan is a trick, which is not handled correctly by all routers

[thorny dew]

Then you can just do this on mobile data and test that way

But yeah it does depend on your router

[deep turret]

then they would need to know how to circuvent certification errors

[thorny dew]

That's fine

That means that epic is trying to serve a cert to begin with

Which means you're reaching the server

[dusky wedge]

Only need to differentiate between a cert error and a timeout

[thorny dew]

What's not good is if your connection just times out

[dusky wedge]

Not like they need to actually see the error response from the API handler

[deep turret]

fair

[naive atlas]

Hello, yesterday in my dedicated server, I have issue items slow in mk2. I don't know if this issue is meeting by my server performance (ram was ok, network was ok), my pc or dedicated server knowing issue.
Do you have meet issue like that ? (Item is hard slow and lag)

[tranquil stirrup]

Hey! Im trying to make a dedicated server on ARM ubuntu 22.04 but seems complicated to install steamcmd, anyone can help?

[green panther]

I have experience with linux, but not steamcmd, could you link whatever guide you used?

[deep turret]

arm is not supported by satisfactory, is not supported by steam
and steam is not even supported by box64/86 virtualization layers

[tranquil stirrup]

There is not a guide i could find... thats why im sking for assitance. I have read using FEX, but im stuck since i have it installed and cant install steamcmd either

[tranquil stirrup]

https://www.reddit.com/r/SatisfactoryGame/comments/1az37fd/dedicated_server_on_arm/

Some people had made it

[deep turret]

where do you see written that is easy, supported, and that it use the steamcmd basic install guide ?

[green panther]

Yea not sure how much help I'll be, never used ubuntu before and im a bit spoiled by AUR 😅

[tranquil stirrup]

Where did i said its easy? im asking if someone who made it already can help.

[deep turret]

What I mean by that is that you are coming asking for common guide on how to do something that is obviously a hack and not generally guided at all

[green panther]

have you looked at this? https://github.com/FEX-Emu/FEX

[tranquil stirrup]

Yes, i have FEX already on the machine

[green panther]

alright lemme do some research rq ill let u know if i think i can provide decent guidance

[tranquil stirrup]

Ty so much

[deep turret]

tldr do not use steamcmd

likely build a container on x86-64 machine, then run it once installed on box64

[green panther]

I will say that one of the resources linked in the reddit post u sent claims that the server crashes when placing conveyer belts, and that they have not found a fix

so not sure how worth it this might be

https://github.com/sa-shiro/Satisfactory-Dedicated-Server-ARM64-Docker?tab=readme-ov-file#-important-updated-for-10

[deep turret]

it's like mods, even if it's possible, here it's the official discord channel for the dedicated server, not the hack hub

[dusky wedge]

Installing is likely the easiest part of the process to run it on a different architecture.

[green panther]

yea i think its a lost cause until someone finds a fix for that crash

[dusky wedge]

Which is not to say it’s easy, but that what follows will be even more difficult

[green panther]

are u trying to run the server on ur own hardware? or are u using a hosting service

[tranquil stirrup]

I was just using this because oracle give it for free

[green panther]

bc it might be worth switching to a different host if ur not using ur own hardware

[green panther]

nvm this answers my question

do u have an old computer lying around? u could try setting up a server on ur own hardware

not easy if u have no experience hosting ur own servers but not too hard if you do your homework

[tranquil stirrup]

I have some other pc, but i dont want to have them running 24/7

[green panther]

understandable

[scenic birch]

I keep getting this error and the server crashes:
LogServer: Error: FMultiBindDatagramSocket failed to read data from underlying socket Unreal: SE_ECONNRESET

[green panther]

also some ISP's dont like people hosting their own servers

[green panther]

not sure but i can tell that its a networking issue

[scenic birch]

yeah it's really weird, just started today

[tranquil stirrup]

i have some hosting provider for my webs, but nothing that can handle a server

[green panther]

u shouldnt need crazy hardware but def a lot of RAM

[past trail]

getting the same. hosting on a DS when anyone leaves it kicks everyone and starts showing as offline and that error comes up in the console.

[green panther]

havent read the whole thing but this might help https://questions.satisfactorygame.com/post/66e08561772a987f4a8a9d67

[tranquil stirrup]

Yes, but its just hosting service, no dedicated machine

Using hostinger

[green panther]

yea i would recommend setting up your own machine to host it, its free and usually runs better than most hosting services, but obv u need to set it up urself and maintain it urself so not the most user friendly method

[lofty lantern]

I found the pinned messages DOH silly me.. But it doesn't really tell me were to reset it on the faulty pc.

[tranquil stirrup]

Its more about the money than the diff or the maintenance. Electricity is pricy here

[green panther]

understandable, then yea i would go for whatever's cheapest, which in this case might be a hosting service

u would prob want at least 8GB RAM for satisfactory but dont quote me, havent set up my own server yet

[tranquil stirrup]

But since its only for my friend and I, i dont want to buy any hosting. So thought i could use the free oracle server, i already made a minecraft server there and thought it could work

[tranquil stirrup]

the one i have has 16GB

[thorny dew]

We do love OCI out here

[green panther]

looking into AWS now, might be a good option

[thorny dew]

I love getting a $2k bill for a satisfactory server

[green panther]

nvm it makes u put payment info to make an acc, which is a bit far for me

its not like i think amazon will scam me, but i dont want to accidentally pay for a service i dont need

[tranquil stirrup]

Dont know whats that

[thorny dew]

oracle cloud infrastructure

The thing you were running your free mc server on

[green panther]

yea not sure what the best approach is for ur case, i would always recommend self hosting but obv thats not for everyone

[green panther]

did u solve this

[tranquil stirrup]

I have an old laptop with 64gb Of RAM and a rtx2080 with an i9-11xx. but dont know if its worth having it run 24/7

[thorny dew]

If you can somehow disable the 2080, maybe

[green panther]

i would try running a cpu stress test for like 10 mins and check the battery before and after

[green panther]

yea that might be good too

[tranquil stirrup]

Dont think you can config th max-q

[thorny dew]

Damn

[green panther]

a lot of gaming laptops have a switch

i think its called a MUX switch?

but if possible yea disable that and use integrated graphics

[thorny dew]

That's for the more expensive laptops xd

[green panther]

damn

[tranquil stirrup]

Well.. its cost 3.7k

[green panther]

yea never owned a gaming laptop so idk

[tranquil stirrup]

almos the same as my desktop

[thorny dew]

Well shiiii I'd double check

[frosty gale]

My cheap lenovo laptop has a MUX switch, I think they're quite common now

[thorny dew]

Nowadays yea

[tranquil stirrup]

its from 3 years ago

[frosty gale]

I've had it for 3+ years

[green panther]

but yea i would try and see how much power it uses and decide if the energy usage is worth hosting the server

[scenic birch]

I believe so, I set bindings to a single IP address instead of 0.0.0.0

[past trail]

seems that error is showing up as it kills the UDP port when anyone leaves the server

[green panther]

👍

[green panther]

ah, not sure what the fix would be then

[thorny dew]

Damn during the gpu crisis lol

Fun times

[tranquil stirrup]

Yep, tried bouilding a desktop and could not

so i went for a laptop

[scenic birch]

nope I didn't solve it

[thorny dew]

EVGA got my back 🙏

[green panther]

i have only ever used arch (which i would not recommend unless u already have experience with it) but i know one of my buddies always recommends POP! OS to people with little linux experience for hosting servers

[scenic birch]

just created a server on linux and it doesn't appear to be doing that, but windows does

[tranquil stirrup]

now i have the desktop, with a ryzen 9 7950x3d and a 4080 super

[green panther]

could be windows firewall

[thorny dew]

W specs

Went with 7800x3d myself because I don't want to be dependent on microsoft's scheduling update whims

[green panther]

i've always gone intel but my 11900K throttles under load with a 3 fan AIO lmao

[tranquil stirrup]

I mainly use the computer for blender and UE5 so its neat

[green panther]

well 2 fan radiator but i have 2 fans on top 1 on bottom

[tranquil stirrup]

I use a 3fan radiator

[thorny dew]

As long as you avoided that whole 13th and 14th gen garbage fire then I'd say you're chilling

I use a 420mm AIO for a 80W CPU.

[green panther]

yea its good for my use cases which is just various compilation tasks and blender

[thorny dew]

Well it really was for my old 5950x running 250W but now my 7800x3d runs with such little power that the rad is practically useless

[green panther]

yea i use a 240mm for a 125W cpu lol

technically can run at 5.3GHz but i have it clocked at 5.0GHz

[thorny dew]

I'm sad that my CPU is capped to 5GHz

[green panther]

no OC?

[thorny dew]

Nop

[green panther]

damn

yea i usually buy OC cpus but i never actually OC lol

[thorny dew]

I oc'd my 5950x and it was fun

[green panther]

also since we're in this channel i might as well ask

how easy would it be to set up a satisfactory server over SSH

at school rn so might as well while im not doing anything

[green panther]

on Arch

ooh looked like theres a satisfactory server package on the AUR

[thorny dew]

Uhh just do docker tbh

But yeah it's not too bad to begin with

[green panther]

whats that

[thorny dew]

containers

Generally an easy tool to use for running apps in a container

[green panther]

o right i forgot my school blocked port 22

🥲

also blocked my vpn recently

[thorny dew]

Just change your ssh port later when you have access

[green panther]

how do i change ssh port?

[glad scarab]

Hi, are dedicated servers stable and easy to work around in the current state for IT newbies ?

[thorny dew]

Not really, but there's loads of information available both here and on the wiki

Like, getting a server up and running shouldn't be too bad

But once you get into firewalls and port forwarding, it's a mixed bag

[tough venture]

aaah, cant wait to get home from work to grow the factory! On my dedicated server! (Edit: realized just now this was dedicated server chat)

[turbid idol]

Me when i'm not able to drive any vehicles because if i do i just disconnect from the server

[naive atlas]

Has anyone encountered errors on MK and items that hang or are slow ?

[quaint vault]

Hi, anyone knows a good server renting for SEA?

[tight spire]

Does dedicated server need a good HDD?

[raven gull]

Ok, are there any settings I can play around with to maybe stabilize the dedicated server? I'm running on a Hetzner dedicated cloud VM, but it seems like the CPUs of those aren't up to snuff to running Satisfactory servers, as when I'm exploring the enemies rubber band hard making it almost impossible to fight them

[deft tartan]

I am using the same provider, but everything is still smooth. Maybe network quality setting?

Also maybe our save is still very light could also be that

Check the tickrate in server manager

[raven gull]

mh, that problem was since starting the game almost

[green panther]

has anyone installed a server on arch

[raven gull]

avg tick rate is 11

[deep turret]

is it your pc ? because poeple here have reported if you do not go through the right connection menu it will show this error ?

[deft tartan]

I am on CX32 with 30 tickrate

[frosty grove]

only for satisfactory 🤣

[raven gull]

Oh, so shared vCPU even, not the dedicated one? I'm running on CCX23
Are you using the docker image? / Did you change any settings from the base settings?

[deft tartan]

nope, just a clean install using steamcmd on debian 12

didn't change any settings that matter for performance

[raven gull]

well
that means somethings up with my install

imma try around, thx for the info ^^

[thorny dew]

Pull up top and send screencap

Maybe you're hitting a bit of cpu steal

Theoretically CCX line shouldn't have any, but you never know

[deft tartan]

don't quote me on that but from my long term usage of hetzner, I am pretty sure you really wanna still go for Intel when going to host a game

[thorny dew]

Not really

It depends on the CPUs used

[deft tartan]

It is just from experience using Hetzner, not something I can back up with proof

[thorny dew]

For CCX, maybe you can ask them to move you to their 9654s

[thorny dew]

For the VPS the intel shared vcpu is better than the amd shared vcpu that's running on epyc 7002

But that also kinda depends on exactly what xeon gold they're using

[deft tartan]

I'd say just try rescaling @raven gull to the same plan I am using currently and see if that helps

[thorny dew]

But the dedicated vcpu should be better on average

[raven gull]

ye imma try that first

[deft tartan]

Currently when someone is on a server, the cpu usage seems to be very evenly spread out, so the 4 vcpus could be outperforming the 2 dedicated ones

[graceful birch]

are rented servers working well ? might just rent and save some energy

[raven gull]

how far into the game are you / how much did you build already? So i can have a sense of scale to my save

[deft tartan]

Early phase 3

[lofty lantern]

No its another computer I have 3 machines and 1 server All seperate . 2 work no problem but 1 just doesnt want to connect and giving me the API issue.

This is what appears on the server when it tries to connect

[2024.09.16-13.15.46:606][ 3]LogSocketsSSL: Warning: Failed to initialize SSL state for Socket HttpRequest
[2024.09.16-13.15.46:607][ 3]LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR

If I change the port on my firewall and change it on a working PC and I get the Sever Security Warking but not on the one that isnt working

[kindred ore]

is there a way to make the sever reset eveny 5 hours then the interval hours ?

[raven gull]

oh man thats wayyyyyyyyyy better

and same

[deft tartan]

Check tick rate

[raven gull]

huh didn't think it would make that much of a difference
will stay with that now

30

[deft tartan]

Set network quality to ultra as well, at least so far I had no issues with that

[raven gull]

yep thats what i was tinkering around with before, didn't think the VM was the problem

[deft tartan]

It should make everything even smoother at the cost of a bit of cpu

[raven gull]

thx for the help ^^

[deft tartan]

you are welcome

[edgy ingot]

Thanks everyone for the help, I was able to get the game up and running. I was confused on how ports work 🤣

[tired ether]

Anyone else have issues with sparkedhost declining payment? I've tried 5 different cards and every single one is declined.

[deft tartan]

You should contact their customer support

[hard lantern]

is there an admin mode fly command?

I see it in console, but doesnt seem to work

[regal marsh]

steam can run on raspberry, but might be trouble 😄
I wouldnt expect the game/server to run in it 😄
managed to run a source game on it, but the physics were all messed up and then it crashed 😄

[kindred ore]

is there a way to make the sever reset every 5 hours then the clock time ?

[tight orbit]

Is there anyway to disable to auto server restart? I enabled my own custom cron schedule to restart the container, but even after setting the restart interval at 0:00 the server still restarts evert 24 hours. Any idea how I can force this behavior to stop?

[regal marsh]

nope, one Coffee guy here said its mandatory (actually you can with a lanuch parameter #dedicated-servers message)

[tight orbit]

What does the interval even change then?

[regal marsh]

its the hour when it restarts

ignore the description

[lofty lantern]

Hello I looked at the pinned message for the API setting I need to reset the certificate on one of my Computers doesnt anyone know where this is located for the Satisfactory game

[tight orbit]

So 0:00 is 6:00pm? What fucking timezone is that lmao

GMT 0?

[regal marsh]

in your server time

but yes 24:00 means midnight

[tight orbit]

0:00 should be midnight, its not using my timezone America/Denver

[thorny dew]

It's prob UTC+0

[ocean bison]

Anyone else have problems driving vehicles? E.g. explorer, sugar cube. Me and a friend randomly get kicked from the server or the game crashes. Sometimes takes just a minute other times 10 mins or not at all.

[regal marsh]

I had default 24:00 and it was my 00:00 on the server
changed to 06:00 and now its 06:00 on my server
while I have CET time on my server, which is different than UTC
windows

[tired ether]

Contacted shockbyte support. They said my card is being declined due to "Fraud chargebacks for buyer's IP in last 90 days" which is BS. I have never issued a charge back. Anyhow, is there a way to access my save file or is that stored on their server? They say there is nothing they can do to continue my trial or allow me to purchase from them? My only option at this point is to move my save to another service.

[thorny dew]

Save is stored on the server

[tired ether]

Alright. PSA NEVER USE SHOCKBYTE

[thorny dew]

I'll forward that to their team 😂

[tight orbit]

Self-host 🤩

[tranquil wedge]

How do you change the server restart time?

[tired ether]

Tell them to go fuck themselves while you're there lol

[tight orbit]

Read above

[thorny dew]

Aww darn the guy I knew worked at shockbyte doesnt work there anymore I don't think

[regal marsh]

#dedicated-servers message
but around that time he also explains that you should restart every day at least

#dedicated-servers message

explanation: #dedicated-servers message

[kindred ore]

is there a way to make the sever reset eveny 5 hours then the clock time hours ?

[thorny dew]

Cronjob

Set the restart time flag to a high number and instead restart the server using a cronjob/scheduler

[tranquil wedge]

Thank you!

[lofty lantern]

Anyone know how to fix cannot connect to server api

[silver orbit]

that is usually TCP not being open

[thorny dew]

Gotta forward port 7777

Both TCP and UDP

[kindred ore]

what is cronjob/scheduler

[silver orbit]

linux version of task scheduler

[kindred ore]

lol its on windows

[thorny dew]

For windows it's Task Scheduler

[silver orbit]

ok..

[regal marsh]

just I guess with manual restart if you just kill it, it might not save

if you kill it during save, bad

unless you gracefully do it somehow
it will not notify anyone its restarting

[silver orbit]

create a Powershell script that shutdown the server and then starts it back up

[kindred ore]

that what i was thinking

[silver orbit]

you can have the pS script read the log file to see when the last save was done

[tranquil wedge]

That's what I'm trying to figure out how to do. To have the dedicated server software gracefully shutdown, then restart the VPS, so it can all come back on line in sync.

my server now that we're getting bigger is needing to be restarted about every 12 hours, or things start getting out of sync.

[toxic anchor]

Is the Blueprint library shared with all the players and can all the players contribute to the Blueprint library?

[silver orbit]

you can use the API to stop the server

then once confirmed down restart it

[tranquil wedge]

API? Is that the console in the satisfactory client on my computer?

[silver orbit]

then use a task to run that on whatever echedule you will need

[kindred ore]

then seems like way to much work then just turning it off and on lol

was hoping there was a console command or a think in the files where u can set more time for it to reset or to set a time to reset cos not once have i used the sever restart interval cos i shut sever down when im not using it

[tranquil wedge]

Is there a way to do that automated?

[silver orbit]

no it is apart of the satisfactory server

[tranquil wedge]

That would be the console manager in server manager?

[silver orbit]

automate restarting the server gracefully, yes

[tranquil wedge]

Thank you for helping btw.

[silver orbit]

no

[tranquil wedge]

That would be on the Linux server side then?

[silver orbit]

in the install directory of the game is the api documentation.
{install dir}\SteamLibrary\steamapps\common\Satisfactory\CommunityResources

to be clear this is the game not the dedi server install

[tranquil wedge]

ok, I'm there now. Looking at it.

[tranquil wedge]

Thank you for helping!

[silver orbit]

if you give the api documentation to copilot or chatgpt it might be able to creat the script for you ¯_(ツ)_/¯

[silver orbit]

no problem

[tranquil wedge]

hmmm really?

[silver orbit]

yea, it is possible, or might get it close enough

[tranquil wedge]

I'm not saying I used bing to build my website, I'm just saying it's better than it was...

😉

[silver orbit]

LOL, I used it to help write a scipt to make backups of the save directory. cause i was lazy and didnt want to type it

[wraith talon]

there any writeups of "known issues with dedicated servers" for satisfactory? I've got a dedicated server running under linux/docker, 16 gig RAM limit, 2 players on the server, and we've found it being glitchy occasionally- sometimes issues with items not transferring correctly between box/inventory, sometimes issues where things we destroy in-game stay there as ghosts.... we can see them, we can't build over top of them, but they're not really there.

[tranquil wedge]

Restart your server more often.

[wraith talon]

mebby there are some common tuning parameters known to help things?

[tranquil wedge]

We had those problems as well. The people here are helping my ignorant a$$ to figure out how to do it automated.

[wraith talon]

oof, ok. I already set up a daily server restart

[silver orbit]

there are also weird issue with the save file that can cause stuff to , but more lliekly the server needs restarted

[tranquil wedge]

We're going to attempt to do a 12 hour restart cycle.

We also have about a dozen people using it.

[wraith talon]

gotcha

[silver orbit]

Mine always loses the personal storage boxes

so they are invisible

[wraith talon]

I mean for me it usually fixes if I simply log out and reconnect to the server

[silver orbit]

but they still work, so not a big deal.

yea that can do it to, sometimes the client doesnt pull all the info down

[wraith talon]

which implies a likely culprit is missing messages between client/server

the fact that there's still UDP being used in the server protocol seems a very likely culprit for that

[silver orbit]

yea, udp isnt the best for that

[wraith talon]

but.... it also seems like restarting the server causes the issues to happen less frequently

[silver orbit]

yea there are definitiely issue with the dedi servers

[tranquil wedge]

Thank you everyone, this will take me awhile to digest the documentation to figure how to do this.

[silver orbit]

There are commons one we see in this chat a lot. the biggest is the server locking up or freezing after the last person exits the server

[tight spire]

Why server lags when anyone travel around the world? (Ping increases from 30 up to 400) (Nothing is overloaded, enough free ram and CPU is on 30%)

[wraith talon]

but I figured it was worth asking if there were any known tunables to adjust that might help reduce the glitches

[silver orbit]

this depends on so amny details. but likely the server needs restarted

[tight spire]

Server is working like 10 min

[silver orbit]

yea, I am waiting to see what CSS says about the dedi servers since they are suppose to be 100% supported now rather than an afterthought

[silver orbit]

huh ?

[wraith talon]

ok, ty for the info!

[tight spire]

I restart server, go travel around the world and ping go 💀

[tired ether]

Ok as an update for anyone who is thinking to try out shockbyte using their trial, there is no way to extend that trial. If the trial expires and you haven't went in yourself and downloaded your save file, it is GONE with no way for them to recover it. Even if you pay them to upgrade after the trial, your save is just gone. Fuck shockbyte.

[silver orbit]

OUCH!

[tranquil wedge]

Well, it's less money to pay them than it is for me to host through IONOS and do all the work myself.

The IONOS server is nice though. I can do anything I want with it.

There inlies my problems...

[gentle shore]

is anyone else having lots of issues when dismantling or recoloring or really doing anything with static objects like walls or foundations?

any objects created in previous sessions, if I try to dismantle them, it looks like they're still there until i leave the game and rejoin

it's super annoying

happens to everyone on our server

[chrome pumice]

If your talking about the restart in the server manager settings that's a clock not an interval. Just fyi.

[vast yoke]

i dont know if this is the right place to post. i have problem with connecting to g-portal server.
copied IP and port, make shure that server is online on the dashboard but gamve sasy "this server appears to be offline"

i have turned game on and off few times, was waiting at the server manager menu for 10 minutes

[gentle shore]

what timezone is that clock? ours said 24hrs

[lofty lantern]

I have 3 machines and 1 server All separate . 2 work no problem but 1 just doesn't want to connect and giving me the API issue. I dont want to have to format the computer for this ... problem there must be a solution

[tough geyser]

I think it is just dedicated servers in general. I have mine on my own PC and it also said the same thing 🤷‍♀️

[gentle shore]

but it restarted at 8pm eastern us

but it seemed like roughly 24 hours after we had started the server

[deep turret]

udp does not mean you can't have safeguards though

it just means it must be handled, instead of ignored/managed/builtin the transport layer

[chrome pumice]

The same as the server time. So 24 hours = midnight server time.

[silver orbit]

for gportal make sure you are using this IP and Port

[gentle shore]

it's a hosted server, so i guess i'll do the math and adjust the time so it's not in the middle of our peak hours

it's on indifferent brocolli and supposed to be hoste in NA eastern time zone, but it definitely wasn't midnight when it restarted

[deep turret]

steam is known for NOT running on any arm

[vast yoke]

thanks, but still same :/

[deep turret]

because it have requirements for old x86 code/libraries

[silver orbit]

then the server isnt running or it is being bloxked outbound for you. probably need to contact gportal

[regal marsh]

well I ran it on my raspberry few years back

surprisingly even source game started up, but super low fps and weird physics (pretty much expected)

[deep turret]

I mean, steam games sure

steam itself big doubt

[regal marsh]

well it did

the arm TwisterOS should have it preinstalled
EDIT: well I guess thats through box86 emulation

[modest remnant]

Does running satis off a dedicated server fix crashes related to playing just p2p?

[rain pine]

is there a way to set the time to restart the server (so like fixed hour for example 04:00 ) instead the slider setting (its confusing that slider) ?

[timber forge]

How do I take my local saved game and port it to a dedicated?

what file do I need to copy over

[regal marsh]

that is THE time!

not interval

and yes the description is misleading

very

24:00 = midnight server time

[raw spoke]

ugh i just hit that error when i reconnected to my server, i was given a new character instead of my correct one. I really hoped this bug was solved for 1.0 but it looks like it isn't.

[gentle shore]

your previous pioneer was determined to be inadequate. please do better with this one. -- FICSIT

[raw spoke]

lol actually yeah it doesn't look like i had that bug. My guy just DIED while i was away.

[fathom saffron]

yet another question about failed to connect to the server api, any idea why I get it on Ipv6 but not Ipv4?

[foggy belfry]

does anyone know what would be a lighter load on a hardware? I have a laptop (which is what I would play on) and a PC (what i would host the game on). I could either set up a server using steamcmd on the PC, and then run the game on my laptop. Or, I could run the game on the PC and them stream it to my laptop.
Im debating which would have a lighter load on the PC (and to some extent on the Laptop too)

[tiny topaz]

i guess a trial is called a trial for a reason

but damn the most advertised mc server hoster being so bad? lmaooo

[fathom saffron]

tbh backups are backups, but yeah that sucks, they could keep the saves like a month in case of issues with payment

[reef gull]

Use the game's "Upload Save" in the Server Manager

[still hollow]

When i delete Blueprints buildings the Server crashes , someone know why?

[bleak vessel]

i would like to let the devs know that flying crabs impact client players from too far away on dedi server (with good ping around 60 ms). at this range it's almost impossible to swat them out of the air with melee since they are hitting at the very edge of melee range. in single player this is not an issue.

[sturdy gust]

Because it binds to localhost for v6 for the API. It's a known bug

[fathom saffron]

is there a workaround?

[sturdy gust]

Workarounds have been sent in this channel, I do not recall it off the top of my head

[chrome pumice]

Not really a bug you are just reconnecting to quickly after getting disconnected. Have to wait like 30 seconds to a minute for the server to know you dropped.

[chrome pumice]

Check the pins for it maybe. Ipv6 is less used and less documented.

[silver cedar]

when i moving around here, the ping suddenly go very high.....any way to fix it?

[heavy furnace]

Hi fellas, just a quick question about networking. Would you say that the dedicated server uses the whole space available according to MTU?

[fathom saffron]

I haven't tried the "-ini:Engine:[HTTPServer.Listeners]:DefaultBindAddress=any" workaround, do I just add it to the docker run or somewhere else?

currently I'm working on just setting up an extra docker to redirect all the traffic to the ipv4

[sturdy gust]

No, that workaround is no longer needed

[undone quiver]

dedicated server started running glitchy after I blew up a rock, went to main menu and back in the game and my player icon changed and then it showed that i picked up all my resources/gear at the same time. then when i logged in next time my player was dead. now all the players icon randomly change color and things feel off on the server. any idea what I can do to resolve this?

[fathom saffron]

the issue is that my ISP is ipv6 only, it runs ipv4 through a nat.

[fervent thistle]

I set up a dedicated server and did the port forwarding on my router (and windows firewall) for all 3 ports listed, and I can't connect to the server via the client, and when I join the server through server manager, it disconnects the server when I log out, anyone else have this issue and is there a fix?

[sturdy gust]

There are two ports

7777/tcp, and 7777/udp

You only join via the server manager. The "join" button is not for dedicated servers

[fervent thistle]

I first setup 7777 for tcp and udp but when I was still having an issue I saw a thread also callling out 15777 and 15000

[unreal crystal]

those are old and outdated.

Also if you're running the server on the same host as the client; this is a known problem

[fervent thistle]

I am attempting to have a friend join the server to see if at least they can connect to the current setup, However, when I login via server manager, the server listed in server management goes offline and needs to be rebooted

[fervent thistle]

ok, yes, I am running on the same client

[unreal crystal]

you can try and run the server on a diff port but idk how well it reacts to being bound to other ports. no one's been pushing that heavily

everythings on 7777 listening to each other by default; client + server when its on same host.

[fervent thistle]

all good, if its a known issue, I'll wait for a fix, in the meantime I have a beefy laptop I can run the server on so I can connect to it from a different client

or just never log out lol

[supple raptor]

Hi guys
Im trying to figure out this "server name pending" issue is solvable or is it a bug?

[twin ember]

Hey when i try to add a server in server manager it doesnt show up i have been trying it for 30 minutes now any idea’s?

[stoic stream]

any dedicated server fixes on the way?

[sturdy gust]

Could you elaborate? What do you see in the server manager? Offline?

[sturdy gust]

Could you provide more information? There are no known issues about this

[twin ember]

O no lol i just figured it out when i coppied the ip there was like :27482 random number yk but had to delete that

[lofty lantern]

LogSocketsSSL: Warning: Failed to initialize SSL state for Socket HttpRequest
LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR

The system that tries to connect that received the Failed to Connect to the server API

[manic rune]

Is there anything I can do against ghost trees?

Foilage doesn't fully update on my client when I join my server.

[proven forge]

Running a server and when people get a client crash they don't seem to respawn in their own body anymore until they go over and kill the old body of theirs and it releases the inventory on that body. Any reason for this?

[hot crystal]

have had this problem too, tends to happen if you rejoin too quickly after crash i think. started waiting 5 minutes after crash before rejoining and hasn't happened since

[midnight harness]

Has anyone here got the API working?
I'm not quite understanding how to connect from the docs.

All I get is a 404 (Not found) at /:

data={'errorCode': 'errors.com.epicgames.httpserver.route_handler_not_found', 'errorMessage': ''}

which I guess is to be expected, since I didn't choose a function.

If I send

{
    "function": "HealthCheck"
}

as Content-Type: application/json to / I receive a 400 (Bad request):

data={'errorCode': 'errors.com.epicgames.httpserver.invalid_content_length_header', 'errorMessage': ''}

I checked the Content-Length Header and it is set to 27 (which should be correct with ASCII and UTF-8, since {"function": "HealthCheck"} is 27 characters long (so 27 bytes).

What am I doing wrong? Must be something really stupid

[tulip iron]

Are you sending newlines or anything?

Since those count.

[midnight harness]

Ah damn, my message is deleted due to links

[midnight harness]

Don't think so. I didn't set 27 myself, that was the library. I assume it strips them

[midnight harness]

but hey, you getting it to work means it is definitely a me problem, nice

[tulip iron]

Nah, I'm not using one myself, it's just a common gotcha when calculating the size

[midnight harness]

For reference, this is my request in python:

api_url = "https:[slashslash]example.com:7777"
headers = {
    "Content-Type": "application/json; charset=utf-8"
}
response = session.get(
    api_url,
    headers=headers,
    json={
        "function": "HealthCheck"
    }
)

This is the debug output

DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): example.com:7777

send: b'GET / HTTP/1.1\r\nHost: example.com:7777\r\nUser-Agent: python-requests/2.27.1\r\nAccept-Encoding: gzip, deflate\r\nAccept: */*\r\nConnection: keep-alive\r\nContent-Type: application/json; charset=utf-8\r\nContent-Length: 27\r\n\r\n'

send: b'{"function": "HealthCheck"}'

reply: 'HTTP/1.1 400\r\n'

header: content-type: application/json;charset=utf-8

header: content-length: 98

DEBUG:urllib3.connectionpool:https:[slashslash]example.com:7777 "GET / HTTP/1.1" 400 98

response.status_code=400

So no, no newlines it seems

[quick mist]

I have some other calls in Python that you can reference if you need them

[midnight harness]

Yes that would be awesome

Was a hassle to get SSL working correctly. If you dealt with that, that would be very helpful too. Currently have to subclass requests.adapters.HTTPAdapter and create an own pool manager

[midnight harness]

(Since the chain.pem has FactoryGame as hostname)

[quick mist]

I have one (the HeathCheck) in my repo, https://github.com/dopeghoti/SF-Tools; here are the others I've had success wtih. Others are failing due to getting no response rather than a no content response:

[tulip iron]

You can generate your own certificates (using lets encrypt or so), but that requires fiddling with all that stuff.

[midnight harness]

Nah, I just downloaded the auto-generated PEM and made urllib3 verify that.
It works, but I'm sure this could be done easier

[proven forge]

Thank you, I wonder if there is somekind of a timeout timer I can adjust for it instead maybe, I might update this once I figure it out after work.

[midnight harness]

I could also just use a reverse proxy, but it is my friends server, so I won't touch his stuff :D

[midnight harness]

Alright, thank you very much.
I was missing the /api/v1 endpoint. I know this is the default, but was this documented anywhere? (must have made a typo before)

POST /api/v1 works (as expected)

[quick mist]

I don't think it's actually documented; I lucked into seeing someone else post that path and it clickd that of course they'd use the standard REST endpoint path

[midnight harness]

No idea why it gave me invalid Content-Length only when I passed in {"function": "HealthCheck"} lol

[quick mist]

one of those "in retrospec, obviously" things

HealthCheck seems to need a data element even though it's not actually relevant

[midnight harness]

I tried that already... but back then I used a GET-request, since they said it should work with some requests that don't require sending data

[midnight harness]

yup, at least the API gives that back as an error tho

Okay nice!

[midnight harness]

If you care, I could add a pull request with adding SSL support.
But only if you want

[quick mist]

this has really been annoying me though:

I'd love any PRs you care to offer

I'll add my REST examples to the repo shortly; I've been meaning to and just needed a round tuit.

[midnight harness]

That's all you pretty much need

But I would definitely make that optional then :D

[quick mist]

and I'd have to document the need to capture that .pem but that's not a huge lift

[midnight harness]

I couldn't find it anywhere in the server directory

I downloaded it using a browser. No idea if there is a better way

Must be stored within the UE sav or sth

[sturdy mist]

Hello everyone, a friend and I play together on a multiplayer server. If one of us logs out, the other one is kicked out of the game and the server is shown as offline by Satisfactory. The server then has to be restarted in order to continue playing. Is this a known problem? I also always get a 520 kb update on Steam after this action, as if it were shooting me a file....

[quick mist]

this is common if the server is being run on the same physical host as the client. a client disconnecting looks to the server like a shutdown request apparently. CSS is working on it.

[midnight harness]

uh I'm an idiot. I just checked with a proxy... the body is dropped. Could be expected when using a GET request (mb)

[midnight harness]

If I understood it correctly, this is just where you place it if you provide your own

[topaz scarab]

server's not responding with a valid json payload, prob

[sturdy mist]

Great, thank you very much 🙂

[quick mist]

I checked it with Wireshark: the server isn't responding at all in those cases rather than giving an HTTP response payload with no content which would be valid.

it should respond with HTTP/204

[midnight harness]

I get a 204

[midnight harness]

Yup, empty response

send: b'{"function": "VerifyAuthenticationToken"}'
reply: 'HTTP/1.1 204\r\n'
header: Server: FactoryGame/++FactoryGame+rel-main-1.0.0-CL-366202 (Linux)
header: keep-alive: timeout=15.000000
header: content-length: 0
DEBUG:urllib3.connectionpool:https:[slashslash]example.com:7777 "POST /api/v1 HTTP/1.1" 204 0
response.status_code=204
data=''

[quick mist]

for me I don't even get that which is why Requests is throwing an exception

maybe you can help with that once I get the examples up

[topaz scarab]

because you're calling r.json()

so it's trying to parse no data as json, which isn't valid json

[quick mist]

🤦‍♂️

[midnight harness]

Dunno, this works for me. With json() I'm getting a similar error

[quick mist]

I need to check r.status_code. Deeeerp

[viral ocean]

hello, i can play and manage the Server when i use the direct connection i become a Token error... how i can fix it and Other Player can connect too.

[midnight harness]

Yeah, exactly

[lofty lantern]

Is there a way to reset caching the rejected certificate and regenerate a cert on the server

[topaz scarab]

🤔 the cert is somewhere
wait, is it fucking baked into the game, i hope not

[quick mist]

I think if you drop in a cert and chain of trust into Certificates/ and restart the server it should Just Work.

[topaz scarab]

it's in the wiki
don't use join, use server manager

[viral ocean]

thank you

[midnight harness]

At least not documented anywhere (I think)
I guess resetting the server entirely shouldn't hurt tho

[topaz scarab]

oh is it in the server sav file

[quick mist]

I am lazy and not fully awake yet and therefore my quick and dirty examples will look like:

[topaz scarab]

yeah i think it's in the server sav file, that's annoying

[topaz scarab]

i wouldn't just assume the data key exists
they can return errors

[midnight harness]

I can't find it anywhere, so yeah I guess

[midnight harness]

This introduces errors, when you hit Ctrl+C at the right time ;D

[topaz scarab]

my fork had a whole bunch of error handling in your post request function

FGGeneratedCertificateData yup

[midnight harness]

uuuh, nice

[quick mist]

this is a quick-and-dirty proof-of concept, maturity will come with time. I don't want to make EnterpriseClassFizzBin out of this (:

[midnight harness]

xd

Thanks for your probeLightAPI. That is a really great point of reference

Now I might use this instead of waiting for timeout when loading a save

[storm violet]

@sturdy gust Updated your image today from 1.0 server to 1.0.0.1, all backups got deleted, and server was fully reset, dunno if it's you or the game. Still sad to loose everything, even though it's only been a week.

[quick mist]

My pleasure. Took a while to build and I still need to make both sides use the same method to construct/deconstruct the package rather than using numpy on one side and struct on the other but it works and that was the important thing on Day One

[midnight harness]

I can imagine :D

[placid night]

Has anyone here used tailscale to share out their server to trusted friends/family?

[frosty gale]

did you reset your docker or something

[midnight harness]

daang, that sounds like a container volume issue

[topaz scarab]

did you not give it a volume

[frosty gale]

sounds like you kinda just deleted your container

[quick mist]

I am still trying to figure out how to properly construct the request payload for this:

[midnight harness]

are you using host bindings?

[storm violet]

It's bind volume, I've been using docker for many years

[fallow linden]

cau u tp someone on dedicated server useing the server console?

[midnight harness]

I think you need to contain it in the data object

[midnight harness]

There is always only these two iirc

[topaz scarab]

is the 'u' uppercase @quick mist

[quick mist]

My first try was putting updatedServerOptions into data (tried with both u and U).

[sturdy gust]

Really sorry about that man: https://github.com/wolveix/satisfactory-server/issues/266

[topaz scarab]

or is that also a doc error

[sturdy gust]

There was an issue with the image released within the first hour of the 1.0 workaround being out

So anyone that pulled in that short window wouldn't have persisted data

[topaz scarab]

oh yeah it should be in a data shouldn't it

[supple raptor]

I'm using steamcmd to start my server. I'm by no means an expert at this sort of thing but i followed most of the info from https://satisfactory.wiki.gg/wiki/Dedicated_servers

I open up port 7777 on my router, added the two firewall settings for tcp/udp(?) for 7777 and made sure the port was open using the nmap script in the guide.

I can connect from my game client using my internal server ip but i cannot connect using the external ip without getting "server name pending, not authenticated"

[midnight harness]

Just checked, this works for me:

response = session.post(
    api_url,
    headers=headers,
    json={
        "function": "ApplyServerOptions",
        "data": {
            "UpdatedServerOptions": {
                "FG.DSAutoPause": "true"
            }
        },
    }
)

[supple raptor]

I didnt really try anything else 😄

[frosty gale]

https://tenor.com/view/easy-button-that-was-easy-red-button-press-the-button-gif-17005111

[sturdy gust]

If you're attempting to use your external IP from within your network, it's very likely not to work. Use your local IP within your network, and let people outside of your network use the external IP

Yes, it may work for other services. No, it doesn't really work for Satisfactory

https://github.com/wolveix/satisfactory-server/wiki/Troubleshooting-FAQ

[random relic]

Having exactly the same issue. Two different dedicated servers, ports appear to be open. No firewall running on localhost. We can connect to eachothers' servers, but a third player is unable to connect to either of ours. Port scans say we're open. curl returns {"errorCode": "errors.com.epicgames.httpserver.route_handler_not_found", "errorMessage": ""} from outside the network. Remote Satisfactory client says server appears to be offline.

[sturdy gust]

That's a different issue

Your UDP port is unreachable

[topaz scarab]

check your port forward for a nat loopback setting

[midnight harness]

But that means it is open and should work

[lofty lantern]

I deleted the server and installed it in another folder and on another drive on my server still giving the same error

[quick mist]

yep, embedding UpdatedServerOptions in to Data worked this time now that I'm properly looking for HTTP/204

Progress!

[topaz scarab]

your settings are in another castle, see wiki config files

[midnight harness]

If outside of your network you get a valid HTTP response, your TCP forward is correctly setup.
Maybe UDP is the problem here?

[vapid cedar]

do you have any suggestions where I can rent a very fast satisfactory server with easy setup process? Supercraft?

[supple raptor]

I didnt know that! I'll try to disconnect my client from the network and try with a mobile hotspot instead

Thanks for the tip

[supple raptor]

Thanks! I'll look into this if wolveix suggestion doesnt pan out 😄

[midnight harness]

@sturdy gust Thank you so much for the docker container btw. :)
Funny github star graph indeed

[sturdy gust]

Lol right 😄

Skyrocketed from 700 to 1.1k 😄

[green panther]

yall ever accidentally break ur ssh config file accidentally and then u cant ssh in to fix it

👍

turns out u cant use port 444 for ssh

[quick mist]

Sure you can if nothing else is on that port

ssh -p 444 [e-mail]

[green panther]

i tried 443 because my school wifi blocks port 22 but then my website didnt load because it used https so i tried 444 and now it just refuses to respond

[green panther]

that worked fine on port 443

and my website is accessible so i know my server's up

[quick mist]

443 is very often already in use because that's the default TCP port for HTTPS transactions.

[green panther]

yea

[quick mist]

Attempting to serve both ssh and https on the same port at the same time from the same host will be an interesting experience.

[green panther]

yea ssh just took precedence

[barren swallow]

Anyone else having major lag issues on dedicated? specially on trains?

[green panther]

i mean my certificate is out of date anyway so i should have just used http so it wouldnt be a problem for accessing my site

[sturdy gust]

Likely due to slow single-core performance on your server's CPU, or low RAM allocated

[midnight harness]

Is 6GB really a good limit with 1.0?

[barren swallow]

Did recommended specs change that much? Same server ran perfect with 4 people pre 1.0.

Running 4 core 16gb ram

[midnight harness]

What CPU model?

[sturdy gust]

6GB is pretty low

[quick mist]

REST API demos uploaded to my repo for those keeping score

[midnight harness]

What are you using? Is that your stuff, or is there already an API wrapper for Python?

[frosty gale]

api wrapper

[quick mist]

@frosty gale went ahead and rolled out an whole API wrapper library while I was working on the UDP protocol and then pivoted to teaching myself how to interface with a REST endpoint

[midnight harness]

Is there an idea behind the limits in your docker compose?

deploy:
  resources:
    limits:
      memory: 6G
    reservations:
      memory: 4G

[supple raptor]

Thanks so much, this worked 🙂

[quick mist]

(Don't think I said this yet but I've been meaning to; that's excellent work you've done there gaymer)

[sturdy gust]

Per the wiki link directly above that. The lower amounts are there as you may be able to get away with them (depending on whether you have server streaming enabled or not), but you shouldn't use them

[midnight harness]

https://github.com/Jayy001/PyFactoryBridge
This one? Or could you link it? :)

[frosty gale]

yea that one

[midnight harness]

I bumped them to 32GB just to be sure anyway. But yeah 6GB seems really harsh

[sturdy gust]

32GB is overkill, but that's great 😄

[quick mist]

there's no kill like overkill

[midnight harness]

Meh, DDR4 ECC has gotten "cheap" :D

[quick mist]

33554432 Kibibytes ought to be enough for anybody.

(:

[frosty gale]

you should add examples for upload and download save files

[quick mist]

Soon ™

[midnight harness]

Oh yeah, regarding that. Has anyone multipart requests working already?

[frosty gale]

ive done download save

thats easy

upload kinda stinks

[midnight harness]

I will try upload save then

[green panther]

yea really not sure what the problem is

[topaz scarab]

instead of examples, why not make it usable as a commandline tool

[midnight harness]

Or better yet, a library 😍

[quick mist]

Implement in code first, make accessible tools later. The tools need backend to funciton after all or the tool is a handle with nothing attached to it

[topaz scarab]

someone already wrote a lib, untested and half baked, but it's there

[midnight harness]

You can ping a port? I thought that was ICMP

[frosty gale]

I wouldnt bother modifying his api lib though, its really not made for that

[sturdy gust]

I was gonna write a Go SDK and CLI tool, but meh, figured I didn't really have a need for it myself

[midnight harness]

If I read that correctly, all you are doing is changing the ping pattern, not the port. That doesn't work. Try nmap

[quick mist]

it is ICMP, -p just specifies a pattern to send. You need something like tcping or nc to test TCP connections.

[midnight harness]

yea :D

[quick mist]

$ nc -w 5 -z henrys.network 444 || echo 'timed out'
timed out

[frosty gale]

is there a good way to ping my network without enabling icmp echo packets or whatever

[green panther]

omg

im so silly

[midnight harness]

Might not be public

[midnight harness]

telnet / netcat? Just check if there is a port open

[green panther]

i dont think port 444 is forwarded to my server pc

although i might have forwarded the entire range

[midnight harness]

Most likely a firewall or NAT issue currently, yeah

[quick mist]

I prefer netcat as it doesn't rely on a telnet handshake but telnet is often used for testing TCP connectivity, yes.

[green panther]

well i get home in 30 mins so ill fix it then

any way to connect through port 443 for example but tell the server pc that im requesting port 444?

[quick mist]

some of it is.

$ nc -w 5 -z henrys.network 443 || echo 'timed out' && echo 'Gibson hax0red successfully'
Gibson hax0red successfully

[midnight harness]

Fair enough

[topaz scarab]

yeah i had a basic no-password auth and claim server implemented in mine, but realistically it's not getting used

[green panther]

i think i did something similar setting up my plex server remotely

[frosty gale]

ok so my game ping is like 80ms and my network ping is like 25ms

[topaz scarab]

that's ridiculous for local

[frosty gale]

im 150 miles away

[midnight harness]

All I want is detect if I'm done playing locally, then start a remote server, upload the save using the API and let it continue running there.

When I start the game from Steam, download the save first and shutdown the server

Should be easy enough

[topaz scarab]

okay that makes it better

[topaz scarab]

sounds doable, prob have to run it before you start a local game, or i guess keep it running if your computer's on

[frosty gale]

im actually 114.74 miles point-to-point

but you get the idea

[midnight harness]

prob have to run it before you start a local game
Sorry, what do you mean?

[quick mist]

since the game auto-saves every 10 minutes, you could probably set up something with inotify on WSL on your game client machine that uploads your auto-saves in effective real-time to the server.

[topaz scarab]

live backups, nice

[midnight harness]

I mean scp also works, or what would be the benefit?

[frosty gale]

secure contain protect

[quick mist]

inotify would tell the program "lookie, a new save", what you do with it (SCP, use the API) is up to you

[midnight harness]

Ahh, get it

Or I just DLL inject XD

[topaz scarab]

VACd
jk

[midnight harness]

hook into the save call 😎

Nah, I just wait for the process list to no longer say FactoryGame with polling. Good enough

[midnight harness]

Will try to fork this and add save uploads

[topaz scarab]

is this jay in here anyway

[frosty gale]

yes

[topaz scarab]

oh is that you
hi

[midnight harness]

oh you are jayy001?

[mental arrow]

is there server for new palyers

that i can join?

[frosty gale]

@ancient bison

[topaz scarab]

#1201555265942724758, careful of bad servers tho

[mental arrow]

bad servers wym?

[topaz scarab]

wait i knew that

dunno, i saw some drama the other day about some guy getting kicked

might've been griefing idk

[frosty gale]

you cant kick people

[mental arrow]

damn

i just downloaded the game and wanted to play it

but not by myself

[ancient bison]

hi thats me

[topaz scarab]

maybe changed the password on them, didn't exactly read it

[frosty gale]

photon engine is like this as well

[ancient bison]

how can i help

[frosty gale]

like who makes a game networking backend that doesnt support player removal

[quick mist]

I mean it's not like SF is a PvP game prone to griefing

[midnight harness]

Didn't want to bother you :D
Did you try save transfers already?

[topaz scarab]

don't you all have phones friends?

[unreal crystal]

every game is PVP when multiplayer is enabled

[quick mist]

and any server admin worth eir salt would have a password given out to a vetted group of individuals

[topaz scarab]

and backups, and backups of backups

game servers are very home gamer tho

[frosty gale]

you cant excuse shitty feature sets with 'well that should never happen so we dont need it'

average api developer

[quick mist]

I'm not saying it's odd that it's not there

but I can see how it was missed

[topaz scarab]

it's because they had snutt code it

[quick mist]

wasn't multiplayer being a thing at all a thing that came into play rather late in SF's development? Or was it planned from the start like pipes and they were just being coy?

[frosty gale]

it wasnt in early releases but it was added pretty quickly

unless we have different definitions of late

[snow saddle]

Are we able to get suggestions for places to purchase a dedicated server here?

[midnight harness]

I would really like endpoints instead of {"function": "..."}, but I guess that's a nitpick

[topaz scarab]

"purchase"

[frosty gale]

and yea that

api dont really work

[topaz scarab]

that would've been real rest

can't have that

[quick mist]

@snow saddle my suggestion would always be to get your own server like a VPS or a homelab and manage it yourself. Cheaper in the long run and you can learn a valuable skillset.

[unreal crystal]

i've no problems with my OVH box the past few years

[quick mist]

I'd have preferred that too TBH

[ancient bison]

hi sorry, no, would you mind opening an issue on the repo if you haven't done?

[topaz scarab]

api docs aren't even correct anyway

[ancient bison]

but it shouldn't be to hard, just need to change the content type

[quick mist]

how do you mean?

[unreal crystal]

OVH or that one german company if you're EU based. Hetzner or w/e

[snow saddle]

For the $20, I'd rather spend my time elsewhere.

[ancient bison]

agreed

[midnight harness]

multipart requests scare me with the amount of faulty errors I've seen so far xd

[frosty gale]

you gotta do a lot of things since your request method relies on the responses being only json

[midnight harness]

Uh sure. I will give it a try, might update you later

[topaz scarab]

upper/lowercase letters on parameter names here and there
haven't tested them all, but on the two i have they were

[midnight harness]

If you are cool with uploading your ID, sure

[ancient bison]

yeah fair, that's why I was wary of putting it in initially

[midnight harness]

I feel like that is not a valid or fun option for most people. Clicking a few buttons on a website and just playing can be very much preferrable.

I couldn't even get some of my friends to download a Minecraft mod pack.

[topaz scarab]

as long as you're not rolling your own, i don't see the concern of using multipart requests

[midnight harness]

I will cite that in a few hours

[topaz scarab]

plus it's python, there's 100% a lib for handling that

[ancient bison]

tell you what, after I've finished with this game I'll have an attempt

will let you know how it goes

[ancient bison]

requests yeah

[quick mist]

And that's is why it's my suggestion which is informed by my opinion and not an objectively correct answer. DIfferent strokes, and all.

[topaz scarab]

natively supported in requests, even better

[unreal crystal]

😭 the hardest part of spinning up a new server is always saving the password to the manager for the new user.

[topaz scarab]

password.txt user here

[unreal crystal]

only for the server steam login

[midnight harness]

Sorry, didn't wanna step on your toes :D
I'm totally on your side

[quick mist]

😄

[granite goblet]

i think i might be experiencing the exact same issue sadly

[quick mist]

All good sir or ma'am or other!

[midnight harness]

I never could get KeePassXC to sync correctly across devices.

Vaultwarden sucks in comparison, but at least it doesn't cause corruption xd

[unreal crystal]

and sometimes i miscategorize them and they get lost. its awful

[quick mist]

I use SyncThing; works like a charm for me

[midnight harness]

Especially on iOS and Android clients, they would always interfere / not sync fast enough

I really like the DB-approach of Bitwarden, and not just one monolithic file

[quick mist]

For Mobile I use KeePass2Droid and have it SSH in to a known location that SyncThing pushes and pulls updates to and from

For iOS I use a disgusted shake of my head.

[midnight harness]

Sorry, off-topic, might wanna move it xd

[unreal crystal]

vaultwarden + bitwarden on errything; no problems

[midnight harness]

this

but I sure do miss autotype

[quick mist]

I love problems that have many solutions that work for many people. Choices are a Good Thing and there should be more of them.

[midnight harness]

and I hate the UI design of the browser extension

[unreal crystal]

then tailscale on your devices with your gaming dedi server as your remote signing approval. or headscale to all in hosting yourself

then your vaults never exposed to da internets

[placid night]

Are dedicated servers backed up to the cloud? If not, what directory do I have to backup?

[topaz scarab]

https://satisfactory.wiki.gg/wiki/Dedicated_servers/Configuration_files#Overview

[placid night]

ty ty

[unreal crystal]

i just use linuxgsm and do ./sfserver backup

so; i recommend that method.

[midnight harness]

I like wireguard. Tailscale is too much fluff imo

[quick mist]

Not to the cloud. The saves are in ~/.config/Epic/FactoryGame/Saved/SaveGames (Linux) or a similar path under %LocalAppData% (Windows)

[topaz scarab]

feel free to help document the windows ones

[midnight harness]

But seriously it should be a crime that toasts cover the entire home bar.

And long, unskippable animations are hurting my soul

[placid night]

the SSD health on my NUC is showing bad S.M.A.R.T. status, so I am a bit worried lol

[unreal crystal]

bad or just hexadecimal?

[quick mist]

0xDEADBEEF

[simple pagoda]

Hiya, how can I limit the amount of RAM a Satisfactory server uses?

[quick mist]

ulimit

[midnight harness]

0xDEADED

[simple pagoda]

so ulimit 16 would be 16gb then e.g.?

[midnight harness]

Depends on your setup. If you are using the docker container by wolveix, you can just specify that in the docker-compose

[quick mist]

ulimit -v 16000000 && ./FactoryGame.sh

[simple pagoda]

Ye using the docker container

[midnight harness]

This should work

[sturdy gust]

😎

[simple pagoda]

Where is that FactoryGame.sh stored?

[frosty gale]

ok file uploads are kinda easy I guess

[simple pagoda]

Sorry still kinda new to using docker but willin to learn :D

[simple pagoda]

Where exactly is that file stored?

[topaz scarab]

no way

[frosty gale]

thats a lotta boilerplate man

[grim mica]

By limiting memory - going to force the instance to swap pages to disk - cant imagine that would be very performant

[midnight harness]

How are you running your docker container?

[midnight harness]

Wait, why would that happen?

[simple pagoda]

Well to be precise, a friend of mine is running said docker container and trying to find out how to limit the RAM usage. As I said, I'm very new to docker, if not to say I'm very much a noob at it.

[grim mica]

Not sure I understand.. if you are limiting memory to 16GB and the application requests more than 16GB, what happens?
It will swap to pagefile and use disk as memory.

[simple pagoda]

I'm just trying to help him by asking for him over here since we wanna play Satisfactory together :_D

[simple pagoda]

So if this is a command he can just punch in somewhere, do enlighten me what directory that'd be in, couldn't find that .sh file with find

[limpid rampart]

Hello everybody! I just arrive 🙂

[topaz scarab]

you're using docker, so disregard that advice

and use that setting in the docker image instead

[midnight harness]

The process should see 16GB as max virtual memory, doesn't it?

[limpid rampart]

sorry for stupid question - how can I looking for -- the stuff -- for windows -- that can monitoring dedicated server?

[simple pagoda]

Ok where is that setting? :_D

[deep turret]

python "stuff" is cross-platform

[midnight harness]

If your friend is using docker compose, it should be placed in the compose file.

If they are running it using docker-cli directly, you could use

--memory=16G --memory-reservation=4G

[simple pagoda]

Ok thanks he'll try :)

[midnight harness]

I highly recommend using the docker compose file for your setup tho