#dedicated-servers
1 messages Β· Page 123 of 1
Their service is a black box. I wouldn't know what they run to detect ddos attacks. It's probably misclassifying satisfactory connections
Run a server with a single client on it. Run a full packet dump on client and server simultaneously. Play until you disconnect. Send both sides of the packet dump to cosmicguard. Say connections are being dropped and to fix it
It's currently holding up fine, haven't had a drop in over 2 hours now. I know of a couple resellers that use their network and run into the same support issues as I have. Their support isn't the best, but the network is pretty good. I will run a pcap on it later tonight, with it's own server, nothing on it except what satisfactory needs to run, have it configured the same way as I do now and see what happens. At the moment I am running a routeros system on a reseller that forwards the connections to our production server, it's cheaper operating costs for a slightly higher latency but it makes it like $350 US cheaper a month to operate. I know sometimes latency can cause slight issues in things, we also have a wireguard filter enabled with cosmicguard's reseller too. We have connections going over GRE to our production servers and have routeros handle the networking over essentially everything. The upstream is possibly where we could have a point of failure but we may also have it in routeros so I will triple check what I have at the moment then start the ticketing process.
If there is any newer updated pterodactyl egg then please share the link. Thank you.
Strange. I've been running dedicated serveres for almost 2 month now without any crash. 1 instance on Ubuntu and Docker. And 1 instance on Ubuntu VM.
But I do notice users reporting old content (saves) make server crash.
I'm not overly bothered, just sharing for the LOLs
The second attempt to join the server crashed the client, I submitted that.
The third attempt to join after the client restarted was successful.
I can live with that.
It's nice to know that we now have reliable buffer overflows :0)
Will the server work if i only have one open port?
I'm not sure what that repository is but I believe the original was superseded by https://github.com/pelican-eggs/games-steamcmd/tree/main/satisfactory
You need two open ports
- Port 7777 UDP
- Port 7777 TCP
The port number assigned to both TCP and UDP must be the same as they are passed with a single-Port=parameter
I mean one port, as its both tcp and udp for me. What im asking is are those query port and whatever 15700 needed?
Query ports are no longer needed. Everything is done over a single port but over both protocols. 15000 range I believe are steam ports but from my testing, they aren't needed since SF uses epic API?
RTFM!
its 10010 for me so that wasnt the question
I meant "can the game work using ONE (both tcp and udp) port only"
But i see yes
If my reply to you don't give you the answere... I'm not sure what will give you the answere then...
Bc this is probably old info
Who said it didnt give me the answer
Because you replied with the same question once more.
I clarified what i meant.
I had no problem in understanding your question and my reply was accurate.
But again. Read the manual. Link is at the top of this channel.
But here is the link to the manual.
https://satisfactory.wiki.gg/wiki/Dedicated_servers
No. You should probably follow the guidance on the official wiki, not out-of-date sources on the internet π
Technically that's two ports because TCP and UDP have nothing to do with eachother except for being IP-based
it just incidentally has the same numbers
If this is a pterodactyl egg, you are using a severely outdated one. I suggest using one that is up to date, i.e. #dedicated-servers message
The old ones are known for overwriting your configuration from the server manager on every restart
no up to date distributions should ask you for the server query or beacon ports
I think i used this one, not sure tho
Nope, that doesn't contain environment variables for the beacon or query ports at all
Well it's still a single ALLOW rule in ufw)
which, funnily enough, maps to two rules in the firewall software ufw is configuring π
sudo iptables -L | grep 7777
UFW is basically a shortcut of IPTables if im not mistaken. It just translates easy to remember commands into IPTables
sudo ufw allow 7777 comment 'Satisfactory Dedicated Server' will by default will allow 7777 UDP and TCP to listen on all adapters
hi can someone help i setup dedicated steam server with port forwarding but when ppl try connect from outside the network we get this
Random-ish Q: is there anything I can do about the fact that the assault rifle is basically unusable with high latency, beyond posting about it on the QA site and hoping they fix what is, ultimately, a relatively niche issue? In particular, the smart cartridges just... don't shoot bullets most of the time with >300ms ping. I can empty the entire clip at a basic spitter or something to no effect at all π
300 ping? π Antarctica?
Europe -> Australia π
There are 2-3 players down under, and just me in Europe, so I get to deal with the pain π
Might want to host in ~India or something, everyone connecting at 150ms would be better π good luck mate
Yeah, I have considered that, but curious if there are other solutions.
hum, you might want to check logs, but it does show up in the server viewer and I was able to connect
thank you for checking
def your client or other client related. Server is fine. I think adjusting client side "network Quality" to match the server's set value "network quality" might help
thank you
Hello, dedicated server is crashing (No mods, tried on both Linux/Windows), on Windows the error message is: [2024.10.18-18.29.25:943][917]LogWindows: Error: Unhandled Exception: EXCEPTION_ACCESS_VIOLATION reading address 0x0000000300000018, any ideas?
heyooooo
maybe im just stupid, do i need to wait?
Your server is one version behind you, need to talk to the guy that runs the server to update it.
i did
still the same :c
steamcmd +login anonymous +app_update 1690800 validate +quit
right command?
Yes, if that command worked, you can check the server logs, it should say something like: LogInit: Net CL: 372858 (correct version)
where are you all getting this from anyway
+force install_dir?
it's in the wiki
You only need force install dir if you are wanting to place the install of the DS somewhere specific
Update process is:
Stop the server process
Run steamcmd (as you have it above)
Start the server process
Run Steamcmd the same way as when you originally installed it and it will just update, adding force install dir will make it download it all again to the directory specified
sure, but they've updated the server, and starting an obv different older install
You'll have to post the remainder of the crash message, not just its memory address
How was that derived? The single SS showing version mismatch doesnt say anything about where the DS install is located on the machine, adding that will just add confusion if it wasnt being used previously. I personally dont use it and made a directory at /home/steam and let it build whatever folders it wanted.. who cares what directory its in as long as you can find it to start the process/service
meh, idk where the command is
the history is too long
Do you have it installed inside the steamcmd directory or just some folder you chose yourself somewhere?
because it doesn't matter (to us) where it's installed, it's up to them to get the dir correct
/SatisFactoryDedicatedServer in my home directory
smth like that
jesus do i really need to study the whole documentary just for updating the server π
No, it's the exact same command as the install command you used. if it's in your home directory like that it will be
steamcmd +force_install_dir $HOME/SatisfactoryDedicatedServer +login anonymous +app_update 1690800 validate +quit
:o its downloading
thank u
My point is that adding the force directory when it wasnt used during initial install will just cause confusion.
sure but in this case it obv was
everything is fine, i got my update now :D
the annoying part then is to have users figure out where it installed, because some annoying distributions (cough debian/ubuntu cough) change the default steamcmd dir
Was it? The original message said "steamcmd +login anonymous +app_update 1690800 validate +quit" which is a perfectly valid command to run as well
if u want to bully me rn, im runnin the server in a screen session on debian lmao
happy
original message was version mismatch showing the server running an older version, and then that steamcmd line
and they said they ran it again, no change
fyi MuffinAmor: you probably created a second fresh server install when you ran the steamcmd command without the force_install_dir parameter in its default location which is a hidden folder which differs depending upon the operating system/distribution you use
hidden?
ah dang, well than it will lying there
Not about bullying.. just pointing folks to the wiki that has specific NEW instance installation instructions isnt helpful, it has updating instructions but this does indeed require the user to know where the install was actually laid down. Again - adding a force install command when you dont know where the original install is located is just adding to the confusion
but we do know that the instance they're running it from isn't the default location
because the server is still showing the old version number after updating
where they installed it is up to them to figure out, and they'll want that force install dir
am i taking crazy pills here
How can i give my friend admin on my steam server?
give him the password
Default on linux is $HOME/.local/share/Steam/steamapps/common/SatisfactoryDedicatedServer where .local is a hidden folder;
debian overrides this to $HOME/.steam/debian-installation/steamapps/common/SatisfactoryDedicatedServer where .steam is a hidden folder
okay that's weird because my debian install defaults it to ~/Steam/steamapps/common/SatisfactoryDedicatedServer
yay more inconsistency #joy /s
I do the same, Debian gang best gang
making it a service isn't hard, you guys should do it imo
so what, you do like
while :; do ./FactoryServer.sh; done
in GNU screen?
I mean it works, but I am getting le shivers
Is it possible to transfer a local game to a dedicated server later on?
Yes, you can upload it in the in-game server manager.
π
Note that saves that are in the tutorial will automatically skip past it
screen -S sfserver sh FactoryGame.sh so I can detatch from my Debian VM GUI and switch between my WSL SSH session
prob should just make it a service tho, but still testing modlist
π but it won't even restart if it crashes
nice
i know, i take the risk, too lazy to setup systemd for this
its not critical so....
true, I might just dockerize my instance once I get my modlist and DS compatible mods working. Just working backwards with a 100+ mod list because I'm an idiot and didn't do one at a time from the start
yea I only turn things into a service once I'm done configuring and stabalizing instances.
Liek my old Insurgency Sandstorm servers
a minimal systemd service is eight lines, including the three section headers you need xD
π i could setup a script that doin this for me, but hella no
i prefer to stick 30 mins around to fix some display issue in a ccs
yea might just keep it baremetal since SMM/FICSIT CLI requires SFTP and I don't wanna deal with docker container ports and networking.
you don't want to install ssh in the container anyway. the default docker bridge network automatically bypasses your firewall (on linux)
on the other hand, the server literally saves to the local filesystem since it mounts the install location anyway (if you follow instructions) so π€·ββοΈ
(also docker is root by default so back to shuddering)
groot? :O
yea true, I also don't want to deal with mounting and permissions lmao. baremetal it is, and good ole .bashrc alias for now
-# eight lines for systemd
On an ubuntu based os it took me literally like 5 minutes
So y'all should do it if you have the option
Is it that bad to run on a root user if your chosen port isn't even in the same port range as default
I'm not running it as root but I'm just curious
is it bad to run a random game server which is not guaranteed to be 100% secure as the superadministrator of your computer that can do absolutely everything including write to your motherboard nvram?
probably not the best idea, no
Fair enough
This computer is literally only for servers so I don't really care about the machine but still
The main issue is network access, but I blocked all incoming and outgoing ports except for the ports I'm using for the server
meanwhile, router/firewall password is "password"
Nah we actually change our passwords
PaSsWoRd
I generally use a different account for each game I host so I don't flood a single "steam" user with a ton of different server installs.
at one point I was running a project zomboid, insurgency sandstorm and valheim server on my NAS and it got messy real fast
apparently the Mk 6 belts are not transporting 1200 since here the miner starts to fill up with even 1185 items/min. The belt goes straight into a sink since we have the problem that our rocket fuel plant was lacking sulfur and the belt were should be 1200 but the blenders are starving. Around 1175/min is a sweet spot were the miner sometimes fill up to 4 or 5 items, this could be the lack between client and server but for 1185/min the fill up is really noticeably let alone 1200/min. The aren't many bug reports about it and wasn't mentioned in the lat patch notes :/
hell yea got 104 mods running and working on dedicated server. Now to test stability before I turn it into a service π
So after setting up the server as a service it completely broke it
It shows as running but also shows offline to the game client
Windows or Linux service?
Both ports are open
Ubuntu
Whats systemctl show for status?
Active running
is there any chance you have the hover pack equipped?
if yes, can you unequip it for like five full minutes and see if it reaches 1200/min?
It shows basically what the one on the wiki shows
(and other players who may have it equipped)
sudo netstat -tulpn | grep 7777
change 7777 to listen port if you changed port
the listener for tcp and the listener for udp
0.0.0.0:[port]
usually bound to 0.0.0.0 and/or ::
For both tcp and udp
if it's showing fine and you can't connect in-game, can you go to https:// <ip address> :7777/ in your client pc webbrowser?
It worked when it wasn't a service but now that it's a service it doesn't show
Exactly what it says on the wiki
what do you see if you go to that url in your browser?
It says exactly what it should
which is what?
"errorCode": "errors.com.epicgames.httpserver.route_handler_not_found",
"errorMessage": ""
}```
can you delete and re-add the entry in your server manager?
did that
what's the server name show in the server manager?
server name pending
check status of port 7777 UDP in your firewall. TCP's going through, UDP is not
I checked nmap and TCP was open while UDP was filtered
nothing's changed since last night when it was running
you have a firewall installed on the server? firewalld? ufw?
Firewalld, I can check again but I haven't touched it since last night
did you set your firewall rules with --permanent yesterday?
Ok so whenever firewalld reloads it resets the ports
I'm using the gui version so idk how tk do that on there
yes, it does that if the rules are not --permanent because then they are ephemeral π
I didn't know how to set it as permanent in the gui version lol
what gui are you using?
Linux lite
It's a desktop linux because I was originally going to use this as an emulation box but it has more use to me as a server
So it's basically just desktop ubuntu
idk the linux lite gui... in command line it'd be something like
sudo firewall-cmd --permanent --add-port=7777/tcp --zone=public
sudo firewall-cmd --permanent --add-port=7777/udp --zone=public
Up at the top it has a drop down with the options "runtime" and "permanent" that I didn't realize I had to change
It's p much just default ubuntu but super trimmed down afaik
I don't usually have GUIs on my server XD
I just didn't bother reinstalling the OS because it still has my emulators on it lol
anyway, guessing that fixed the connection issue then?
My emulator dreams of it died when I realized that a ryzen 3 4350GE isn't enough to emulate ps3 games
Yep, thanks!
trust me, I learned firewalld rules weren't persisted by default when I rebooted 1 of my new servers for the first time xD and there went my ssh access poof lmao
Yeah I'm still fairly new to linux but I'm using it more and more wherever I can
Windows just stays with me because gaming on windows is easier than gaming on linux
Some games I play aren't supported on linux or proton
So windows gets to live...for now
yay ring0 anticheats /s
The anticheat is battleye, which supports proton if they told them they wanted it to
Not rockstar, bungie
actually yes, does the hoverpack cause some bugs on servers?
whenever the hover pack switches between power connections it tends to disrupt the power grid for a single tick
if nobody has it equipped you should see full 1200/min throughput
(the belt output on the miner and input on the sink will shut off for an extremely brief moment every time the hover pack switches)
sorry if this has been answered; physical server (Win11) lost power, reboot and update GameServer and Steam game, three things happened: 1) GameServer showed a faulty cert 2) GameServer says it is unclaimed 3) All .sav files related to the loaded savefile at the time of power loss are completely missing (except the init from Sept13, set to ONBOARDING phase) based on all locations listed in the wiki and an in-depth file search.
Is this normal (enough to justify offsite .sav backups)?
btw not a problem on lower mark belts. the inventories can displace items fast enough to compensate on slower belts
you use docker ?
If you're running it as a service you may have the saves stored in a different user's or the system's folder
no
what method did you use to run the server last time prior to the outage?
it was through nssm. I looked in those relative directories per the wiki, no dice
If you ran it as system instead of as a user the saves might be somewhere along the lines of
C:\Windows\system32\config\systemprofile\AppData\Local\FactoryGame
yes, that's what I meant by "per the wiki"
Not sure what you just linked, but this is the official wiki, which is whitelisted: https://satisfactory.wiki.gg/wiki/Dedicated_servers
that plus #Loading_a_save_file
so the folder C:\Windows\system32\config\systemprofile\AppData\Local\FactoryGame doesn't exist?
it does, empty
you've started the server with nssm again and it has completely reset? and you also checked your current user's save directory?
correct
okay so this powershell is hopefully right since I'm not on windows....
Get-ChildItem -Path C:\* -Include ServerSettings*.sav -Recurse
in an admin powershell
two results: one from tonight (presumably from trying to load a save file) and one from Sept (presumably the original that I was trying to recover, as its in Windows.old)
did you upgrade your windows OS?
no, just the power failure
check C:\Windows.old\system32\config\systemprofile\AppData\Local\FactoryGame
does that have any of the recent saves in it?
there are unrelated saves, from a single-player run
you can try
Get-ChildItem -Path C:\* -Include *.sav -Recurse
to find all savefiles, though you might find savefiles from other games that also end on '.sav'
it's probably from an earlier server run since you can't run the game client in singleplayer as localsystem unless you did something really weird
same results, missing the important one
did it finish running?
I think I loaded it into the server back in Sept to test 1.0, you are probably right
yes
I'm gonna conclude that I should expect to make a backup remotely, and I'll attempt to recover from the disc in case it was something with the power loss that corrupted data
and maybe stop running the game server as localsystem and instead run it as a normal user
thank you for the help
that way your folders don't end up in C:\Windows
You can configure NSSM to not run as localsystem by setting the correct objectname parameter, but I'm not an expert
I'll look into it
in the UI it would probably look something like this, you'd check "this account" instead
yeah and I can manually set the service to run as "this account" like in your image
thanks again βοΈ
good luck!
There is an issue with the hover pack if you have your grid seperated into two with switches and you switch over with the hover pack, all miners in that grid stop for 3 seconds I believe
can you test that and report back? there is an issue somewhere, but I am too lazy to find
Anyone else got that annoying bug that's been happening on dedicated servers since update 8 where the game disconnects you when you enter tractors?
Better than that... Game clipped me out of bounds when I tried logging back in.
why
Does anyone have any guidance on connecting to dedicated server externally? I just did a fresh install of Debian, server is running and can access from inside my network but says "offline" when viewed externally. Port 7777 is being forwarded and verified with a port forwarding tester. When looking at what ports the server was actually using there were a couple others and I forwarded those too, still no connection. Thanks in advance!
Its 7777 TCP and UDP, no other ports needed - both types are forwarded?
Yup both are forwarded
You can always use a convenient interface for server installation.
https://github.com/asidsx/SatisfactoryPyServerInstaller
Hello all ! I have a weird beaviour on my dedicated server on linux, I don't know why but now when I start the server (Satisfactory server) it makes my Linux Crash ! Anyone has/had the same issue ? It worked well for 800+ hours. I Have that red line in the log : LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR - I don't know if it's THE cause
I'll test it later when I am alone on our server.
So all the sudden I'm getting a bug where a ton of the folliage is coming back after it was fully destroyed. The little tress, stumps and stuff show up as if they are visually there, but you can walk thru them causing tons of rubber banding, and when you try to blow them up or otherwise nothing happens cause they are not actually there.
Has anyone else had the issue? Any ideas for potential fixes? A ton of trees I destroyed just popped back into the world, straight thru my factory
Server just restarted and its all still glitching out like I described above
Yeah it seems like basically all the trees/folliage around this area are glitching out like this
All of this is glitched
If too much foliage is destroyed in an area the game netcode fails to correctly communicate what foliage is destroyed to game clients
The foliage is still gone on the server. You can even walk through it
Yeah it just makes you rubberband to shit, anyone know of a fix?
Destroy less in a given area and hope it doesn't happen
An actual fix would need to come from the developers
Someone claimed it occurs less quickly on medium network quality settings than on ultra
Interesting, thanks for the info, hopefully the devs apply some sort of recursive fix
The server/save file is fine. It's a network code issue. If you load the save in single player you'll see what the server sees: foliage is still destroyed
That's why the server lets you walk through it c:
Yeah that part makes sense :p
Unhandled Exception: SIGSEGV: invalid attempt to write memory at address 0x0000000000000003
Assertion failure. please post the backtrace
Anyway without the backtrace it could literally be anything
The server was only using like 7.5G ram at the time of the crash.
were you/a player doing something with a manufacturer during the crash?
So we're supposed to document what every single player is doing on a server every second someone plays in order for satisfactory to run properly? I don't know, they're just playing the server. No mods nothing out of the norm.
No, but it may have been useful in identifying why the crash occurred. If the crash doesn't occur again you're probably fine either way
This is what we get when we crash randomly.
There are some known issues which can cause consistent game crashes which happen more easily in multiplayer/servers
We've crashed with this error multiple times now.
specifically with Assertion failed: Pair != nullptr [File:Runtime/Core/Public\Containers/Map.h]?
if it's always on a AFGBuildableManufacturer you probably have a manufacturer with bad save data
If you want and don't mind you can DM me your save file and I can take a look
How do you upload blueprints from the web onto a dedicated server? i'm hosting on my computer. i can't seem to find a solution on it because the savedgames/blueprints folder doesn't show up in my server files. i've tried creating a blueprint folder manually as well as making a blueprint ingame. any help is appreciated!
windows or linux server?
Windows
%LOCALAPPDATA%\FactoryGame\Saved\SaveGames\blueprints\SessionNameHere\
Sent.
you may need to create one blueprint in-game to have the server create the folder
note: server may need to be restarted (or save reloaded) to load new blueprint files
π
Has anyone seen this issue? Was finally able to see my server externally but still getting this error and cannot connect. Seems to work fine on LAN
I do get this error when doing https:// IP : 7777/
This is good, means it's working fine since there's no route handler for the root path
are you connecting to your WAN IP from outside the network or is this from one of your users?
I am technically on the same LAN and can access it just fine from there (putting 192.168.xxx.xxx into the server manager) the issue is when I use my external IP (47.150.xxx.xxx) then I get the issue shown in the picture (bottom server is local IP, top server is external IP)
If you are trying to connect to your server that is inside the same LAN as you through the WAN IP you need to correctly configure NAT reflection, which is advanced material beyond regular port forwarding, and usually not enabled by default (and sometimes not available at all) on consumer grade routers.
Is there a reason why you are trying to connect to your server via the WAN IP?
just to test that the ports were all setup correctly, me and a buddy have the same server and his attempt failed to see anything and then mine was failing until this issue just now. But alright I can try to get a friend to see if they can connect. If that is the only issue then that would be great!
if you correctly forwarded port 7777 for TCP and port 7777 for UDP it should normally work (if you configured the server for the default ports, and the server's firewall is correctly configured too)
ok ill try it truly external, I always tried that on different servers and it worked fine (minecraft/factorio)
ok, the server is debian and from what I read ports are enabled by default. I ran a couple lines to tell it to open 7777 on udp and tcp anyways. pretty new to all this stuff!
changed my game to IP hosting, but telnet localhost 7777 failed
Connecting To localhost...Could not open connection to the host, on port 7777: Connect failed
If the wiki for regular multiplayer is up to date, port 7777 TCP is not used so I'm not sure why you're trying to connect to it.
I only see a listener on 7777 UDP with IP multiplayer so you won't be able to connect to it with telnet.
What are you trying to achieve?
That's for installing dedicated servers, but your earlier message mentioned you loaded a game in the game client and changed the multiplayer session type to IP.
Which one are you trying to do?
If you've effectively launched the dedicated server with its default settings, you should be able to connect to it on port 7777
Are you launching FactoryGame.exe in the dedicated server installation directory and seeing the server console?
I have my server running in unraid on a docker. I get the following error in mod manager when I try to map tthe path...
Failed to list directory
failed to get partitions: Error 0: Access is denied. Error 1: The network name cannot be found.
\TOWER\appdata\satisfactory-server1
#1243688087578738770 ask on the modding discord. this seems to be a question about the mod manager, not dedicated servers
My friend tried enter my public ip
In the server manager?
Yes
Did you configure the firewall and port forwarding as described on the official wiki?
My friend can test it with curl -k https:// . . . :7777 ?
For the tcp port forward yes. There's no easy way to test the udp port forward
He should be expecting {"errorCode": "errors.com.epicgames.httpserver.route_handler_not_found", "errorMessage": ""}, correct?
Yes
Are you sure because I'm not sure server manager can throw this error until after you join
I installed the dedicated server directly from Steam, it seems to be working but crashes every now and then, especially when dealing with vehicles and paths. Is the Steam dedicated server tool an issue or is that just everyone?
Anyone know how to open the command console? keybind isn't working and setting it in the client config doesn't seem to change anything
You need to edit the client config while the game isn't running. What keyboard layout do you have configured?
And what are you trying to set as shortcut
Is the server crashing or is your game client crashing?
Just the server
US keyboard layout, i have 'ConsoleKeys=F6' in my input.ini file
Can you share the latest crash log?
Yes where does it live?
You're trying to open the game command console, not the server's, right?
Depends. What operating system are you running the server on?
Correct, for things like turning off fog on the fly
Windows 11
if your F-keys are shared with media keys, make sure you use fn as appropriate for your keyboard device. are you sure you quit the game before editing the Input.ini file in the folder %LOCALAPPDATA%\FactoryGame\Saved\Config\Windows\?
Note, you may have an old WindowsNoEditor folder if you played in early access, this is no longer used
%LOCALAPPDATA%\FactoryGame\Saved\Crashes
do you want the FactoryGame.log (4.4mb) or the UEMinidump.dmp (710kb)
Solved, turns out my modern blocked 7777
the FactoryGame.log, at minimum from the "Critical Error" line onwards to the end if you don't want to send the whole thing
I double-checked, i don't have a WindowsNoEditor folder, i'm in %localappdata%\FactoryGame\Saved\Config\Windows\input.ini and i've changed the key to '0' to avoid any fn or media key issues. Still no luck. I have this at the end of the input.ini file
[/script/engine.inputsettings]
ConsoleKey=0
Actually could not find Critical Error in the file. I scrubbed my brothers IP out with xx.xx's I am the internal 10.0.. IP
the section name is [/Script/Engine.InputSettings]. I'm not sure if it's case sensitive
in addition, the config key is ConsoleKeys, including the last s
so
[/Script/Engine.InputSettings]
ConsoleKeys=F6
looks like opening/closing the game reformats the file and removes the custom section, since that header is at the top of the file anyway. I'll double check the 's' at the end
yes, the game overwrites the configuration on quit. this is why I'm stressing to only edit the file while the game is closed
Right. Closed game, set console key, save/close file, then launch game. so far, no luck
if you close the game again, is the section with the ConsoleKeys config change still in there?
yes, I can dm you the input.ini file if you'd like to double check? Doesn't look like it's working on local games either so it must not be tied to the dedicated server π€
Oh wait, i just got it to work on a local game
you're free to share it here or to DM it to me. it's strange you can't get the console to open on any key since it works fine in both single player and on dedicated servers for me using the ~ key on an ISO keyboard, and I've seen other Europeans with wilder keyboard layouts get it to work with ConsoleKeys changes
also everything you can temporarily configure in the console you can permanently configure in the ini files
Ah, it's working now. I must have been missing the 's' when the keybind was set to F6. then in changing the keybind to '0' and adding the 's' to then config correctly, i think there was keybind conflict since 0 is technically bound to the hot bar by default. Changing it back to F6 and having corrected the missing 's', it's now working on with the dedi server too. Thanks for helping me realize my mistake π€¦ββοΈ
no problem, glad I could help π good luck on Massage-2(AB)b, pioneer!
π there is no crash in that log. Can you send the accompanying CrashContext.runtime-xml file please?
Yes thanks!
is this save new in 1.0 or is it an older Early Access save you've continued?
new in 1.0, in fact I had not even installed or played Early Access
can you stop the server and verify the game files (of the server)?
Using the Steam verify ?
Not that I've seen. Sometimes they are colliding with other vehicles (we're sorting that out still lol)
My friend crashed
I am just guessing at this point, but this crash looks like a bug in processing the vehicle colliding with another actor it can interact with so it may be caused by those
Verification has completed. 780 files success validated
oh we are building a race track, currently it had 5 factory carts, we planned to do over 100. We're short on fuel. Is this a bad idea?
I've not seen that happen before. His game crashed while ticking the HUB
I can't think of anything involving the HUB that would lead to that xD
If this leads to the server crashing frequently, I would probably consider it a bad idea for the time being π
does sound like incredibly fun if it did work xD
@mild quiver you may want to leave feedback on the QA site, including the impacted save file and the full crash report https://questions.satisfactorygame.com/
HAHA ok - we even named it Daytona and had started the grand stands and lights π we can convert it to a train thanks for the look at the logs today.
You're welcome, pioneer. Good luck on Massage-2(AB)b!
hello , i'm new here , trying set dedicated server for more friends more than 4 defaultly set , can someone help me how to increase max player limit?
I came here to say I am giving up on Satisfactory. Its just too buggy. I setup a server for me and my 5 friends, were having a great time except our clients keep crashing and when we log back in we are new characters and our stuff is almost always lost. we have to get all new invintories, reset our color and remake all our utility hot bars, its just not fun. Also, as a professional programmer, i cannot understand how this is a hard bug to solve. So while we love this game, the lack of stability is ruining it for us. I was hoping 1.0 would come with better network code but sadly its just not ready for serious coop. im really sad to post this and i hope you keep working on making coop playable. ill give it another shot if there is ever another major release. If I could ask for a refund after 81 hours of gameplay I would. I think we deserve it given how buggy the game is. π¦
please read the wiki?
telling us this here is talking into the void
afaik your intended audience doesn't read this discord
try a ping or ten tho
okay, thanks for letting me know
I agree the respawm as an other character is very annoying yet should not be so hard to fix if the connection tracking code was more mature
though, I don't know if you reached here before ever
since we usually (me and other community members) advice to make sure to not reconnect too soon to server after any crash so you stay on your old player body
as well as providing links to tools to go around some of the more permanent issues like inventory recollection (maybe it would be cool to add a nospoil mode and inventory recover tool in scim)
which make me feel like you also didn't really try the workarounds that are making the annoying things .. much less annoying (yet still frickimg annoying but with much less impact on gameplay)
I really appreciate your comment @deep turret Thank you. I am open to using mods to make the side effect of this bug more managable. i dont want to cheat but i also dont want to waste time getting my invintory back.
I didnt realize you could wait to reconnect and avoid this issue. how long do you wait? is this written up in the wiki?
Youre right i have not tried the workarounds. i just had it happen twice in two hours of game play this morning and basically i rage quit. ill give these workarounds an honest try. Thank you again for replying to me
the thing happening is the old connection to server is still active
and as the server is not allowed to cancel an ongoing connection to give the select player lead to an [other] client
it will create a new character for it instead
the workaround is to wait for connection timeout time, when server will notices the [old] connection is broken and shut down
then connecting will not overlap with an ongoing connection
and so the server will be able to give back control of the then [offline] character (your character)
and voilΓ
it sounds like if i restart the server it should also prevent this
timeout can take some time
you can usually wait a whole minute after crash (default timeout is 30s but there may be some back and forth before actual shutdown)
yes
but that means canceling for everyone too and usually take more than 2 min to get fully back up
knowing just that might make the game playable again for me, thank you so much for the further clarification
imo the current timeout should actually be the retry/reconnect window timeout
and the actual connection timeout should hard occur within seconds
(but the client would be able to transparently reconnect until retry window timeout)
that would make the player lead report as stale within seconds too and fix the problems
(and would be accurate since client effectively lost control of the character within that window)
Are there tools today for recovering inventory?
that's a bit of a big tool for the thing but is currently the only good tool I know
it's a bit annoying to use (need full save save/reload download/upload cycle)
but give you the power
it's the save editor scim
https://satisfactory-calculator.com/interactive-map
it's a community driven tool
the recommended thing to do with crate and other equipment is to change their location to somewhere you can access in the game
it is very raw tool
but is reliable if you input the correct coordinates
for Players I think you can transfer inventories in the character editor tab
then delete the other/old characters
then you save
upload the edited save to your server and reload it (don't override the old save for safety)
Possible Solution:
send the command:
"net.MaxNumberOfAllowedTArrayChangesPerUpdate 4096" then rejoin your dedicated server.
I ran into this issue earlier this week and decided to investigate today, really hoping someone can test this out to see if it works for anyone other than just me
Did not work for me unfortunately.
Joined, typed the following in console:
net.MaxNumberOfAllowedTArrayChangesPerUpdate 4096
Left, and rejoined.
Also tried with restarting client instead of just leaving.
alrighty, there is also: "net.MaxNumberOfAllowedTArrayDeletionsPerUpdate" I had also bumped that up to 4096
also after sending the commands, send them without a value, to ensure that they have been updated
also there is a chance that 4096 is not large enough if you have destroyed enough stuff, I have no idea what the maximum might be but could try setting both to even higher than that
Issue: When too much foliage, trees, or flora are destroyed in an area, the game netcode fails to properly communicate what foliage has been removed to game clients. This can cause synchronization problems where these objects visually appear but are not actually present on the server, you can walk through these objects but it will typically cause quite a lot of rubberbanding.
Solution: Run the following commands in order, then leave the server and rejoin.
net.MaxNumberOfAllowedTArrayDeletionsPerUpdate 4096net.MaxNumberOfAllowedTArrayDeletionsPerUpdatenet.MaxNumberOfAllowedTArrayChangesPerUpdate 4096net.MaxNumberOfAllowedTArrayChangesPerUpdate
This will help the server correctly manage foliage destruction updates.
For anyone interested in how I came to this solution:
Started having this issue on my dedicated server this week and decided to dig into it today, here is what I found:
While reviewing server logs I kept seeing the following line:
[2024.10.19-19.45.27:200][ 61]LogNetFastTArray: Warning: NumChanged > GetMaxNumberOfAllowedChangesPerUpdate: 2315 > 2048. (Write)
so I decided to take a look at what "MaxNumberOfAllowedChangesPerUpdate" was, so I started typing that in the console to see what popped up and there were two commands:
"net.MaxNumberOfAllowedTArrayDeletionsPerUpdate" and
"net.MaxNumberOfAllowedTArrayChangesPerUpdate"
There was no description in the console about what these commands were for so I checked the current value for each and they were set to 2048, so I set them each to 4096, exited back to menu, then rejoined, and all of my phantom trees were gone.
To ensure this was not a fluke I set them both back to 2048, loaded back into the server, and all my phantom trees were back. Changed them back one at a time, DeletionsPerUpdate had no impact, however toggling ChangesPerUpdate between 2048 and 4096 reliably gets the phantom foliage to return and disappear.
Conclusion:
MaxNumberOfAllowedTArrayChangesPerUpdate appears to be an array used to update the client about the foliage that has been destroyed or gathered. I am assuming that most of us are running into this issue when we go to the swamp because as we start clearing that area we quickly fill that array up (Maybe? idk I didn't make this game).
After further testing it has been found that both DeletionsPerUpdate and ChangesPerUpdate need to be changed for the solution to work
Possible Solution (tested twice):
send the commands:
"net.MaxNumberOfAllowedTArrayChangesPerUpdate 4096"
"net.MaxNumberOfAllowedTArrayDeletionsPerUpdate 4096"
then verify they worked by sending them again without a value
"net.MaxNumberOfAllowedTArrayChangesPerUpdate"
"net.MaxNumberOfAllowedTArrayDeletionsPerUpdate"
then rejoin your dedicated server.
this might also work in the non-dedicated multiplayer as well, but I have not checked.
Hope this helps!
Lena thank you for being awesome and helping so many people enjoy this game! πππ
Anyone know if the restart schedule is necessary? Does it improve stability?
hey can someone help me?
whenever i leave my server it stops and my friend gets kicked off
this happens every time i leave and i need to restart the server
and i have the service thing on
known issue, supposedly fixed but not yet released
is there a temp fix?
host the server on another device or in a vm/docker
wat
is gportal or indifferentbroccoli a better provider? Anyone got experience with that ?
Scalacube is good
just get some old office hardware
run something headless
not helped. when using steamapp for host cant type in CMD.
edditing game.ini cause none changes even is set read only to nor rewrite it on start
cant find second solloution where exacly
you stopped the server first before editing, yes
yes i did
when will it be?
no clue
oh
i see
todays update seems much more stable
today?
Hello, Is there a way to increase the ReliableBufferOverflow?
(Server Side)
I have a very special Problem and can't understand why it's not working. I'm hosting the deticated server version of satisfactory inside a docker container on my home server because It cant run on kvm cpu's any more. So I installed a Vpn Server (Wireguard) on my external kvm server and an Vpn Client on my home server. With an Nginx I bind my external Server Host Port 7777 tcp and udp to my homeserver 7777 via the VPN (So to say an Reverse VPN tunnel). I tested this whole thing with an minecraft server too. There is all working. But The Satisfactory Server is only reachable in my local network and not via the VPN. So my question is what is the satisfactory deticated server doing that it is not working through the VPN. It makes no sence to me.
Since you reach your dedicated server locally there is nothing wrong with the server.
The answer to your problem lies elsewhere than this channel.
Yes but i would like to know if the server is binding the local ip ore what it is doing that the vpn is not working. So in which channel i have to go to get my answer i need.
The server binds it to 0.0.0.0:7777 which means any ip-adress you get...
I'm not sure why you want to force VPN connection to port 7777.... It's not a VPN SERVER.
Yes so far I know and it uses tcp and udp now im using an external server so my frinds can insert the ip of the external server and their connection get routed over an nginx through a vpn to my homeserver but then the server appears as offline and i dont't know what the server does to appear online. Ore what it needs to do
Your server appears offline since your network configurations/solution is wrong/isn't working.
If you have an external server, why not have the dedicated server there?
Because its an vps (Virtual Private Server / KVM) and Satisfactory needs Host CPUs and not virtual CPUs in the new 1.0 Version. My Vpn Network Connections works with a Minecraft server. So I thought it's maybe thomething diffrent in the things Satisfactory does or how it handels network traffic. So that NATing don't work as an example.
I don't know where you got this CPU info? Unless you are talking proxmox has a weird CPU setting that causes issues.
I've sarted Satisfactory on Proxmox and got this error
IF your last sentence was right then ALL would have problem with Dedicated server for Satisfactory....
I had one server running on VPS with no problem, but I used Docker. That CPU isn't reported as kvm CPU.
Tested on Proxmox where CPU is reported as kvm and server didn't started...
Changed CPU to host and ran without problem.
Change the CPU setting for the proxmox VM.
Ja that have i done on my homeserver too. But on my external vps i have no control over that settings
Strange. I just did all of this and it changed to 8 users.
Not using SteamApp. Using steamcmd....
That settings has only to do with dedicated server.... Which you have stated works fine locally...
So troubleshooting should be networking then.
not necessarily. A minecraft server has worked with the same network settings. And as long i don't know what satisfactory does diffrent i'm not able to fix the netowrk issue. Like opening diffrent other ports after first connection etc.
The problem isn't that your solution works on an Minecraft-server.
The problem isn't dedicated server to Satisfactory.
It's a fair guess that Minecraft-server and Dedicated Server for Satisfactory is not the same....
But trying to find solution to problem that don't exist.... Will never bear fruits...
Wrong folder.
"C:\satisfactory\FactoryGame\Saved\Config\WindowsServer\Game.ini"
If you installed the server in that location.
You need to connect to the server and create a game.
Or upload a game to the server.
Then edit....
we play on that server
Does it show when chosing Server Manager on Game screen?
Then shutdown the server and edit/see if the game.ini file is there.
what edit what?
game.ini
but in that path isnt there
where have you installed your server then?
i install right throu steamapp
it instal and mantain it selfs
dedicated server from app
(cant write in cmd manualy)
Easiest is then to use it like this.
The steamapp version isn't listed as an option in the wiki.
And it can be started like this too.
are the pipes having strokes a server thing or just satisfactory after 1.0 thing?
Yeah but like everything worked before 1.0. Like no efficiency issues and stuff like that. and right now to power a small fuel burn facility that takes 373 m3/min i need an imput of at least 400 m3/min (yes it has backflow preventers, bufors, and pumps (pumps are useless in theory cause its all on the same level))
I didn't play this game before 1.0.
But there was some hassle with unnatural flow within pipes yes and had to "redo" them all to be sure flow was correct.
Fuel powered generators take more fuel per min in 1.0
Before 1.0 they took only 60% of the fuel they do now. But they also only produced 150MW instead of 250Mw which is also only 60%.
So you might adjust the number of generators or their clockspeed.
Can anyone helo me with this??
IPV4 and IPV6 (in that order) and none are connecting
Check UDP firewall/port forwarding rules
They are gtg
What operating system is the server running on? Is it in your LAN? If it is, are you connecting with the internal IP?
Windows, using WAN (Public IP) and connecting with Public IP. Friends aren't able to join me using IPv4/6
That error is usually that the UDP isnt making it
Where is the server in relation to you, thats an important fact you left out of Lenas question
You're connecting using the server's WAN IP from outside the server's local network?
From inside, I am the server host
Can you connect using the server's LAN IP?
Yes, but others are not able to join
Is that screenshot from someone on the outside trying or when you use the WAN IP when adding server?
If you look in your router control panel, does the WAN IP range somewhere between 100.64.0.0 and 100.127.255.255?
You not being able to join with the WAN IP is possibly a separate issue from others being unable to join with the WAN IP, since connecting externally is regular NAT whereas doing so from inside the network is NAT reflection, which is something you can choose to figure out once we solve the basic connection issues
Not even sure where to find that
You logged into your router control panel to configure port forwarding earlier, right? It should be visible somewhere in the same control panel
In the WAN section all it says is automatic IP, no range is listed
is there a status page or something? without the make and model I can't pinpoint the specific location you'd need to look at
Can you tell me which specific port forwards you configured?
usually show on internet connectivity status pages
Hello,
can dedicated server run on RockyLinux?
Yes
Can you recommend hardware requirements?
8GB RAM, 2x CPU core, 48GB SSD will be okay?
!wikisearch Dedicated+servers
This page provides basic instructions on setting up and operating a Satisfactory dedicated server.
You may want to double that RAM since 16GB is recommended to reach the end game
Yes, apart from the physics engine, the main gameloop relies heavily on a main simulation thread
Note: if you run it in a KVM you need to configure a relatively high instruction set CPU. the default qemu/kvm64 cpu won't launch the dedicated server. Set it to "host" or manually configure it to suit your needs
thank you for tips
More cores dont hurt though.. there are a ton of background workers that can easily load ~8 cores at even just midgame, add more players and that just goes up even more
i am more prepared for this now
You may need x86-64-v3 or -v4, I believe there's an AVX requirement but I'm not entirely sure which one
good to know
there's a table here for the implemented features https://en.wikipedia.org/wiki/X86-64#Microarchitecture_levels
very knowledgeable
7777
if you're seeing failures due to SIGILL/"Illegal instruction" it'll be that
on which protocol(s)?
TCP and UDP
i am running on old 2x Intel Xeon E5-2630L v2 @ 2.40GHz; hope it will work
and the destination IP is the internal IP of your server, correct?
Next would be verifying your WAN IP in your router is your actual WAN IP to eliminate CGNAT or double NAT setup. If you're having trouble locating it in your router's software you can let me know the make and model
That would be my LAN IP?
of the machine the server is on, yes
I'll try that
hey guys, im sure this is probably laid up on here all the time, but im having a hell of a time getting my dedicated server setup here. I have done all of the port forwarding stuff, made sure windows firewalls rules are set and even tried to turn all firewalls off in troubleshooting and still cannot get a connection from online
on a second machine in the local network it works.
any ideas?
"from online" => you are trying to connect to your server's WAN IP from outside the home/local network?
Hey, can you explain how you got it working?
how do this pls i dont want to loose my save again
Having the same problem, Minecraft works fine but satisfactory wonβt. Did you find a solution?
I'm not sure how this is related to losing your saves. Do you happen to be running under NSSM in Windows?
yea
You're connecting to a machine running Docker.io through a Wireguard VPN and experiencing some sort of connection difficulty?
NSSM runs services as LocalSystem by default, which puts the saves in the systemprofile directory, which gets cleared when certain windows updates are installed. Don't run your satisfactory server as localsystem
You can retrieve your save files from the systemprofile directory to safeguard them and load them back into a server running as a regular user
Running Tailscale on a VPS and on my local server. Iβm behind CGNAT so not able to get public ip. So routing incoming traffic for player, to the vps and then that proxies over the Tailscale connection to an amp vm on my home server. Works great for Minecraft but canβt get the satisfactory server to resolve
where do i find?
If you're running your server as local system,
C:\Windows\system32\config\systemprofile\AppData\Local\FactoryGame
or
And as a regular user
%LOCALAPPDATA%\FactoryGame
ty
your VPS is doing IPTables NAT Masquerade?
what port forwards have you configured?
Yes, with Nginx as stream proxy for making the iptables configuration
UDP TCP 7777
what OS is recomended for my VM to run servers?
what do you see in-game when you try to connect?
Not at the pc at the moment but it canβt resolve. Just gets server offline information in server manager
But have previously used ZeroTier where it works, so I know the server works. Itβs just the nginx part that seems to do something wrong
... wait you're using nginx as a reverse proxy for both tcp and udp on your server instead of just masquerading the traffic? π
Yes, Iβm sadly not that strong in CLI and configuring iptables. So nginx just has worked perfectly previously
easiest way is to put it in your game client's save folder and use the server manager's upload save feature
and where would that be?
Normally I would say whatever you want, though some people are reportedly experiencing more networking issues and an event log issue on windows
ok so maybe ubuntu?
refer to my prior message
oh i see
ugly hack.. but if it works for you π€·ββοΈ
if you remove and re-add the server through the server manager in-game, does the server name load properly? (even if the server itself shows as offline)
I have a guess of what might be wrong but I've never done such a weird setup
is my server up for anyone 104.153.105.95
testing port forwarding
its my vps ip.. so its safe to post
yes
server called Keystone Earth on game version 372858
probably, I'd have to launch the game to check that
it's apparently loading a map right now
did you want me to try to join?
Yes tried readding and everything. Just canβt seem to connect to the port, but can fine connect to the MC server with same configuration (different ports) and can ping the server and see the ports are assigned and open
all good
I'd need an answer to the specific question I asked, not generics, unfortunately, to identify what specifically is going wrong
you say you can't connect to "the port" but there's two bind ports for the game server: 7777 TCP and 7777 UDP
did it let you in?
Look at chat and the xeno-zapper I dropped c:
ok thx for the help
Booting now
btw anyone is free to join.. ill try to keep it up
You'll have more luck in #1201555265942724758 if you're looking for players.
This is the error
And server name pending in server manager
do you get this error as soon as you add it to the server manager? π€ I didn't know it could show that error there
if you're seeing Server Name Pending and not the server name you configured there's something wrong with the UDP proxy or it's being blocked by a firewall. on the server's machine, do you have any firewall installed? (i.e. firewalld, ufw, ...)
Nope, by trying to join directly
I'm using hetzner and allowed the following
All outbound is allowed by default
You can't use the Join Game feature to connect to dedicated servers at all.
You can only join through the Server Manager
you bet, from outside local network.
any firewall/netfilter installed on the server/vps itself though?
Is the WAN IP in your router's control panel the same as your WAN IP on e.g. checkip.amazonaws.com ?
Nope, none at all
good question, let me check
your reverse proxy has a listen config including :7777 udp reuseport ?
No, i don't think so. Where should i add that? the nginx stream conf?
you said you configured this in nginx:
Yes, with Nginx as stream proxy for making the iptables configuration
UDP TCP 7777
how did you specifically configure this?
what GUI is that π
Nginx, and then the seconds is just nano
I don't think it's absolutely necessary, it just seems more sane to me imo. not sure what's going wrong but I know nearly nothing about nginx stream as a proxy
aha. getting responses would be useful yes π
change the config and then systemctl reload nginx.service
or "nginx -s reload"
That makes a lot of sense ahahaha will try now
Just restarted the docker compose
Sadly still server name pending, but have to go now sadly. But will look more into the reuseport part, or if you have any other tips i should investigate please share π
just to clarify: don't check only the vps you're running the vpn on, but also the server where the satisfactory dedicated server itself is running on
since you have a multi-server setup here
192.168.87.58 is the wireguard peer IP of the server that has the satisfactory server process running right?
The server works fine over ZeroTier and over LAN
Local ip of amp server and configured as advertised route for Tailscale vpn
well you're dropping the UDP packets somewhere, so just to doublecheck, on both the vpn server and the satisfactory server machines:
sudo ufw status
sudo firewall-cmd --state
Okay perfect will try tomorrow! Thank you for the help
is there an offical docker image anywhere?
I found the wolveix one, Im not sure if that is the recommended
CSS does not publish any container images for the satisfactory dedicated server. The only official distribution channels are through Steam and Epic Games as far as I'm aware.
alrighty
The wolveix image appears to be the de facto recommended image at this time
Been running the wolveix for years, its been stable and problemfree π hosting 3 servers for me n friends
So my computer is starting to struggle late game hosting a server for me and 3 friends. Am I able to migrate a local hosted game to a dedicated server easily?
As easy as it could possibly be. You just go to the ingame server manager, 'manage saves' tab -> 'upload save' sub-tab -> select a save -> hit the 'upload game' button. Then you need to go to the 'load save' tab, select the save you want to load on the server and hit the 'load game' button. It should even load automatically after restart (check the server settings though).
Wouldn't it be nice if there was a list of available/open public servers under join game?
no
Thank you!
Hey everyone, I have recently set up a dedicated server using AMP on my Linux server. It was working for awhile, even had a friend able to join and we've both been able to put about 8 hours each on the server. All of a sudden yesterday I had some weird stuff start to happen when I was mass deleting a bunch of foundation, some of them started to not disappear but were technically "not there" and interactable. Then some of them wouldn't even be deleted, I could just delete them over and over and they were still there...
I logged out and restarted the server, turned up the Tick Rate and turned up the Network Quality, seeing some things saying that maybe the server wasn't able to "send enough information for the size/density of stuff in the world". Now when I try to join my server it starts to load in then it kicks me back to the main menu with a reliable buffer overflow error.
Looking in the server console I see that error and a "LogGame: Warning: AFGFoliagePickup::BeginPlay: Could not find a local player controller"
And as I write this, I have the console open still and it just output
LogSocketsSSL: Warning: Failed to initialize SSL state for Socket HttpRequest
LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR
20:24:49
LogSocketsSSL: Warning: Failed to initialize SSL state for Socket HttpRequest
LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR
I have no idea what is going on...it was just working.
My friend is still able to join, but says that sometimes it gets stuck loading the server to display in the server manager screen but he backs out and goes back into server manager and it shows up.
Anyone have any ideas what's going on or any steps I can do to figure out what is going on?
Update:
I just got the closes I've been to getting back in. I was at the loading screen for much longer and went through a bunch more of the scrolling updates before getting in but then kicked me out with this error:
Network Error
UNetConnection::Tick: Connection TIMED OUT. Closing connection.. Elapsed: 30.41, Real: 30.00, Good: 30.00, DriverTime: 67.55, Threshold: 30.00, [UNetConnection] RemoteAddr: (My publickIP):7777, Name: IpConnection_2147476811, Driver: GameNetDriver IpNetDriver_2147476812, IsServer: NO, PC: BP_PlayerController_C_2147476494, Owner: BP_PlayerController_C_2147476494, UniqueId: Steam:1 (76561198025969930)
console is outputting this, it seems it says I connected successfully (apparently too large and put it in a text file)
Did your host provider rate limit you?
I've heard of some people, running lower end servers, getting rate limited by their provider if they push it.
You should be able to check by logging into the host provider.
Do you mean my ISP? Or do you mean Cloudflare where I purchased my Domain?
You're running AMP, yeah, so are you running it locally? Or on a server?
I'm running it on my own local linux server.
It's strange that my friend can still join. He's actually in the server now. But I keep getting the errors when I try to join.
Have you tried joining via localhost instead of via ip?
Or are you cloudflair layered?
I have my whole setup routing through cloudflare, but I still should be able to join locally, right?
Depends on how you set it up, i would try.
All of my local services are impossible to access via local. Its essentially a seperate network for me.
I know I've joined one of my other servers with my servers local IP.
I just tried Satisfactory and when using my local IP and port I get the same UNetConnection::Tick error from above.
Except this time it has PendingNetDrive and IsServer: NO PC: NULL, and Owner: NULL, UniqueId: INVALID
When I try to join with my friend in the server he gets notified that I joined. And in the console it even says I connected.
Console output:
Yeah idk man, i would restart everything and see what happens. Computer, Linux, router, just restart all of it. Doesn't really make since why you're getting rejected.
I'll give it a try. It's some cruel joke that my friend can play on my dedicated server but I can't lmao.
Restarted everything, still a no go.
Another thing I'm noticing, it seems no matter how many times I do "Update All" Satisfactory always has an update. It's strange.
Firewall is inactive on both, but the VPS is done behind hetzner so there i have allowed for both tcp and udp on port 7777
hello how can I add BPs to a server with a host? because I have a server BP folder because I created BPs but when I add them it doesn't work and also even restarts the server. knowing that it's not my server but I can do it by asking the guy who hosts the server
nope
that's the dedicated server files
what you want is config/user_setting/data files
%userprofile%\AppData\Local\FactoryGame\...
Are we able to host multiple sessions on a single dedicated server, i assuem we can just designate a different port? Is there any documentation for this?
Any fixes/workarounds for MP DESYNC issue on dedicated when it comes to deconstructing?
Multiple sessions on a single dedicated server? You mean in a single server process or in the same machine?
You can host multiple sessions on the same machine on different ports. You cannot host multiple sessions in the same dedicated server process.
Not sure about the MP desync issue, please make a ticket on the QA site or give a link to the existing ticket
In the same machine
yes offcourse, which ports or range or ports can we use...i know default is 7777
can do
i run multiple servers on my homelab using a vm for each server instance and just change the port count +1 fir each instance 7777, 7778, 7779, and they are run fine
any. there are no restrictions.
thx
I have a problem with my dedicated server [Experimental] (Epic Games)
Ive set a custom port (25565) cuz i wanted to dodge doing port forwarding on my router again and I already had that port set up for older minecraft servers which worked fine
I also set inbound and outbound rules for port 25565 UDP and TCP
The server is hosted on a windows machine. The same machine also runs the game.
I can connect via my local IP but not via my public IP and people from outside also cant do so
I tried since yesterday and at this point I have no more ideas other than coming here for help
make sure you also forwarded for both UDP and TCP
I do
this is outbound but the same thing is inbound
so router yea?
yes
ok
Last tab iirc
it's setup automatically with wiki provided script
Also outbound rules are useless
you mixed local and remote port
anyway if you want to make sure it's already you can delete all these
and execute the provided script instead (they will appear then anyway)
yes
I just did this
so now i can try to use one of the provided websites which check if the port is reachable
So you know on normal firewall setup outbound rules are only used for restrictions
here you want to allow so no need for outbound ones
still no worky
did you change the port number in the script ?
yes
on both lines ?
yes
it said OK and you see it
show 25565 ?
very right of the img
In manual settings, as your screenshot
this company lol
so i used the provided script and changed the port to my custom set one
i set port forwarding on my router
which worked a few months ago with an MC server
double check your machine is the target of the forward rules still
i pinged it and its reachable
after that only I see is either need restart every/something because it is not triggering or
you use / behind cgnat
or you failed testing it and it actual works already
been getting this a lot client side recently. Thoughts?
Fatal error: [File:C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Private\HAL\MallocBinned2.cpp] [Line: 1438] FMallocBinned2 Attempt to realloc an unrecognized block 000002104D0C0000 canary == 0x0 != 0xe3
FactoryGameSteam_Core_Win64_Shipping!FMallocBinned2::FreeExternal() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Private\HAL\MallocBinned2.cpp:1106]
FactoryGameSteam_Core_Win64_Shipping!FMemory::Free() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Public\HAL\FMemory.inl:141]
FactoryGameSteam_RHI_Win64_Shipping!FRHICommandListImmediate::FlushPendingDeletes() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RHI\Private\RHICommandList.cpp:1024]
FactoryGameSteam_RHI_Win64_Shipping!FRHICommandListImmediate::ExecuteAndReset() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RHI\Private\RHICommandList.cpp:732]
FactoryGameSteam_RenderCore_Win64_Shipping!FRDGBuilder::EndFlushResourcesRHI() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RenderCore\Private\RenderGraphBuilder.cpp:387]
FactoryGameSteam_RenderCore_Win64_Shipping!ParallelForImpl::ParallelForInternal<TFunctionRef<void __cdecl(int)>,TFunctionRef<void __cdecl(void)>,std::nullptr_t>() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Public\Async\ParallelFor.h:411]
FactoryGameSteam_RenderCore_Win64_Shipping!FRDGBuilder::CreatePassBarriers() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RenderCore\Private\RenderGraphBuilder.cpp:3019]
FactoryGameSteam_RenderCore_Win64_Shipping!FRDGBuilder::Execute() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RenderCore\Private\RenderGraphBuilder.cpp:1906]
FactoryGameSteam_Renderer_Win64_Shipping!RenderViewFamilies_RenderThread() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Renderer\Private\SceneRendering.cpp:4578]
FactoryGameSteam_Renderer_Win64_Shipping!TGraphTask<TEnqueueUniqueRenderCommandType<`FRendererModule::BeginRenderingViewFamilies'::`80'::FDrawSceneCommandName,`FRendererModule::BeginRenderingViewFamilies'::`80'::<lambda_3> > >::ExecuteTask() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Public\Async\TaskGraphInterfaces.h:1265]
FactoryGameSteam_Core_Win64_Shipping!FNamedTaskThread::ProcessTasksUntilQuit() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Private\Async\TaskGraph.cpp:648]
FactoryGameSteam_RenderCore_Win64_Shipping!RenderingThreadMain() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RenderCore\Private\RenderingThread.cpp:412]
FactoryGameSteam_RenderCore_Win64_Shipping!FRenderingThread::Run() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\RenderCore\Private\RenderingThread.cpp:537]
FactoryGameSteam_Core_Win64_Shipping!FRunnableThreadWin::Run() [C:\BuildAgent\work\b731a33f2a691e17\UE4\Engine\Source\Runtime\Core\Private\Windows\WindowsRunnableThread.cpp:149]```
Unstable cpu or ram
Router restart done
I'm gonna try to launch the Minecraft server and see if I can connect via public ip
interesting. my OCs have been stress tested before and never had issues. Should I just remove my overclocks?
You're crashing in malloc so I would start by going back to base clocks and non xmp ram speed. Then you can slowly ramp back up to values that remain stable. UE5 tends to make workloads that are more sensitive to memory issues than most RAM stress tests nowadays
Mostly because the load fluctuates a lot more than a constant high in most stress test scenarios
my server is listening on this port but it still cant be reached
Is this on windows or linux?
is the WAN IP shown in your router control panel the same as your actual WAN IP on e.g. checkip.amazonaws.com ?
Also can you run this in powershell?
Get-NetTCPConnection | Where {($_.State -eq "Listen") -And ($_.LocalPort -eq 25565)}
Get-NetUDPEndPoint | Select-Object LocalAddress, LocalPort, @{Name="Process";Expression={(Get-Process -Id $_.OwningProcess).ProcessName}} | Sort-Object -Property LocalPort,Process | Where-Object{$_.Process -like "FactoryServer-Win64-Shipping-Cmd"}
I should really change the udp query next time to port number as well xD
will do
okay bind ports look ok. can you check the WAN IP in your router control panel?
just did and pinged it again and its reachable
no I mean, is it the WAN IP in your router the same as your actual WAN IP? Just trying to eliminate dual NAT/CGNAT here
its been crashing now, though now I spawn in as a new character and not my character
same IP worked on a mc server minutes ago
You may have reconnected to the server before your old session disconnected so you got a new pawn. Also you can try posting in #1038092680493801533 where baldurfi can maybe help you. He's more well versed at client crashes like that
If it's 13th or 14th gen Intel cpu it might be fubar. Might also try to ddu and reinstall graphics drivers
It's a simple yes/no question which would be useful if you could answer it since minecraft is TCP only so the connection protocol is different
Oh I definitely reconnected before my old session disconnected
I saw myself log off after I had connected
you can kill your old pawn for its inventory
yes it is I guess
hotbars are unfortunately lost unless you grab them in SCIM before you kill it
never fiddled around with WAN
other than killing is there no other option?
loading a previous autosave on the server
do you know where the status page/internet connection status/connectivity/... page in your router control panel is?
yes and i copied my IP from there
i also used a website to double check my WAN IP and they are identical
okay great! are you connecting to your WAN IP yourself to test or do you have a friend testing for you?
both happened already and both fail
i also tried with the default port being 7777 and same result
obv i rebooted the server to listen on that Port before testing
Okay so it should be important to note since I don't know the specifics of your router that testing from inside your LAN using the WAN IP may not work, since this involves configuring something called NAT reflection, which may need additional configuration for the UDP port to work, so we'll want to continue testing from outside your local network
i also set firewall rules and port forwarding before testing
ok im down for that in a call or smth
why is this shit on 58171
You'll have to ask a friend to help you with the testing, use a separate device on a separate internet line (e.g. mobile hotspot) to test, or you can share your IP with me. I won't be available for a call for another several hours unfortunately
it's an automatically assigned return address. this is normal
thought so
i can share my ip no problem if you dont attack me or share it further
UDP is a one-way stateless communication protocol so a return address is established
thats not attacking tho
I'm not going to be doing anything with your IP address other than testing for as long as you want my help. I'll delete it once we are finished
ok cool gonna dm it to you
can you check if your local IP address for your server and the local IP address configured in the port forwarding are still the same number? Your router may have assigned a new IP address via DHCP if it's not statically configured
note, that screenshot contains your WAN IP
yee anyway
so my German is like mega rusty uhh that screenshot is for both firewall and port forwards right? it just says freigaben
thats the router port forwarding basically
firewall would be on my pc no?
I don't recognize the router brand, some routers have separate port forwarding and firewall sections, but if you got minecraft to work without doing anything else it's probably not a worry for now
AVM Fritzbox is the router type
can it be that UDP is somehow restricted through some windows settings?
Can you run following powershell script as admin?
New-NetFirewallRule -DisplayName "Allow Satisfactory default inbound port udp" -Direction Inbound -Action Allow -EdgeTraversalPolicy Allow -Protocol UDP -LocalPort 25565
New-NetFirewallRule -DisplayName "Allow Satisfactory default inbound port tcp" -Direction Inbound -Action Allow -EdgeTraversalPolicy Allow -Protocol TCP -LocalPort 25565
done that just now
hm still getting connection timeouts
yea
so since mc servers work tcp is fine but udp seams to have an issue
or am i wrong here
TCP's timing out as well
I fr have no idea what im doing wrong
What port is your Minecraft server running on?
default so 25565
but not running atm obv
Do you have any firewalls installed other than Windows Firewall?
no
i also have no vpn or proxy set up
@magic jolt can I recommend trying to set this up using the default ports instead of attempting to remap them?
i tried that with no success either
but sure we can do that again
I think it'll help eliminate variables
Are you using your public ip address to connect to the server
works on local IP but not WAN IP
Make sure to update your satisfactory config as well
can you try removing the UDP forward and applying the settings? I want to test if forwarding only one protocol works
server now listening on 7777
sure
rule deactivated
so tcp should still work
when all else fail, grab wireshark and see if windows even receives the packets. you'll see them even if windows firewall them off. if you dont see them at all, it's the router
How recently did you have your Minecraft server up? Some ISPs block inbound ports, and it's possible they started doing that since you last hosted a server.
And what happens if you ping you external ip address with your port number so ping externip:port
I tested today
you cant ping ports
You can't ping ports
no on the TCP either. you can activate the UDP rule again, though I did notice a different interesting issue. your router is allowing the TCP connection in on 7777 now but it's not reaching the destination internal IP. is the internal IP properly set in the port forward? either the local IP is set wrong or the firewall of your server is misconfigured
ICMP is also it's own protocol and wouldn't be forwarded by a TCP/UDP rule
(your router hard rejected the tcp connection on port 7777 when it wasn't forwarded and now it just times out instead)
Test-NetConnection ip_address_here -p port_number_here
That's a TCP connection test, which you would need to do from outside the LAN if you want to test the port forward... since NAT reflection and all that
how even if the serevr is running?
destination port or IP is wrong in the forward or the firewall is misconfigured on the server
on the server?
i have the server running on PC or are there any config file lines i needed to edit?
"the server" being the PC you're running the server on. sorry for the confusion
np just wanted to make sure
can i somehow reset my nic drivers easily?
just to cross that out as a fail point aswell
the screenshots from the fritzbox manuals I can find in regards to port forwarding look very different from what you've shared. are you under internet -> permit access -> port sharing?
Yes
It's just internet/port sharing nowadays tho
Your manuals are probably old
I have a 7590
Maybe I should try to download the server on steam or steam cmd
But it should be the same files so idk what that would do
the screenshot you sent earlier was cropped. can you DM me the full screenshot? I should see, at minimum, the internal address, the port to device, external assigned start port, external assigned end port
Good morning everyone. It's time for my bi-weekly "Is there any word on a fix for the desync issues with foliage and blueprints?"
not sure what you've got going on with blueprint desyncs but someone posted a config change you can do to fix the foliage issue the other day
oh, i missed that.
You can try finding it, something something array size something? Or I can check later once Iβm not on a mobile device again :P
Was in this channel I do think
Thank the smart person who found it π
Hello guys, I have a problem with a dedicated server. Port forwarding seems to work. Because when I type in the URL and Port in the browser from outside the network, I get a prompt that the connection is unsafe. When viewing the certificate its the self signed certificate from the server I am expecting. But when I try to add the server in game it just says "server name pending" and "not authenticated". Oh yeah and I use duckDNS.
What is your network topology like? Are you on the same LAN as your server? Have you tested both TCP and UDP connectivity?
What's the setting that need to be changed?
: Issue: When too much foliage, trees, or flora are destroyed in an area, the game netcode fails to properly communicate what foliage has been removed to game clients. This can cause synchronization problems where these objects visually appear but are not actually present on the server, you can walk through these objects but it will typically cause quite a lot of rubberbanding.
Solution: Run the following commands in order, then leave the server and rejoin.
net.MaxNumberOfAllowedTArrayDeletionsPerUpdate 4096
net.MaxNumberOfAllowedTArrayDeletionsPerUpdate
net.MaxNumberOfAllowedTArrayChangesPerUpdate 4096
net.MaxNumberOfAllowedTArrayChangesPerUpdate
This will help the server correctly manage foliage destruction updates.
I had to run it in both the server console and local console
and then rejoin
but it worked
thanks
The Server runs at my parents home. I am connected throght the internet of my university. So they are not on the same network. I don't know the topology of the university network. I didn't check UDP and TCP individually, how to do that?
okay, knowing you're remote helps. I have a tool I wrote to check both ports at https://github.com/dopeghoti/SF-Tools/ or if you give me your server address I can test it for you
I will try it first. I'll let you know if it worked or I need your help. Thanks!
that's a lotta steps to just pip install and run
your code's only as good as your docs yknow
I do need to work on my docs, but I tried to keep the requirements light
other guy's
Fair, but I'm willing to concede that I need to spend some time on documentation
So I tried running it. On the server it worked properly and displayed the stats. On my side the connection timed out. So it might be the dns and multihome settings?
it's not likely to be the DNS but that can be verivied by testing with your IP address rather than your hostname
The most likely issue is that you are not forwarding both of the ports you need to forward
(by default, TCP/7777 and UDP/7777)
slightly less likely is that the server's LAN address may have changed and you need to update the router to point to the new internal IP address.
If you can test from another computer on the same network as the server, that would be a good way to determine whether it's a problem with the server host's firewall rules
I checked the firewall, I have TCP and UDP Ports in the Inbound rules. So that should be fine I guess.
I tried to put the URL and Port in the Browser again without https:// at the beginning and the Server actually display something:
LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR
[2024.10.21-17.29.54:926][504]LogSocketsSSL: Warning: Failed to initialize SSL state for Socket HttpRequest
[2024.10.21-17.29.54:926][504]LogHttpListener: Error: Error accepting expected connection [0] SE_NO_ERROR
LogSocketsSSL: Warning: Failed to initialize SSL state for Socket HttpRequest
So I think the Ports are actually working.
Wait. Why does it say its 17:29? The time should be 18:29. Or is this just because computers don't use daylight saving time?
The SSL handshake will fail if the server's timezone settings are wrong- both the server and client have to agree what time it is for the SSL connection to establish.
It happens more often than you'd think
(the logs generally are time-coded to UTC)
I think the time zones are set correctly. The remote control software I use for the server also uses this. When I switched time zones I couldn't connect anymore or startup satisfactory from epic games launcher.
when attempting to go to your server address through HTTPS on port 7777, you should see this:
that will confirm the TCP connection is going through also
Yes I see that
okay so the SSL and the TCP connection is good. of the CLI tool is still showing a timeout that's a UDP connection issue
Still time out. So its the UDP
What do you think? Is there a way to get it working?
well, other than double-checking (in order from the outside in):
- Is the server running?
- Is the server host reporting the server bound to both the UDP and TCP port?
- Is the server host firewall correctly configured for both UDP and TCP?
- Have you verified the LAN IP address?
- Have you verified the router's port forwarding rule for UDP is pointed at the correct LAN address?
- Have you verified the router's port forwarding rule for TCP is pointed at the correct LAN address?
- If you are using IPv6, have you tried across the board only using IPv4?
I can't think of anything else to check off the top of my head
the SE_NO_ERROR is still pretty active globally
the fact that you can connect from inside the LAN but not outside points to either a router issue or possibly an internet issue.
Actually there is one thing I can think of, which should never actually help but sometimes does
- Have you tried turning it off and on again? (the server, the router, everything)
Okay, I found the issue reading through your suggestions. Probably the Port forwarding is not setup for UDP. Gotta change that but this will take some time until I get home again and then see if it helps. Massive thanks for your help dude! Much appreciated!
pleased to have been of help!
There used to be a section on the Wiki I wrote called "Seven Steps to Satisfaction" that had that as one of the callouts; at some point that was written out I guess.
Interesting... But it's a good tipp!
good luck!
Folks'll still be here if you're still having issues after having looked at the router.
which page was that on?
Do you have any experience with the server running through nginx stream and a VPS for a public ip?
No, and it would never occur to me to try to use an HTTP server daemon to route traffic for a service that is not exclusively HTTP. Shoehorning it in like some people do is mystifying to me.
works great π
(ignore the haters)
I'm not a "hater" per se. I just see that as an ask and think "but why tho?" and move on with my day. If it works for you, that's great! That's the wonderful thing about systems administration: it's a cat with so many skins.
Can you share how you got it working? Configs or etc in dm?
Can you explain the proxy_pass part? Server is just the server ip correct?
yes. replace <Server>with IP, DNS-name or name of a upstream block
it passes the proxied connection to the specified target ip:port
Okay perfect, will try again tomorrow. My head is scratched almost through to the scalp π
tbh nginx have good stream capabilities
but I hate when people call it "nginx reverse proxy"
it's not a nginx reverse proxy
it's simply connection/stream relay/forward
it's not even tunneling either
Probably stems from trying to define things in terms of an HTTP server
saying "connection proxying" is not wrong
saying only "proxying" already imply wrong things (border line is tls termination relay)
which it does not operate in this mode at all
I get that, but when you take a pickup truck to the drag strip people are going to look askance at you until you pop the hood and show off your roll bars (:
well it is acting as a reverse proxy. even if it is basically just a layer3/4 switch/router is this case π
it's not a reverse "proxy"
without any specific term or context
it implies it overlay the game server
which it does not
in this mode it overlay nothing
it only forward at connection level (without handling application data)
it does not even handled then repack
it just send as is
"Proxy" implies "I am an endpoint for service X but I am instead going to forward your traffic to that service X endpoint over there". In this case it's not even at that layer of communication but just routing traffic at the network protocol level rather than the application protocol level
TL;DR: words mean things (:
Wouldn't proxying raw packets necessitate rewriting the ip header and keeping track of the response port much like in regular connections? Can satisfactory even see the ip from the users or will it only see the nginx ip?
that's what routers do
I imagine it's going to think all users are at the nginx ip if I understand it right
in this case the HTTP server is acting as a router
They only modify the ttl unless there's NAT right?
yes that's why it need reuseport to work
and yes, unless DS can understand proxy protocols it'll only see the nginx IP
SF speaks HTTPS and its own UDP API protocol and whatever black magic Unreal brings to the table.
proxy protocol need explicit handling
here it is not used at all even on nginx side
nginx here work as router with nat
I effectively do the same but with nftable (Linux netfilter)
I like compartmentalization. I use my firewall to firewall, my router to route, and my HTTP server to serve HTTP traffic.
Others like one-stop shopping and stuff everything they can into one dæmon like nginx.
With NAT only the destination ip would be rewritten. With a proxy it's both source and destination. So does more than a router. Since the downstream app needs to know to respond back to the proxy and not directly to the sender ip
To each their own (:
NAT is for network address translation
there are multiple type of these
there is dnat and snat
usual routers does both dynamically (snat on reverse path)
and nginx act as both enforced bidirectionally
source or destination ip, depending on direction. switch + NAT = nginx stream rev-proxy
Yes that's what I meant to write. My bad xD
for what's wanted here you want automatic reverse path with dnat
like an usual router consumer NAT
I'll leave troubleshooting the nginx users to the nginx pros
sometimes I wish the original terminology caught on more. "IP Masquerading" makes is so much clearer what's going on.
I can setup nat in iptables no problem but I'll die if I have to troubleshoot why it doesn't work for someone in nginx
NAT is more than ip masq though? it changes ports too
IP Masquerading is part of it.
