#dedicated-servers

1 messages · Page 230 of 1

wind saffron
#

Sup y’all, currently acquiring some second hand parts. Would an i7 7700k with 16gig ram be sufficient to run my own server?

bright robin
#

yeah, that would be enough, and with some more RAM you could run several servers too.

balmy smelt
bitter sentinel
#

swanky NeuralBlender moment

toxic sentinel
#

I had a server running on one of my machine, but this machine was on WIFI. I took the savegame and put it on another machine, changed the router ports forwarding, lauched the server , connect from local client , loaded the savegame. Added the ports to the windows firewall as well , has I did on the previous machine.
But I can't connect remotely from my location to the newly started server.
Anyone knows why ?

#

"This server appears offline."

toxic sentinel
#

ok I get it , I restarted my router, resetting NAT

#

that works

glass zealot
#

Hey guys, is there any way to get older satisfactory server version?

timid cobalt
#

So following situation: I have a dedicated server running on a secondary computer and when I want to access save files on my main computer, I have to move them from the save file folder to a network folder and from there to my pc. Can I, instead of using a folder on my secondary pc as the save file location, rather use a shortcut to a folder in my network storage and rename it to the original folder name? Will this work, ie. will the save files be stored in the network folder and not on the local machine?

whole dome
#

If both are on windows you can map a network drive and easily copy back and forth that way

fervent carbon
#

Can you create a symlink to the network drive?

gloomy pilot
#

Hey guys, are there any public servers that you can visit ?

frosty gale
#

Hey, It appears that my satisfactory server is offline even though I have port forwarded ports 15000, 15777 and 7777 on my router (UDP)

frosty gale
celest shell
#

hello, i need some help. I can see my server "not authenticated" i can try connecting but just ends up timing out... is it common?

#

trying to set up a dedicated server^^

frosty gale
#

There's no doucmentation so...I'm having no luck myself either.

sturdy bough
celest shell
#

I did also

#

I managed to connect with my local IP...

frosty gale
#

okay I got it running and can connect from teh host pc with local ip but not with external IP (says server appears to be offline), and def have all 3 ports forwarded to that pc from router. USeless if my friends can't join that way...may as well just launch the game and idle with that PC then play from another PC. This is MUCH more trouble than it's worth. Maybe in a year it'll be less frustrating.

crystal eagle
#

On the server pc Google 'what is my ip' to get the correct external ip

frosty gale
fallow flame
#

I still have no idea how to set up a dedicated server after reading instructions

frosty gale
#

the first step is choosing where, windows/Linux, then choose type of install like steam/steamcmd/epic

#

also docker, but thats not on the wiki i linked

fallow flame
#

Yes but I read it on a day when I didn't take my meds so I'll try again

limpid granite
#

Are all vehicles bugged in servers or is it just the factory cart

opal compass
#

So i ran a multiplayer session and I'm still getting this doubling shadow with vehicles not sure what's going on there

civic lantern
#

Hi, anyone with experience using sf with zerotier?

#

Actually, I just realized, I have a spare laptop I can try this on nvm.

vernal loom
#

How do i join a dedicated server?

whole dome
vernal loom
opal compass
#

Anyone running a dedicated server having a black ghost from vehicles when spawned in the min. I drive it leaves one behind

naive cypress
#

I tried running a docker dedicated server, but whenever I try and connect to it it just disconnects. Am I doing something wrong?

lean heart
whole dome
#

Dedicated servers don't support mods (yet?)

crystal eagle
timid crypt
#

This error has been preventing me from running the Epic dedicated server.... any insights? literally out of the box, i get this error

unknown.png
hasty girder
#

Will we ever be able to save a Server Save Game localy? Dedicated hasn't got any Updates in a while, hope Update 6 adds something

trim hearth
#

you can, you can transfer a save to your local computer in the server manager section, provided you have access

#

failing that, ask the server owner nicely and they might give it to you, its not hard to pull

trim hearth
trim hearth
opal compass
#

I restarted my dedicated server but it's even happening on the multiplayer side to I tested it

trim hearth
#

im pretty sure its a known bug

opal compass
#

Doesn't seem to want to go away

#

Not hurting anything right now

trim hearth
#

yeah my thoughts exactly

#

hopefully update 6... and with it some better admin tools for dedi's

timid crypt
vocal latch
#

this is OLD but did you ever find a fix for this? Getting the same issue.

gritty elbow
#

does anyone host this on a cloud and experience AWFUL rubber banding? Is this a host issue, or just this terribly unsupported ass server

timid crypt
pulsar bramble
#

Dedi servers are trash period atm

vocal latch
#

Fwiw, we are hosting it in Azure's cloud running Win10 Pro. Wanted to get it up quick and I'm more comfortable with w10 than ubuntu, but if ubuntu is more stable to run as a server, I'm not opposed to trying it out

tall jay
vagrant turtle
#

If you are running windows, have you allowed the ports on windows firewall on the server itself?

#

I have never had the issue in linux .. always windows servers, where ports are blocked by the OS unless allowed.

#

linux port allowing is easy (at least for me hehe)

vernal loom
#

anyone know why this happens?

unknown.png
sturdy bough
#

whenever I re-join my dedicated server game, I spawn in the hub and find 'myself' sitting on the ground outside, I have to kill myself to retrieve my inventory. I just started a new game this morning, but I've not had any issues like this on my past playthroughs. (FYI I log out normally from the game each time)

untold prairie
frosty gale
#

I'm getting bugs where items don't flow as they are supposed to

#

the speed

#

and visual bugs

trim hearth
#

yep

#

falls under "experimental"

frosty gale
#

yep I know

#

i have found a way to fix tho

#

downgrade belt and upgrade again

wintry stone
#

it definitely wasn't an "every log out" sort of thing with mine

fresh saddle
opal compass
trim hearth
#

stinking programmable splitter wont keep its config 😦

wintry stone
#

when I add the rules and change the rule I add in the same "session" it doesn't stick around

sturdy bough
sturdy bough
trim hearth
#

Let’s hope! But also - let’s not pin all our hopes on one update

remote ferry
wintry stone
#

interesting, I've had a bunch running in my train drop off for a long time at this point

#

guess I'll take the risk and use my backup saves if it comes to that 😄

blazing halo
blazing halo
# trim hearth stinking programmable splitter wont keep its config 😦

Yeah, the programmable splitter is broken right now. Can you use the smart splitter? It's good for most purposes you'd use the programmable one for (I guess), and it does keep its config. You will need to select options before it lets anything through though (the defaults don't actually auto-apply at creation).

round glade
#

heyo, what are the most important things that happen to be bugged on dedicated servers?

muted wyvern
#

I'm trying to diagnose a conveyor issue on dedicated server satisfactory.
20 Iron plate smelters, 30 output per (600 total), being backlogged on a level 5 conveyor (780 potential).
I dont see anything in the FAQ, anyone have advice?

#

The effective rate is ~480

frosty gale
muted wyvern
#

Originally, a sink (connecting a train station atm), but in the meantime i have it hooked up to a storage unit

frosty gale
#

so its just a line of smelters all merging into one line

muted wyvern
#

oh no

frosty gale
#

then you send it to storage unit and sink it?

#

oh no

muted wyvern
#

I built that factory i linked. Then send it to a sink

#

the issue is in the smelter line at the start.

frosty gale
#

so iron ore is backing up?

muted wyvern
#

yes

#

no

#

Iron bars

#

Iron ore is full backed on input of smelter

#

smelter has full 100 pending output. Only 16 of 20 on each row are running.

#

the other 4 are backed up

#

Checked the conveyer belt, its running. If i remove and replace, it clears.

frosty gale
#

so the iron ore is runnning at 780 too?

muted wyvern
#

720

#

both are capped at 720

#

on the 780 line

frosty gale
#

yes

#

i meant belt, sorry

muted wyvern
#

yes

frosty gale
#

so then it goes to 20 smelters

#

then smelters go to...?

#

your chart shows 96 smelters lol

muted wyvern
#

I have 96 smelters total XD

#

its several rows of 20 XD

#

and a row of 16

frosty gale
#

ok where is this row of 20 or 16 connected too

muted wyvern
#

both feed and load balance into the next row of processing. The chart i sent shows the output breakdown

frosty gale
#

yeah but you cant combine all those smelters onto one line

muted wyvern
#

i can post a screenshot if you want, its not the cleanest in the world..

frosty gale
#

sure

muted wyvern
#

oh no, load balanced into multiple out

#

4 780 lines out

frosty gale
#

what do you mean load balanced out?

muted wyvern
#

all 4 lines, split into 4, then merge into the 4 out

frosty gale
#

So you have 3 lines of 20 smelters which would be 600 ingots

#

and 1 line of 16 smelters or 480 ingots, correct?

#

3 lines of 600 ingots each

muted wyvern
#

yes

frosty gale
#

Ok, so what happens to those lines?

#

you are merging?

#

or you are keeping them all separate and splitting

muted wyvern
#

each splits into 4, then merges back into 4 (1 from each line)

frosty gale
#

So you split a 600 line, into 150 each, then back into 600?

muted wyvern
#

I'll redo the load balancer section. its a mess anyway..

frosty gale
#

I don't understand what this load balancer section is for lol

#

If you are just matching input with output, nothing needs to be load balanced

#

Like, if you do a line of 20 smelters

#

you just run the ingots into a line of constructors that total 600 ingots

muted wyvern
#

honestly, it was for paranoia XD

frosty gale
#

paranoia of what

muted wyvern
#

incase i was wrong

#

or if 1 line gets backfeed in the chain, the overflow will go to the other lines

#

oh, its 4 rows of 20, not 3.

#
  • a 5th row of 16 at 30 output
#

and the output needs to be 4 conveyor lines, thats why.

muted wyvern
#

switched the 4 200 lines, to just receive a split 480 line (480 split by 2, by 2 [aka 4]) then output the remaining line. I'll update if that works, will take about an hour

#

(to clarify why i asked in this channel, i ran the map locally, no issues)

silent venture
#

Hey guys - Is there a way on the Windows build to set the IP being used? I can connect locally, but it isn't broadcasting correctly, even with the port forwarded / DMZ on my server.

frosty gale
frosty gale
vagrant turtle
#

And have your router point those ports to the IP of the server.

#

and using multi-home is also a valid point..

silent venture
vagrant turtle
#

and that's on the router, correct?

#

if you can connect via internal network, then the server is working fine.

silent venture
#

Yeah, that's what I'm gathering at this point. Server > Dumb switch > Router. Just thinking out loud at this point.

vagrant turtle
#

ok, log into your router .. switch is irrelevent.. Make sure that those 3 UDP ports are forwarded to the IP of the server.

From what I'm reading of your posts, that sounds like what your main issue is.. the router is not configured properly

silent venture
vagrant turtle
#

static IP on the server, or using DHCP?

silent venture
#

But it's also odd because my webservers are all functional.

#

Static

vagrant turtle
#

hmm ok, same machine I'm assuming for the web pages...

silent venture
#

Yeah, port 433, 80 are fine

unknown.png
#

All from local .66

vagrant turtle
#

ok, I had to do this...

#

15777 - 15778
15000 - 15001
7777 - 7778

silent venture
#

Oh? I'll try that then

vagrant turtle
#

yeah, I had to do a range instead of just the single port

#

depending on the router, you may need to have to be outside the network in order to actually test... some routers don't like internal NAT routed out and back in... I'm lucky that mine doesn't do that .. I had one like that years ago though

silent venture
#

NAT detection failed 👀

vagrant turtle
#

One other thing that happened to be a PITA for me recently, is that SatisFactory didn't shut down properly one time, and the port was blocked, and the game would not fire up ... I had to reboot the whole server to release the port. But at least that VM has only the game server on it.

silent venture
#

Not sure what the TCP connection is. Random pings if anything though.

vagrant turtle
#

yeah, that's windows... mine is a linux vm 🙂

#

yeah, tons and tons of random hits to every IP in the world happens.. people trying to see if they can hack their way in. heheh

silent venture
#

Nah, that's one of their servers from what I can tell. Always hits the same IP

vagrant turtle
#

hopefully the port range helps getting you going.. I have to take off to go to the store.. don't need wife agro HAHAHAH..

silent venture
#

Thanks for the help. I'll let you know if I get it going. No dice on the range 😄

vagrant turtle
#

bummer on range 😦

#

ok, I just looked at my ranges.. not just one... but 10
7777 - 7787
15000 - 15010
15777 - 15787

#

ok, now I'm out the door heheh

limpid granite
#

are vehicles basically unusable on dedicated servers?

silent venture
crystal eagle
silent venture
crystal eagle
#

Google on the server pc 'what is my ip' that is the ip you should connect exteranlly

silent venture
#

I have to use the local IP because it's on LAN. External people have to use my external IP.

#

Works on the 192 IP, port showing closed on external

crystal eagle
#

Yes

#

Make sure in windows firewall rule in allowed

silent venture
#

Maybe it's working properly and just doesn't show the port open..

crystal eagle
#

Apologies but cannot I'm editing music atm 😦

silent venture
#

No worries.

crystal eagle
#

PM me the ip and pw (you can change later) see if i can see the server

silent venture
#

Sent you a message. Not worried about it being people knowing it - I use it to host a bunch of stuff.

crystal eagle
#

kk

silent venture
#

@vagrant turtle TURNS OUT I've had it right the whole time, but the port checker must use TCP instead of UDP. I'm so smart that I'm stupid.

crystal eagle
#

Its always better checking it physically rather than software port checkers

vagrant turtle
fading quarry
#

Anyone have experience with using ramdisk as storage location for saves? Getting quite annoyed with the very long savetime and constant connection drops due to saving timing out connections. (I know I can increase the connection timeout)

#

Or rather have anyone attempted this, and is there actually any benefit by doing so.

broken ravine
#

Writing the file is a tiny fraction of the save time. The vast majority of it is the serialization of the world, so faster disk IO isn't going to shave much of the overall save time off.

fading quarry
#

Thanks, I thought that was the case and not a IO problem when writing.

broken ravine
#

I think to really speed it up, they're going to need to come up with a method to snapshot game state and fork off a process to serialize and save it. Factorio (on linux only) does that by having the server just fork and save while the parent keeps going. Windows doesn't support that kind of forking though, so I could see CoffeeStain not wanting to have super inconsistent behavior between platforms.

#

welcome. Hopefully they come up with a way to handle by 1.0 in a better way. The biggest annoyance for me during saves is building belts as they'll show up after the save, but most of the time aren't actually connected. 😦

#

Very annoying to go back and find out which of the 15 belts I built 30 minutes ago aren't working.

fading quarry
#

belts and pipes during saves for sure is troublesome, especially when they seemingly look all correct.

broken ravine
#

Yup, pipes are even worse since it's not quite as easy to see the issue as belts that are empty.

late coral
#

When the save happens, I just delete the pipes that are placed during that process. (You can tell by the blue hologram being displayed for more than 1 second, before the actual build happens).

gritty elbow
#

I figured out where to find the Game.ini file for hosted servers. Since no one can seem to answer the question "Where can I find game.ini?", it can't be accessed from FTP. You need to access it from the hosts FIle Manager.

unknown.png
median kindle
blazing halo
silent venture
blazing halo
silent venture
#

Have someone try to connect to your server. If you opened the firewall and router you're good.

blazing halo
#

I forwarded the ports correctly (tested with external testers, netcat - all three ports are fine). There's no firewall set up on the server itself. Still can't connect from external or via NAT hairpin.

#

It's as if the server is trying to do something funky with the NAT, fails, and then somehow falls back to a configuration that happens to be wrong, or something.

#

Or, it could also be reporting my internal IP as external - which obviously wouldn't work. I get a log error about it being unable to resolve some host address - if it's trying to use its own DNS servers, that might explain it (I require all software to either go through my DNS, or sulk in the corner).

silent venture
blazing halo
#

Debian Linux

#

(on both, actually)

silent venture
#

iptables has 1500, 15777, 7777 open?

#

Never used Debian, I'm assuming it's the same as CentOS

#

What helped me was noticing I could connect from Lan without problems. If you can get on through Lan, it's port forwarding at fault

blazing halo
#

15000, 20000, 20001, but yes

silent venture
#

Vc? I'll walk you through it

blazing halo
#

And I know it's somehow the forwarding's fault - I just don't know how.

silent venture
#

check the server log

#

See if it's binding the correct ports

blazing halo
#

Nothing that would indicate a specific failure.

#

Um, binding works - I can connect via LAN fine.

silent venture
#

Oh then it's a port forwarding issue 100%

#

What type of router are you using?

blazing halo
#

Also Debian.

silent venture
blazing halo
# 
[2022.05.07-02.26.55:647][  0]LogNet: Created socket for bind address: 10.0.1.22 on port 20001
...
[2022.05.07-02.26.55:647][  0]LogNet: GameNetDriver EOSNetDriver_2147482558 IpNetDriver listening on port 20001
...
[2022.05.07-02.27.37:534][  0]LogNet: Created socket for bind address: 10.0.1.22 on port 15000
...
[2022.05.07-02.27.37:534][  0]LogNet: IpNetDriver_2147452409 IpNetDriver_2147452409 IpNetDriver listening on port 15000

So yeah, looks like no issues there.

blazing halo
silent venture
#

You did this then?

blazing halo
#

Not using UFW. Using iptables.

silent venture
#

screenshot?

blazing halo
#

Did this though:

iptables -t nat -A PREROUTING -m mark --mark 0xd111 -m conntrack --ctstate NEW -p udp --dport 15000 -j DNAT --to-destination 10.0.1.22
iptables -t nat -A PREROUTING -m mark --mark 0xd111 -m conntrack --ctstate NEW -p udp -m multiport --dports 20000,20001 -j DNAT --to-destination 10.0.1.22
#

(mark 0xd111 is forwarded from WAN, to distinguish from NAT hairpin which is 0xd222)

silent venture
#

Did it apply to the iptables file? I'd vi into it and make sure it's there

blazing halo
#

Yes, it's there. Further, if I shut down Satisfactory and run netcat on those ports, I get replies as you'd expect.

#

Satisfactory does not reply, however, when running on those ports.

#

(you can nc -u server.ip.address bconport and if you send anything 4-byte, it should reply with BCON)

#

Also, this configuration is exactly the same as for every other server I run - ARK, Conan, Atlas, OOR, 7Days, DnL, Memories of Mars, Cryofall... all of them work, just not Satisfactory.

silent venture
#

Okay.. dumb question

#

Reboot?

blazing halo
#

While I doubt it'll do much, I can do that. Sure, let's give it a shot.

silent venture
#

I've seen so much shit fix itself on a reboot, although I'm betting it's a configuration issue moreso than fluke

#

Could also try dmz temporary to see if it works.

blazing halo
#

You know what, since I'm already about to reboot, I'll run updates.

#

... or, I would, if there were any. :D

#

Okay, let's see.

#

Okay. Router rebooted. Now the server.

#

Hm, interesting. Now I'm paying more attention, and here's a fun fact:

#

packets make it to the server.

#

Replies don't.

#

So it's clearly a misconfiguration that only applies to UDP, somehow. Let me dig into this.

blazing halo
#

Ah, no. It seems that the remote I was using didn't want to hear the replies. Tried another remote and got replies immediately.

#

It seems that it's replying to remote queries. I can see both BCON (on 15000) and GAME (on 20001). It still doesn't work.

#

Here's what I'm getting in the log:

# 
[2022.05.07-03.09.29:892][188]LogNet: Server accepting post-challenge connection from: 10.0.1.1:65143
...
[2022.05.07-03.09.29:892][188]LogNet: NotifyAcceptedConnection: Name: OnlineBeaconHost_2147452409, TimeStamp: 05/07/22 05:09:29, [UNetConnection] RemoteAddr: 10.0.1.1:65143, Name: IpConnection_2147451283, Driver: IpNetDriver_2147452407 IpNetDriver_2147452407, IsServer: YES, PC: NULL, Owner: NULL, UniqueId: INVALID
[2022.05.07-03.09.29:892][188]LogNet: AddClientConnection: Added client connection: [UNetConnection] RemoteAddr: 10.0.1.1:65143, Name: IpConnection_2147451283, Driver: IpNetDriver_2147452407 IpNetDriver_2147452407, IsServer: YES, PC: NULL, Owner: NULL, UniqueId: INVALID
[2022.05.07-03.09.31:851][305]LogDTLSHandler: Warning: Empty certificate identifier
[2022.05.07-03.09.32:979][367]LogDTLSHandler: DTLSCertVerifyCallback: Verified
[2022.05.07-03.09.38:958][724]LogDTLSHandler: UpdateHandshake:  Handshaking completed
[2022.05.07-03.09.46:929][201]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.05.07-03.09.46:929][201]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.05.07-03.09.46:929][201]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.05.07-03.09.46:929][201]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.05.07-03.09.46:929][201]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
...
[2022.05.07-03.10.02:734][799]LogNet: Warning: UNetConnection::Tick: Connection TIMED OUT. Closing connection.. Elapsed: 27.05, Real: 32.83, Good: 32.83, DriverTime: 338.54, Threshold: 30.00, [UNetConnection] RemoteAddr: 10.0.1.1:65143, Name: IpConnection_2147451283, Driver: IpNetDriver_2147452407 IpNetDriver_2147452407, IsServer: YES, PC: NULL, Owner: NULL, UniqueId: INVALID
#

So basically it looks like the client either doesn't get the replies, or ignores them.

#

I would say Warning: Ignoring handshake packet received after completion. is a fair clue here, but as to what's causing it? Still no clue.

blazing halo
#

So, bottom line:

  • ports are forwarded (and it works)
  • I can netcat into the server, and get BCON on port 15000, and GAME on port 20001 - so Satisfactory can clearly both hear me, and reply to me
  • I can also verify that the server is indeed listening on the specified ports (query 20000, game 20001, beacon 15000)
  • for whatever reason, connecting to the server doesn't work - it looks like the client isn't getting the responses, or is ignoring them, causing the server to time the connection out
blazing halo
#

So unless there's something really funky going on with port 20000 (and basically every port set up via -ServerQueryPort=<num>), I have no further clues.

#

... waaaait a second.

#

Connecting on port 20000 nets me: LogServer: Opening a new route from '10.0.1.1:52134' to '127.0.0.1:15000'

#

Is the server rerouting the connection, and attempting to reply from 15000 on attempts at 20000?

#

Ah, only if I try a BCON there. Trying GAME gets me 20001 in the log.

#

Now would be a very good point to ask a dev about this, because it's clear that something weird is going on. If I assume that the server is attempting to reply from a port different from the one it received communications on, the issue makes perfect sense - since it's not the remembered port, iptables doesn't mangle it (because it doesn't know how), and it never arrives. The client thinks the server is dead (never receiving handshake accept), and the server thinks the client is not listening (>> ignoring handshake packets after completion).

supple night
#

Where can I find the Game.ini file on linux?

frosty gale
#

out of curiosity has anyone found any better instructions on how to setup a home server, esp with Epic Store version of game? I had it working for a while on local lan....except now it doesn't and I don't think I ever got it to work with external IP address (and local game on that PC kept giving me InvalidSession for the ID when I asked it in that game. Maybe in a few months they'll have improved the instructions. Then again, I worked out leaving my PC on 24/7 would probably cost me at LEAST $10 a month in electricity use of the PC itself AND the AC running a little extra to deal with the heat from the PC. thinking of either trying out the CubeCoders AMP app which is a flat $10-ish once, or going with 4netplayers.com seems to be the most inexpensive hosting solution for up to 8 players.

silent venture
fading quarry
#

@frosty gale Generally speaking setting up any game server from home is entirely dependent on each individual setup in most cases its a matter of how networking is set up that is the main culprit of any trouble, double nat is by far the most common problem, and heck even triple nat in some instances. (2-3 routers) more an more ISP also start to lock down their routers leaving little to no options for home users to easily customize port mapping.

frosty gale
#

yeah I have a LinkSys Velops system inside my ISP modems thing so that prob complicates it fro my other PCs trying to connect to it outside theVelops network but still internal. (game host Pc direct connected to ISP equipment to try to reduce complications) I Did port forwarding for 7777-78whateever UDP to the host plus 15k and 15700 both udp and tcp jsut in case. meh. I'm kinda over it. I also made SURE the firewall rules on that PC allowed the traffic in but like I said, kinda over it. I'll either Pay $10 for cubecoders AMP program to try it out or I'll say eff it and host remotely. considering the electrical costs at home in the summer, hosting remote might actually be less expensive.

supple night
#

Getting an error when trying to update, anyone who can help me out with this one?

steam@Satisfactory:/usr/games$ steamcmd +force_install_dir "/home/steam/SatisfactoryDedicatedServer" +login anonymous +app_update 1690800 validate +quit
WARNING: setlocale('en_US.UTF-8') failed, using locale: 'C'. International characters may not work.
Redirecting stderr to '/home/steam/Steam/logs/stderr.txt'
[ 0%] Checking for available updates...
[----] Verifying installation...
Steam Console Client (c) Valve Corporation - version 1648077083
-- type 'quit' to exit --
Loading Steam API...OK

Connecting anonymously to Steam Public...OK
Waiting for client config...OK
Waiting for user info...OK
Update state (0x0) unknown, progress: 0.00 (0 / 0)
Error! App '1690800' state is 0x204 after update job.

frosty gale
#

Is it possible to run a dedicated server on the "regular" build and not the experimental build? Once you get the server config setup

vapid zephyr
#

If the server auto updates while a player is in a tube, it crashes that players game. Any Ideas?

vast mural
#

can dedicated servers have mods yet?

quick mist
quick mist
quick mist
vocal latch
#

Okay I've searched around and I feel dumb not finding an answer: I'm running a server from an extra computer at my house. Everything is running fine except it won't accept the password to claim the server. I enter the password and nothing happens, no error or anything. Is there a way from the server computer to reset the password? It's just running w10 and running through steamCMD

wicked knot
vocal latch
#

I really missed that. Thanks @wicked knot

unkempt crescent
#

Anyone have recommendations for server hosting services? was thinking aws but mby you know about a better offer/setup?

feral osprey
#

Anyone else running into Vehicle Glitches? We can drive the vehicles partially but the vehicles seem to be duplicating now, visually the explorers for example leave behind a non-interactable duplication of the vehicle

#

Bringing the save into single player seems to solve the issue but we are getting werid issues with vehicles now

unkempt crescent
#

Jesus, dont go to shockbyte for server hosting. Marketing servers for satisfactory but supplies specs below minimum requirements. Scummy af

vagrant turtle
frank rapids
#

I’m not sure about the others but the port 15777 can be changed

vagrant turtle
#

yeah, I believe one of them can be, but 15000 I believe is one that can't.

blazing halo
#

The Wiki clearly states that the only port that can't be moved is the beacon port, which is hardwired to 15000. Every other port, as I understand, is freely assignable. Unless the Wiki is wrong (or imprecise?) somehow...

#

From what I can tell, the assignments do go through (both netstat and ss agree that the game is listening on the specified ports). But I can perform a quick test here, to see if it does indeed try to communicate from the default ports - by blocking ports on the server machine itself.

#

Okay, so my setup right now (on the server machine) is as follows for this test:

Chain INPUT (policy ACCEPT)
target     prot opt source               destination
ACCEPT     udp  --  0.0.0.0/0            0.0.0.0/0            multiport dports 15000,20000,20001
ACCEPT     tcp  --  0.0.0.0/0            0.0.0.0/0            tcp dpt:22
DROP       all  --  0.0.0.0/0            0.0.0.0/0

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination
ACCEPT     all  --  0.0.0.0/0            0.0.0.0/0            state RELATED,ESTABLISHED
DROP       all  --  0.0.0.0/0            0.0.0.0/0

Explanation of the rules: allow input for port 22 (SSH), and then 15000 (beacon), 20000 (query, assigned), 20001 (game, assigned); allow output ONLY on ports that are "related" to an established "connection". Drop everything else. This means the machine won't be able to make connections of its own to anywhere, unless it was first contacted on that port.

#

The server responds to queries. I can make a connection, and play on the server.

#

I can also confirm (via nc) that I can't send anything out from 7777.

#

So, that theory is busted. Both the QueryPort and GamePort are freely assignable, and communication happens only on those ports (i.e. these ports are sufficient for administration and gameplay to happen).

#

If I had some information on the protocol in use, I could investigate further (for instance, I could see what the "query" response is). Sadly, Satisfactory doesn't seem to use the same protocol STEAM uses for communicating with the various game servers -- as evidenced by the fact that I can't get a status on it in the "Servers" view regardless of what I input.

#

My current (and only) clue is that the server does reply to the client, but the reply never arrives or is otherwise ignored. The client tries to authenticate with the server a few more times, and then the connection times out.

#

I'm actually leaning towards the "ignored" part - if the reply is somehow malformed or contains bogus data, the client might ignore it. One such case might be that the client connects to "my public IP", but receives a reply claiming it's coming from "my server's private IP" - this mismatch could be interpreted as "bogus" and ignored, if this is indeed the case.

#

It would also align with the server stating that it can't detect NAT type due to some issue with name resolution. If the failing attempt is related to the server getting its public IP resolved (because how would it know it otherwise?), it could be the failure that's preventing it from working.

blazing halo
#

Interestingly, NAT detection probably won't work over the EpicGames NAT detection Satisfactory is using, because the CNAMEs direct to euc-NUM-prod.stun.euc1a.on.epicgames.com for me (where NUM is 1 or 2), which don't exist (NXDOMAIN). Further, I cant find anything that would exist in that namespace, from any nameserver. DNSDumpster reports a number of similar domains (mostly in use1a, which I understand is "US East"), but they also don't resolve.

blazing halo
#

Okay, so I did a little digging, and redirected the queries (via CNAME) to whatever jumps out on querying for turn.rtcp.on.epicgames.com (for me, it was turn-euc1.us.rtcp.on.epicgames.com). Not sure if that was an issue, but it's clearly not THE issue, as I'm still getting the "NAT detection failed" message, and I still can't connect to the server. Whatever the issue is, it's still unresolved.

#

If I had the damn query protocol, I could at least verify what the server is replying about itself. :/

whole dome
#

Sounds like a double nat issue 🤨

#

Can the client connect to a server not behind nat? And can you connect to the server from a client not behind nat?

blazing halo
blazing halo
# whole dome Sounds like a double nat issue 🤨

I'm not sure where this second NAT would be coming from. The WAN-facing interface has a public IP, and I can confirm that I can open reachable ports on it. I host a number of other game servers, and the setup works for all of them - except Satisfactory. The web server is also reachable, and it happens to be on the same machine.

whole dome
#

It’s the same issue FTP data used to have - it’s a nat/udp issue when both server and client is behind different nat gateways

blazing halo
#

Yeah, but if you forward the correct ports, the issue ceases to exist.

whole dome
#

Not entirely

blazing halo
#

(for FTP specifically: as long as you know to instruct the client to to request PASV instead of waiting for a back-connection)

blazing halo
#

Still, the point is moot: if I shut Satisfactory down and then run netcat in listening mode, I can "talk" to every port and get replies back.

#

So, whatever the issue is, I think it's specifically with Satisfactory.

whole dome
#

All I mean is just because you can connect to the port, doesn’t guarantee it will work if the way satisfactory uses udp doesn’t support server -nat- internet -nat- client

#

Since udp is connectionless.. at least with tcp you get a proper connection set up

blazing halo
#

Yes and no.

#

iptables relies on conntrack, and conntrack does track UDP connections.

#

(it considers each set of source IP + source port + destination IP + destination port to belong to the same "connection")

blazing halo
#

But just to be sure, I just did another test from a remote machine (over the Internet), and I can

a) netcat over the ports, receiving messages on both ends, and
b) once the game is running, test both Beacon and Game ports (one replies with BCON, the other with GAME); the only port that doesn't have "simple" replies is the Query port, which I don't know how to query (and neither does STEAM, it seems)

#

To make sure I got unskewed results, I also disconnected the VPN link between us, so that it didn't have a straightforward (non-double-NAT) line back to the server if it wanted to try that. Didn't affect anything.

#

So I'm back to the problem being with Satisfactory, somehow.

quick mist
#

I have a tool that polls the query port with the extremely limited documentation received so far on its protocol

blazing halo
#

Share? Or better yet, share docs?

quick mist
#

tool is in the same repository

blazing halo
#

Yeah, okay - I can see that reflected in the capture I made yesterday.

#

So yeah, it's clear that it's not using the same Query protocol STEAM uses for its games.

quick mist
#

no, not remotely

#

probably an artifact of it not having been developed for Steam initially

blazing halo
#

So, I assume it gets further game data (like game stage, runtime, player count) via BCON?

quick mist
#

or rather of having been developed to be platform-agnostic

#

either that or though message types I have not gotten specs for on the Query port

blazing halo
#

Does it reply with anything other than msg type 1?

#

Does it reply to anything other than msg type 0?

quick mist
#

Not that I have been able to elicit by poking it

#

it may not respond to messages not conformant with the protocol it expects

blazing halo
#

That may be the case. It doesn't respond to anything I send it through netcat, but it clearly responds when the game asks.

#

BCON does reply, and it sends a ton of data - I just have no first clue as to what it actually is, apart from a similar "check" value echoed back to the client.

quick mist
#

if the query protocol is any indication, the data are byte-encoded rather than being plaintext

blazing halo
#

No, I assumed as much, and I actually did figure one more part of it out -

#

there's a message counter in the message, but for some reason it's counting in multiples of 8,

#

and it seems to be separate for each message type?

quick mist
#

is it counting multiples of eight or is it big-endian when you are looking for little-endian?

blazing halo
#

Hm, I did entertain that idea for a short while there. But then it would mean that the Query port uses a different endianness?

#

Why.

quick mist
#

it'd be a FICSIT thing to do (:

blazing halo
#

Possibly...

#

Actually, it'd be mixing endianness within the message itself.

quick mist
#

that would be sanity-taxing

blazing halo
#

Ugh.

#

Yeah, because I get BCON, then the 8-byte sequence, and then what I assume is a 4-byte sequence counting in 8ghts :D

quick mist
#

wait, even if it were an endinness issue, it'd be counting in 16s, not 8s

blazing halo
#

But yes, if we assume big-endian, then the whole "8 byte sequence" ending in mostly 0's would make sense. (of course I meant little-endian)

quick mist
#

the numerical data such as the server version is sent little-endiian on the query port, but you know what they say about assumption (:

blazing halo
#

Hmmm... Actually, if you treat the "8 byte user data" as big-endian little-endian, the subsequent queries are rising values.

#

(for me: 2764404486, 2764404936 and 2764405602 - within about two seconds, all three)

quick mist
#

that's a pretty good correlation

blazing halo
#

Yup, they track. It's a timestamp.

#

2764404486 = 0.000000

2764405271 = 0.785 134
diff 331 = 0.331 055
2764405602 = 1.116 189

#

(left: value; right: time in capture)

#

The first value to the last one diff is 1116

#

It's a timestamp in milliseconds. But since when?

quick mist
#

abjectly guessing, but:

  • since server start?
  • since server start not counting paused time?
  • since server start, reset on the 24-hour-restart?
  • since 0000 local time?
  • since 0000 UTC?
blazing halo
#

Can't be since server start, because it's the first message sent by the client ever.

#

Can't be 00 of either time, because it works out to a little over 31 hours.

quick mist
#

ms of gametime since start of world?

blazing halo
#

How does it know if it's literally the first message the client sends?

quick mist
#

oh it's a client message. Hrmm

blazing halo
#

Yes, it's present in the first message.

#

And it tracks between them, so it must use some universal base.

#

Unless we're wrong about it being an 8-byte value - it might be a 2-byte value.

quick mist
#

be interesting to ask Bogdan about it when he's awake and not in DnD

blazing halo
#

As for the endianness, I suppose that makes little difference if it's straight memory structures. Because why not.

#

It'll come out OK on either end. ;D

quick mist
#

true, if the endianness is mixed and it's coming straight from Unreal, the stinkeye goes to Epic, not CSS

blazing halo
#

Yeah, I know.

#

Though if you've dabbled in the messy hell that is blueprints, you're well-aware that this is the least of their offences... :D

quick mist
#

:snrk: I can imagine

blazing halo
#

Gimme a sec, I actually did a simple parser on .uasset files (looking for stuff like blueprints, asset names, things like that). Let's see if I have comments.

#

Okay, so in ARK assets: 2-byte values are little-endian, 4-byte values are big-endian, and 8-bit values are little-endian.

quick mist
#

🥴

blazing halo
#

Strings are also encoded as UCS-2LE.

#

:D

quick mist
#

at least they're not CP437 (:

blazing halo
#

You don't know encoding hell until you've tried decoding Japanese.

#

Especially if you don't actually read/write it. :D

quick mist
#

I've come close. I was one tasked with parsing a highly irregular XML file feed using nothing but awk

blazing halo
#

> using nothing but awk

There's your problem right there.

quick mist
#

with respect, I disagree- awk is great -- if the input can be parsed with regular expressions

#

XML, being by definition not regular, is not the data for that toolset

blazing halo
#

> with regular expressions

#

:D

#

How did that saying go?

#

"If you have a problem, and you're trying to solve it using regex, you now have two problems."

#

Yes, web-related mostly, but heh.

quick mist
#

I've heard the same, but with s/regex/JavaScript/ and s/regex/Perl/ and...

blazing halo
#

True, true.

#

I believe it was originally said about trying to parse HTML DOM with regex.

#

Which, admittedly, is a problem you should really solve differently.

quick mist
#

it's all about using the right tool for the right job. Regex is a great tool but not for HTML/XML

blazing halo
#

But hey, we now have luxury, to be fair. We have cli DOM parsers, simple DOM parsers, complex DOM parsers, running the DOM in a virtualised browser...

quick mist
#

indeed

blazing halo
#

And parsers that return meaningful error messages instead of crapping their pants on a misplaced >.

quick mist
blazing halo
#

Actually, I was wrong about it working out to over 30 hours.

#

It's actually in days.

#

If I ignore the last three digits, it's 35.5 days (I just ran a log a moment ago).

#

So that gives me the value of 0 at 2022-04-02T19:15:12+00:00

#

Don't know the significance of that date.

#

Also, @quick mist corrected my words above, I of course meant little-endian. Not sure how I got confused there, but it can be confusing, so I got confused.

blazing halo
#

But yes, now I'm sure: unless it switches bit endianness, it does count in multiples of 8. I'm seeing 00 00, then 00 08, then 00 10, then 00 18 and so on, on subsequent messages. When I lay it out in bits, it makes no sense unless they scramble the bits on purpose (this goes beyond bit endianness or high/low nibble scrambling), or there's something clever going on. If I remove select bits, then bit-reverse some nibbles, and rearrange the whole thing - I get one working counter (on 11 out of 21 bits). What the hell.

quick mist
#

I wonder if it mightn't have been some sort of data-interleaving for some hyperoptimized data parser or something

karmic merlin
#

Currently running a server with a buddy, we are going for a mega base. How much RAM would you recommend? We are currently using 95% of the 8 GB. Upgrade by 5 GB or 10 GB? We are facing little lag issues at the moment

quick mist
#

if the lag issues are not caused by swap utilization, you're fine. my experience is that the game server seems mostly CPU-bound.

#

Look at the system metrics to see where the bottlenecks are

late coral
#

The game server is very CPU centric and really benefits of higher single core performance.

cyan rock
#

speaking of that, planning on moving our at home dedicated server from a xeon to a threadripper, do you think thatll make a difference? just starting at tier 5 not much built but the belts are really out of sync and ghosting alot

#

the xeon is pegged at 100% on one core rn so figured to give the threadripper a try for higher clock speed and better cpu architecture

quick mist
#

what are the individual core speeds of the relevant CPUs?

cyan rock
#

going from a E5-2650 @2.0ghz to a threadripper 39555wx @4.3ghz with single core boost to 4.5

#

i think thatll do

quick mist
#

that will serve you quite well indeed, yes

shrewd zinc
#

My dedicated server was timing out for a remote user yesterday, it was only him and I on the server at the time. I have a 10mbps upload connection, he was able to reconnect immediately. System has a 2.9Ghz dual core processor with 12G of RAM. I think I need to upgrade the CPU, IDK, thoughts?

quick mist
#

if it was only that person, every time, getting kicked, I would suspect an issue with that person's connection or client. That said, what is the resource contention picture like on your server?

shrewd zinc
#

it's linux, it's not running anything else, it's an older mini system (HP 400 mini I believe), SSD for the main drive and linux is baremetal installed. I had an SSH open to it, running TOP, I didn't see CPU peak past about 85%

#

I was getting some strange stuttering around the same time, seemed to persist longer than an autosave would.

#

I also can't quite figure out how to get it running as a service, but I want to test a few things before I dive in too deep.

quick mist
#

Getting a little bit of a performance hit when it gets to be autosave-o'clock is a known issue- that is CPU- and storage-bottlenecked. But if you're not seeing your CPU get to capacity the gain you'll see from an upgrade is going to be limited

shrewd zinc
#

thanks. I'll keep an eye on it, if I start pushing more than 95% load or so, I'll consider making the change. until then, Chalk it up to internet shenanigans.

trim island
#

Hello, I'm having some trouble running the DS on my ancient workstation/server

#

It's a Xeon X5680 (equivalent to 1st gen high-end i7), 24 GB RAM, pretty much a fresh install of Windows 10, I've ran the redist and I was even able to run the UE5 editor on it, but the Satisfactory DS simply does not launch at all

#

On my primary PC, a Ryzen 2700, 32 GB RAM it runs, but I'd rather have it on the server so that it can run 24/7

#

One maybe important thing to note, as I have no idea if this really impacts it, is that my primary PC uses a network connection to the server to access hard-drives, since the server has significantly more ports and space for them, and as such, with the Epic launcher on the main PC, I installed the DS on the server drive, and tried launching with that

#

However due to this setup, connecting the server to the Epic launcher yields timeout too often so I can't confirm if that's the missing link

#

To avoid some confusion, I refer to

  • "Satisfactory Dedicated Server" as "DS"
  • And to my "server PC" as "server"
#

So, essentially, my problem comes down to: Is this an unsupported CPU or should I try and get the Epic launcher to work? If anyone has experience running on a similarly ancient CPU I'd appreciate some confirmation

quick mist
#

Do you know what CPU opcode extensions your older CPU supports?

trim island
#

I can download a CPU-Z and see what it supports

quick mist
#

I thought there was a list of required support on the Wiki but it is either a misrecollection or was edited out

trim island
#

I was wondering about that too, the requirement of "Recent x64 processor" on the wiki does not help, but the store page seems to have copy-pasted the requirements of the game itself

quick mist
#

Looking through the history now

#

which is not terribly convenient to do on that wiki engine ):

trim island
#

While you do that, main PC supported instructions, will need a moment for server's

unknown.png
#

A noticable number fewer

unknown.png
#

The thing that perplexes me is that the server can run UE5 of all things so Satisfactory shouldn't have been an issue in my mind

quick mist
#

oh the old one does not support x86-64? Pretty sure that's a core requirement

trim island
#

It absolutely does, it's a 64-bit OS

quick mist
#

I vaguely recall mention of SSE4A but am not sure if it's Satisfactory or another game I help out with

#

so EM64T includes basic x64 instructions including a lot of amd64, but it's not comprehensive. Comprehensive support is IIRC what x86-64 implies

#

so one of those two may be a roadblock

#

DS simply does not launch at all
Also it might be helpful to see what if any error output there is when it fails to launch; that may lend insight into the exact issue

trim island
#

Takes a few seconds and goes back to this

unknown.png
#

When running normally (on main PC) the console "hangs" and the server window opens

quick mist
#

try it without the -log or -unattended options so that the output goes into that terminal

trim island
#

Didn't help either

quick mist
#

but also no output? Shame

trim island
#

Even tried changing multihome to one of the various IPs

quick mist
#

yeah sounds like it's not getting as far as the Unreal engine itself to even get to dealing with the network stack

#

the lack of an affirmative confirmation is suboptimal of course

#

on a lark it didn't happen to drop any log files did it?

trim island
#

I could reroute the network again but it'll destroy my latency to the drives, which is rather essential

#

To at least have the Epic launcher running and perhaps the game itself

quick mist
#

yeah I'm not going to ask you to do that

#

but it'd be worth checking for logfiles

trim island
#

No log files were written from launching on the server

quick mist
#

drat

#

I was hoping to get some actual confirmation about my CPU incompatibility hypothesis, but alas.

#

Are you running Windows 10 or better?

trim island
#

Yes, latest version even

#

Perhaps there's a redistributable not included with the generic UE4 package but I'm rather doubtful

quick mist
#

okay. This is kind of stupid, but just for grins, you could try installing WSL, using the Linux version of SteamCMD and seeing if the server at least gives you some error message from there

#

(WSL is the Windows Subsystem for Linux- basically a native Linux distro running inside Windows as an application-- shockingly not through emulation; it's an odd beast)

trim island
#

I don't require a Steam purchase of Satisfactory for that, right?

quick mist
#

Correct, the dedicated server is free on Steam

#

the Wiki has the directions for using SteamCMD to acquire it

trim island
#

Just making sure since some games' servers aren't free

#

Or well, require you to login with a purchased game

quick mist
#

not an unreasonable question, but I would never tell someone to spend money to troubleshoot something, especially for such an off-the-wall idea (:

#

but yeah, you don't even have to log in; you can get it by logging into steam anonymously

#

tl;dr:

N:\Path\to\steamcmd.exe +login anonymous +force_install_dir N:\GameServers\SatisfactoryServer +app_update 1690800 -beta public validate +quit
trim island
#

Downloading now, will report later

quick mist
#

good luck!

trim island
#

Actually I'll retry tomorrow since it's a bit too late and I want the server to be done with things before restarting to enable WSL

shrewd zinc
#

Just commenting on the above: The X5680 does have SSE4.1 and SSE4.2 but lacks AVX. I had a Xeon system of the same generation I had to retire since I needed to run software that required AVX. It's a shame, those systems are powerhouses.

#

I'd be willing to bet that the dedicated server has a require AVX flag somewhere. It's likely that mehtehtrollface will find illegal operation errors in his windows logs.

#

the official minimum specs require a 3rd gen Core i5 (specifically citing the i5-3570), which is an AVX capable CPU.

quick mist
#

I should have thought of the Event Viewer, that's a good call

#

I'll add that to the Wiki about the specific minimum Intel CPU spec

shrewd zinc
#

Yeah, I upgraded from a Xeon L5460 to a Xeon E5-2650L, and I haven't hit a problem since. Dunno why everything seems to require AVX now, but things do.

quick mist
#

probably from the department of "this is the default"

shrewd zinc
#

I'm not even totally sure how AVX helps, but okay, who am I to question the software overlords? I'm just a lowly IT support person.

quick mist
#

Based strictly on AVX, this may be helpful for future conundra:

unknown.png
shrewd zinc
#

Agreed. I'm unsurprised for the lack of 32bit support, even the dedicated server takes 4-6 G of RAM depending on the situation. but I'm hoping ARM64 support ends up happening eventually. I'm not holding my breath on it, but it'd be nice to load up a dedicated server on an SBC as long as it has the memory required.

quick mist
#

Time will tell but as you said, I'm not holding my breath either. And if they do add it, I can't wait for people complaining about RasPis melting

shrewd zinc
#

yep. probably won't run decently on a pi unless you OC it to high hell, and slap one of those overkill cooling towers on it.... even then, once the map gets big enough, the game will probably lag out every time it does an autosave.

quick mist
#

The mind reels at how imperformant saving to an SD card would be || yes I know you can connect a USB cable to an SSD but I would bet a dollar it would not be up to SSD speeds ||

shrewd zinc
#

The only way it would work well is to use one of the non ras-pi SBC's that have higher clocks and an NVMe port.

#

(or at least a SATA SSD or something)

fading quarry
# trim island Hello, I'm having some trouble running the DS on my ancient workstation/server

I have been running Dedicated server on a Xeon X5650 for half a year now with multiple Operating Systems both Windows and Linux and with and without a hypervisor. All i can say is that Xeon X5... series is not the best CPU as it has poor single thread performance compared to even the cheapest newer gen CPU's out there. So expect quickly to encounter a huge loss of tickrate and every problem that comes with that.

shrewd zinc
#

it runs on an X5650?

fervent carbon
#

I run it on a e5-2680 inside an LXC container and the server runs fine

#

I think it's usually between 26-30 tps, which is not ideal but it is playable

shrewd zinc
#

That CPU has AVX.

toxic sentinel
#

Hi everyone , is someone able to help me to have my friend join my server ?

#

I can join it from abroad but my friend can't

#

When adding the server he see it offline

#

no matter what we do

late sandal
#

The (epic version of the) dedicated server is driving me crazy.
I can reach my server from the LAN, but it's unreacheable from my external IP
I checked the wiki and forwarded the ports to the machine running the server.
This machine is hosting a valheim server (and some others in the past) but won't be reached by satisfactory.....

#

ports declared, ports forwarded to the server, traffic allowed on these ports everywhere, ports listened to on server

unknown.png unknown.png unknown.png unknown.png
#

everything matches

#

and there's no other service using these ports either declared on the router

fading quarry
# shrewd zinc it runs on an X5650?

Yes, however once a save reaches a certain threshold its like hitting a brick wall and tickrate plummets to unplayable state, and this unfortunately happens very quickly. I have given it a dozen attempts to increase performance such as: Hyper Threading on/off, dedicating cores for different threads (messy AF as main thread likes to jump all over the place), insane amount of memory (192GB) different OS's: Ubuntu 18.04/20.04/22.04, Debian, Windows Server 2019, Windows Server 2016, Windows 10. Each and every both as virtualized and bare metal.

trim island
#

Or at least be able to run for the foreseeable future

#

We don't exactly progress too fast in the save we have so I hope to have it running for a while longer on the server

fading quarry
#

Main source of trouble we have is when we are 3-4 players playing at same time and building in close proximity to each other, CPU will then just grind constantly at 100% until server restart. I have it setup to restart daily to mitigate the worst of it, and also increased auto save to 60 minutes and connection timeout to 120 seconds instead of 60 so clients don't get kicked when game saves.

trim island
#

We'll be 3 at most and usually we spread the tasks so that we don't really have more than 2 building at once so I'm not overly concerned about that

late sandal
#

I tried the Steam version of the dedicated server and it works just fine (once I bind it with -multihome, otherwise it doesn't find an interface -_-)

#

I wonder how I can start the epic version with -multihome

#

probably with just the same command as the steam version but ibn the Epic folder

trim island
#

Opening the event viewer and there's nothing relating to the DS at all

unborn wind
#

I’m running my server on dual e5-2690v2s, and currently everything runs fine, but what issues will I run into once my friends and I expand our builds to other biomes, and make them overall bigger? The processor has 3ghz base and 3.3 boost, server is running ESXI, with a windows server 2022 data center VM… thoughts on how it’ll run in the future? TIA!

#

Are there any tests I can perform to get a guesstimate on how well it’ll perform as things get bigger?

#

Please ping when respond 😁, might be a bit slow to respond myself sorry

trim island
#

Update: The DS ran after I installed it from the server itself, which took a number of attempts and hopes and dreams but it seems like there's something in the installation process that is missing if you don't install it natively

#

Attempting to launch the old installation results in the Epic launcher being rather forcefully opened and then nothing more

late coral
#

Recently my friend updated the CPU of the Dedi to a 4790K, and this resulted in a significant performance boost. All issues we had with performance are now gone, except the short autosave delay.

late sandal
#

Running mine on a dual E5-2658 😅

unborn wind
late sandal
#

Barely notice any load on the CPU 😂🤣

opaque forge
#

Hey guys im new to satisfactory i was wondering if anyone wanted to play with me and help me learn

unkempt crescent
#

Hey! Does anyone knows what may be causing this issue when trying to load a bigger save?

ULevelStreaming::RequestLevel(/Game/FactoryGame/Map/GameLevel01/Tile_X0_Y0) is flushing async loading
09.05 19:44:43 [Server] INFO [2022.05.09-19.44.43:440][ 0]LogStreaming: Display: ULevelStreaming::RequestLevel(/Game/FactoryGame/Map/GameLevel01/Tile_X0_Y1) is flushing async loading
09.05 19:44:43 [Server] INFO [2022.05.09-19.44.43:520][ 0]LogStreaming: Display: ULevelStreaming::RequestLevel(/Game/FactoryGame/Map/GameLevel01/Tile_X5_Y2) is flushing async loading
09.05 19:44:44 [Server] INFO Killed
09.05 19:44:44 [Server] INFO Game process exited.

quick mist
#

is this on a server you administer or one you utilize but is administered by a vendor?

That looks like the kernel is killing the server, either the OoM killer or the user resource quota monitoring killer

#

next time it happens you might be able to see something interesting if you immediately run /bin/dmesg

misty valley
#

sorry just saw this in console. made me chuckle..
IsAwake isAwake = ture

#

and equiv..

#

IsAwake isAwake = flase

#

so if it possible to be neither, would it be nlul?

opaque spruce
#

Can I make my server auto restart when it crashes or my internet goes out for a quick second?

crystal eagle
#

Yes friend of mine wrote this script for me:

#

@echo off
::run satisfactory server batch file

:run
@echo Run Satisfactory Server
FactoryServer.exe -log -unattended -NOSTEAM

IF %ERRORLEVEL% EQU 1 goto end
IF %ERRORLEVEL% EQU 0 goto end
IF %ERRORLEVEL% NEQ 0 goto error
goto end

:error
@Echo server exited with error code %ERRORLEVEL%
goto run

:end
@echo exiting

#

replace what you have in your .bat file

#

Add the file to the windows startup, eg I have a windows schedule for a restart every morning

opaque spruce
fervent carbon
opaque spruce
opaque spruce
#

I got something that restarts my pc when internet goes offline

round glade
#

do we need to be on the normal or experimental version to use dedicated servers?

late coral
#

Dedicated servers can be used with both branches.
But both server and client need to be the same build.

plush ravine
#

does the server stay running even after i turned off my pc? if not how do i keep it up

quick mist
#

The physical computer that is running the server does in fact need to still be on for any software, including but not limited to the Satisfactory Dedicated Server, to be available.

opaque spruce
#

Is there something that I need to do to solve the issue of the conveyor's visuals messing up?

plush ravine
#

its not letting me change my session type , every time I try to switch it to friends it reverts back to private (using dedicated server btw) ,

#

also says "invalid session"

quick mist
plush ravine
quick mist
#

is the server running on your friend's computer?

plush ravine
#

no its running on mine

quick mist
#

okay, he is using the wrong address

#

127.0.0.1 is the network address for "Me"

plush ravine
#

thats a screenshot from my pc

quick mist
#

you need to give him your LAN IP address, not the localhost address

plush ravine
#

ohhh ok

quick mist
#

the easiest way to see your LAN address(es) is with this powershell command:

Get-NetIPAddress -AddressFAmily IPv4 | Format-Table -Property IPAddress
#

most of the time, the address you want will start with 192.168

plush ravine
#

also says Nat detection failed for me

quick mist
#

Let's perhaps start from the beginning. I say this because Session IDs and Types are not at all relevant to dedicated servers.

#

Step zero: Do you have a dedicated server installed and running right now?

plush ravine
#

yes

quick mist
#

And it is running on the same computer you are using as the game client?

plush ravine
#

yes

quick mist
#

Swell. What is your LAN IP address?

plush ravine
#

dm

quick mist
#

(let the record reflect that J has identified eir IPv4 address)

#

okay, in the game client, go to the Main Menu, and once there, go do the Server Manager in the lower left. Click Add Server, and punch in that LAN IP address. Leave any port settings at the default values.

plush ravine
#

ok

#

done

quick mist
#

okay, good. now, in the Server Manager, you should see the entry you just added for your server.

plush ravine
#

yes i see it

#

game ongoing

quick mist
#

if you click on it you may be asked for a password if you had previously set one up

plush ravine
#

yes done

quick mist
#

cool- you should see a Join Game button

plush ravine
#

ok cool seems to be working for him now

#

thanks a lot for your help

quick mist
#

you are quite welcome

spring sage
#

Need help hosting a server on my computer I have epic games version

round glade
#

i tried setting up a server, and while i am able to join it, my friends are saying the server appears to be offline

#

i think i port forwarded properly, is there something else that's causing it?

plush ravine
fervent carbon
fervent carbon
#

Has anyone managed NAT penetration through an openvpn instance? I have a pfsense firewall running infront of my satisfactory server which I could use to route the traffic through an OpenVPN server running on Oracle Cloud, is it possible to redirect the traffic on the required ports through this tunnel?

tacit junco
#

Is it possible to run the server on multi cores? I have ond core going towards 100% After a while.

late coral
#

The main thread will run on one logical processor only

tacit junco
#

Ok. Then I guess there is not much to do. Something in making the game use a lot of resources after a while. It starts quite low but after some hours it maxing out for me. I will try windows server.

quick mist
#

That won't make any (or much) difference; the Linux and Windows servers are largely built from the same code

mint elm
#

hello guys i need some help, i play on dedicated server and after maj i can't join
where i can update my dedicated server ?

quick mist
#

please define 'after maj'.
See the Wiki for directions on updating a Dedicated Server instance.

late coral
quick mist
#

that cronjob actually isn't needed - the server has a built in restart after 24h of uptime. It doesn't hurt though (:

late coral
#

Tested that, and it only reloads the save, not restart the actual server.

#

Also, the cronjob makes steamcmd check for an update.

quick mist
#

well, yes. But it does as a result spin down and spin back up the game session which does clear a lot of cobwebs. Like I said- restarting the whole process tree certainly doesn't hurt

#

(nor does checking for software updates (: )

mint elm
#

maj = update sry im french la baguette

quick mist
mint elm
#

Steam>app_update 186638 ERROR! Failed to install app '186638' (No subscription)

#

my satis is 186638 but my dedicated is 177909

quick mist
#

where are you getting the directions to include 186638 in the update command?

#

please follow the directions as written

mint elm
#

to my steamcmd

quick mist
#

The app ID for the Dedicated Server is 1690800

tacit junco
#

Not sure if this is normal for everyone or if its something unusual.

unknown.png
quick mist
#

sustained that would be unusual but it's not all that strange for CPU load to occasionally burst to more than one core's worth.

(bear in mind that the CPU% shown by top is 'percentage of one core being utilized)

tacit junco
#

yes i know but its constantly at that level, never dip sbelow

quick mist
#

what CPU are you running on out of curiosity?

tacit junco
#

amd 1950x at 4 ghz

quick mist
#

hmm. that is a half-decade old but should be up to snuff

fading quarry
#

@tacit junco I would recommend htop its a bit easier to get some useful information from it.

topaz tendon
#

so

#

there was an server update today ?

topaz tendon
#

i cant connect to my server anymore
i restarted it 3 times now
but ingame in the server manager it is shown as offline

#

any ideas ?

quick mist
#

well let's start from the beginning. In order:

  • Is the server running?
  • is it listening on the correct ports?
  • if it's on a LAN, has its internal LAN IP address changed? Check that the forwarding is still correct
crystal eagle
#

If you dont have a fixed ip and your router reboots your internal ip will change

shadow verge
#

would a digitalocean droplet or azure vps be able to handle a server

#

I still have student credit on both and its getting close to expiring (if it hasnt yet) so I want to know

fervent carbon
#

Droplet doesn’t have that much ram iirc?

shadow verge
#

you can select the amount

red belfry
#

your credit won't cover an Azure VM for the month. used to run a 1 proc 4GB DC in Azure and it was about $100/month. you'll have to go with a larger size that will likely go beyond your $150/month credit

shadow verge
#

jesus

#

ok then

#

probably cheaper for me to gget a server for $10/mo on a rented gameserver company

#

maybe linode could work? idk I think linode is expensive too

opaque spruce
#

why is there a server restart on my dedicated server?

quick mist
#

I have been happy with Linode. they're spendier than some but you get what you pay for

#

Tdawg, could you be more specific?

opaque spruce
quick mist
#

The server automagically stops and restarts the game session every 24 hours of uptime

opaque spruce
#

Hmm.. It didnt have 24 hours of uptime. It was only at 16hours

shadow verge
#

also I believe they give a little bit of free credit just like digitalocean

#

linode has docs for gameservers so I hope that means they have some servers optimal for it

#

apparently linode's dedicated cpu servers can actually handle game servers

fervent carbon
unique bay
#

is it normal that the conveyer belts are laggy on dedicated servers?

unique bay
#

the items stuck on the belt and move after some seconds :/ and there is no reason why.

runic siren
#

I rent a server (zap hosting) but experience strange things. Like boulders on nodea that I already removed. They are not physically there but shown. Also i see snails and when i get to them, they disappear. On belts there are parts of the belts where the items disappear. Are these issues normal?if yes I'd rather continue play offline and not having the production running while my pc is not running.

fervent carbon
#

I think that’s just an issue with dedicated servers

fading quarry
#

Imagine the client sees the world pristine and untouched until multiplayer host/ dedicated server tells the client otherwise.

gentle heron
#

Hello, I was wondering if there was an officially sponsored dedicated server host(or one you would suggest) for dedicated servers? Preferably one that I can upload my save to? Thank you in advance!

fervent carbon
#

Check the pins

gentle heron
#

I did see that we can't suggest any, I was asking for an official 'tell'.

proven nexus
#

does any1 here by chance have experiance with oracle always free VMs?

#

i need help with an issue im having setting up a satisfactory server on one

#

ive installed the server (im pretty sure i did but there is a chance i did something wrong) any help would be very much appreciated

unknown.png
fervent carbon
proven nexus
#

its nothing but a console no gui or anything everything is done through either sftp or putty

#

at least for me anyway

fervent carbon
#

is it the ARM based or the AMD based ones?

proven nexus
#

amd ampere something something

fervent carbon
#

I’ve messed with both in terms of Minecraft servers

#

I don’t think the dedicated server supports the ARM cpu architecture

#

I’ll go take a look

proven nexus
#

i have 2 cores and 12GB of ram dedicated

#

but i can add more if i need to

#

just starting wtih this to see how it runs

fervent carbon
#

There are currently no plans for an ARM CPU-compatible build

#

the AMD ones will not work with Satisfactory, the memory is way too low

proven nexus
#

12GB of ram wont work?

#

i can go up to 24

#

if i need to

#

honestly i can figure out the specs and stuff later i just want to know what im doing wrong because i know im doing something wrong just unsure what

#

its been bugging me for over a month now

fervent carbon
#

Do you know anything about CPU architectures?

proven nexus
#

bare minimum

quick mist
#

The directions on the Wiki should be accurate for a Linux VM instance on Oracle Cloud, provided it has sufficient horesepower

fervent carbon
#

The one with 4-cores and 24gb of memory is using a cpu arch called ARM

#

cpu arch is essentially the type of instructions you can give to a cpu

quick mist
#

your screenshot shows you attempting to run sfserver from what I infer to be the directory you installed SteamCMD to

fervent carbon
#

satisfactory only runs on x86-64

fervent carbon
quick mist
#

that would also be an issue, yes

proven nexus
quick mist
#

and is that the directory you're telling SteamCMD to install the SF dedicated server to?

fervent carbon
#

Unfortunately I don’t think you could reasonably create a dedicated server using oracle cloud always free tier

proven nexus
#

nope

#

im not on the arm proccessor

fervent carbon
#

Are you using the AMD one?

proven nexus
#

yes

#

from what i know

quick mist
#

if you run arch, what is the output?

proven nexus
#

aarch64

fervent carbon
#

yeah that’s ARM

proven nexus
#

ah ok

quick mist
#

^

proven nexus
#

good to know

fervent carbon
#

the always free AMD ones are way too weak to power a satisfactory server

#

<1gb of memory and 1 thread

quick mist
#

x86_64 is the only valid response to arch on a host that can run the SF server

proven nexus
#

uuh

#

the always free have 4 cores and 24GB of ram

fervent carbon
#

that is the ARM Ampere instances

proven nexus
#

ah

#

im completely new to this whole style

#

thanks for the info

fervent carbon
#

anytime ^^

quick mist
#

ARM does make beefy 64-bit CPUs but they're not compatible with the x86-64 instruction set; binaries have to be compiled specifically for them

proven nexus
#

hmm

quick mist
#

think of ARM64 as diesel to the x86-64's petrol, but don't take that analogy too far

proven nexus
#

got it

#

is there anyway to get a free satisfactory server up aside from local hosting? cause im beyond broke lmao

fervent carbon
#

Not a permanent solution

quick mist
#

most free hosting tiers won't have the horsepower

fervent carbon
#

You could spin up something like google cloud witj their $300 free credits

#

but that would only last you a few months

quick mist
#

you might be able to get it running on a free tier Google Compute instance for about 2 mintues before you burn enough resources to no longer be free

proven nexus
#

lovely

#

is there a version on here that can run it ?

#

i have a buddy who has access to paid tier stuff who is looking into doing the same thing im doing

quick mist
#

maybe your buddy can offer you access to one of his hosts

proven nexus
#

ight

fervent carbon
#

Speaking of which.. I have an i7-4790k i could host mine on instead of my dual e5-2680

proven nexus
#

not very good for hosting and playing at the same time

opaque spruce
#

Is there a way to fix the visual bugs with the conveyors? Most items glitch out and disappear, just to reappear in a few seconds.

nova drift
#

SOMEONE KNOW IF FIX THE ISSUE WHERE YOU CANT IVITED FRIEND TO YOUR PRIVED SERVER

crystal eagle
#

you cannot invite friend to your privet server you just share the server ip and password (if any)

#

Please do not write in caps it is considered rude

floral meadow
#

Hello everyone, I setup a new Satis Dedi server yesterday, and loaded a 2 day old save game, works like a charm - but we want to have mods on it, is it possible by any chance ?

tacit junco
#

Not yet.

floral meadow
quick mist
shadow verge
#

how much ram do I need

crystal eagle
shadow verge
#

yes..

crystal eagle
#

minimum 16gb

shadow verge
#

damn

#

even for like just 4 people?

fervent carbon
#

8gb is recommended

nova drift
karmic merlin
#

I've unfortunately died under the map due to hyper tube cannons. I edited the file via the calculator and uploaded it to the server. But i can not start from it

#

i loaded it but nothing happens

#

is this normal or did i do something wrong?

#

do i have to remove the _calculator?

crystal eagle
#

when drag the save to the server make sure you delete the other saves

coarse crag
#

probably useful to run a few servers in parallel if it can handle one.

fervent carbon
#

doesn’t run at 30 tps, but 26-30 range

late coral
#

@fervent carbon 4790K certainly copes well with the Dedi server.

fervent carbon
#

would have to check with the owner of it if it’s okay

coarse crag
#

my home pc is still a 4790K 🙈

blazing halo
#

You'd have to anyway if you don't want to do shenanigans with beacon ports, right?

blazing halo
#

I have a different question: sometimes (when exit is requested, when it crashes...), the server process doesn't terminate completely. Instead it just sits at LogOnline: Error: OSS: Leaderboards Interface Requested. Can I somehow make it exit all the way, or is this a known pending issue?

crystal eagle
blazing halo
late coral
blazing halo
frosty gale
#

probably a stupid question but do you need to port forward to use the dedicated server?

blazing halo
#

If you're behind NAT - yes.

#

In general, in every situation where the computer you want to reach does not have a public IP - you will need to port-forward.

#

If the machine has its own, internet-reachable IP address (as would likely be the case in a correct IPv6 setup, as discussed above), you will not need to port-forward.

frosty gale
#

damn it
okay, thank you

#

follow-up, what ports do i need to open?

fervent carbon
#

i think you only need to open 15777 with udp, the rest will dynamically open iirc?

frosty gale
#

alright, thanks
optimum being stupid anyway lol so i guess i cant port forward yet anyway

#

"Sorry, we're having trouble accessing your router settings."

crystal eagle
fervent carbon
#

If you can’t get the portforwards to work, there’s always the option of a vpn. Although would require your friends to install it to play

blazing halo
#

Speifically using:

while true; do logger -p user.notice "FICSIT_WATCHDOG: Updting and starting the Satisfactory server"; ~/satisfactory_dedicated_server_update.sh; sleep 10; /var/NVMe/SATISFACTORY_Dedicated_Server/FactoryServer.sh -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000 2>&1; logger -p user.notice "FICSIT_WATCHDOG: Satisfactory server exited with code $?"; done
blazing halo
fervent carbon
blazing halo
#

By default, yes.

blazing halo
#

(the other are - having full control over the device in my house (as in, admin access), actually owning the thing (meaning: I can swap it out, throw it out a window, whatever), being able to update it whenever I want (because both ISPs and consumer-facing terminator manufacturers seem to have some trouble keeping their crap up-to-date, for some reason), and nipping all sorts of "free" WiFi APs they may deem to want me to provide - at my own expense (electrical bill) - in the bud)

crystal eagle
crystal eagle
#

In my Start up of the server basically it will restart the server in case of a crash

blazing halo
#

Which is essentially the equivalent of what I do.

#

Except, when the server crashes, it doesn't exit for me.

#

It just... hangs there.

coarse crag
blazing halo
#

Uhhh, what?

#

You want to try that one again? :D

coarse crag
#

you probably want something like:

while ! logger -p user.notice "FICSIT_WATCHDOG: Updting and starting the Satisfactory server"; ~/satisfactory_dedicated_server_update.sh; sleep 10; /var/NVMe/SATISFACTORY_Dedicated_Server/FactoryServer.sh -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000 2>&1; logger -p user.notice "FICSIT_WATCHDOG: Satisfactory server exited with code $?"; do echo "Server crashed. Restarting."; sleep 1; done
#

Well, not really, I hope you get the idea. :P

blazing halo
#

That literally makes no difference.

coarse crag
#

let me try to make it proper

blazing halo
#

Ahh, I think I know what you're trying to do.

#

You're trying to make it cycle unless it crashes, right?

#

But again, it doesn't matter, because... the process doesn't exit. :D

#

It never gets to "server exited".

#

The server process just zombifies.

#

Well, actually not exactly zombifies, because it reacts to Ctrl+C.

#

But why it doesn't exit before then? No clue.

coarse crag
# 
clean_quit=false
while ! ${clean_quit}; do
 logger -p user.notice "FICSIT_WATCHDOG: Updting and starting the Satisfactory server";
 ~/satisfactory_dedicated_server_update.sh;
 sleep 10;
 /var/NVMe/SATISFACTORY_Dedicated_Server/FactoryServer.sh -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000 2>&1 
 exit_status="${?}"
 logger -p user.notice "FICSIT_WATCHDOG: Satisfactory server exited with code $?";
 if [ ${exit_status} -ne 0 ]; then
  logger -p user.notice "FICSIT_WATCHDOG: Server crashed, restarting it.
  sleep 2;
 else
  clean_quit=true
 fi
done

So yeah, until the server stops quitting with error status, it will loop restarting.
Closing the server gracefully should make it quit with exit status 0, unless ficsit has their own set of exit rules. :P

blazing halo
#

Cool.

#

Except it does not terminate.

#

/var/NVMe/SATISFACTORY_Dedicated_Server/FactoryServer.sh -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000 2>&1

coarse crag
#

ok, so if you're saying it never quits the script above is trash

blazing halo
#

this line sticks.

#

It finishes shutdown, and then just sticks around.

coarse crag
#

when you ps aux | grep the process what do you get?

blazing halo
#

Gimme a sec.

coarse crag
#

this .sh probably spawns a child process, so you want the .sh -and- what it calls

#

(ps auxww if the line is long might help)

blazing halo
#

(it only calls the factory server, and doesn't do anything beyond that, so if anything is spamming children,

#

it's the server itself)

coarse crag
#

Yes. So (a) the child process (the actual binary) might be stuck (b) the shelll script might be the culprit

blazing halo
#

No, the shell script comes with Satisfactory DS and literally just sets a few env vars and launches the executable with the stuff you provide.

coarse crag
#

and there are some flags from ps aux that tells what status the process is in

blazing halo
# 
#!/bin/sh
UE4_TRUE_SCRIPT_NAME=$(echo \"$0\" | xargs readlink -f)
UE4_PROJECT_ROOT=$(dirname "$UE4_TRUE_SCRIPT_NAME")
chmod +x "$UE4_PROJECT_ROOT/Engine/Binaries/Linux/UE4Server-Linux-Shipping"
"$UE4_PROJECT_ROOT/Engine/Binaries/Linux/UE4Server-Linux-Shipping" FactoryGame "$@"
#

That's the entirety of it.

coarse crag
#

like, actual zombie, stuck in i/o, or just idling.

blazing halo
# 
steam     117916  0.0  0.0   2420   524 pts/1    S+   00:30   0:00 /bin/sh /var/NVMe/SATISFACTORY_Dedicated_Server/FactoryServer.sh -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000
steam     117923  119 18.5 10672524 6075332 pts/1 Sl+ 00:30   2:47 /var/NVMe/SATISFACTORY_Dedicated_Server/Engine/Binaries/Linux/UE4Server-Linux-Shipping FactoryGame -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000
coarse crag
#

so yah, what's the state of that UE4Server-Linux-Shipping when it becomes stuck? It goes away if you just kill it (without -9 or SIGTERM)?

#

hmm it looks pretty busy in CPU

blazing halo
#

Ctrl+C kills it, no issue. It just doesn't self-terminate.

coarse crag
#

in this ps you caught, it is in that "no-self-terminate" state, right? You issued the console command to quit, and it stuck there, showing that in ps?

blazing halo
#

Yeah, I said "exit" through the console.

#

The final lines in the output are:

[2022.05.14-22.32.39:352][373]LogDeviceProfile: Device profile begin destroy: [0x7f75575d8440] Linux
[2022.05.14-22.32.39:352][373]LogDeviceProfile: Device profile begin destroy: [0x7f75575d8380] Default__DeviceProfile
[2022.05.14-22.32.39:352][373]LogDeviceProfile: Device profile begin destroy: [0x7f75575d82c0] LinuxServer
[2022.05.14-22.32.39:354][373]LogOnline: Error: OSS: Leaderboards Interface Requested
[2022.05.14-22.32.39:354][373]LogOnline: Error: OSS: Leaderboards Interface Requested
#

and it's just sitting there.

coarse crag
#

it is just running (as far as the OS is concerned). No zombie or anything. It is in a deep loop as it seems to be taking a good load in the CPU

blazing halo
#

Which is weird, because htop says 0.0

coarse crag
#

try another ps , it should report the same as htop

#

maybe when you did it, the game was still in the process of shutting down?

blazing halo
# 
steam     117923 44.7 18.5 10676636 6075456 pts/1 Sl+ 00:30   2:49 /var/NVMe/SATISFACTORY_Dedicated_Server/Engine/Binaries/Linux/UE4Server-Linux-Shipping FactoryGame -multihome=10.0.1.22 -ServerQueryPort=20000 -Port=20001 -BeaconPort=15000
coarse crag
#

hah I think htop is a bit off... well, or ps, but I'm more acquainted with ps anyway

blazing halo
#

top says the same, basically

#

(0)

coarse crag
#

it even increased the process cpu time (2:47 -> 2:49), I think htop is not getting the right process

blazing halo
#

ps aux | grep -i satis is what I use

coarse crag
#

the line is getting the right process... that UE4Server-Linux-Shipping.

#

Okay, so if you go ahead in the terminal and ctrl+c it, then the process should be gone from ps, and that would ensure you're getting the right one

#

the time of the logs is consistent with the process start time (GMT-2h should be your local time) -and- its CPU time (it's not the same as wall clock but assuming the game consumes a lot of cpu cycles it could be somewhat close to wall clock)

blazing halo
#

The weird thing is, if I run these commands one after the other, the supposed "CPU" line just keeps dropping down at a rate of 0.1 every 2s or so.

#

It's weird as hell.

coarse crag
#

It is probably internally waiting for a thread to return before it can quit... it may mean waiting user connection to be dropped from the server

blazing halo
#

Except there's no user connection.

coarse crag
#

I mean, you're connected to console in-game and it waits for you to disconnect from the game. You issue the command and leave at once?

blazing halo
#

No, it literally says "This server appears to be offline" in-game.

#

Only thing I can get from ss is UNCONN 0 0 10.0.1.22:20000 0.0.0.0:* users:(("UE4Server-Linux",pid=117923,fd=32))

coarse crag
#

It really looks like it is waiting its threads to finish and some are still stuck around

blazing halo
#

Yeah, it kinda looks like that.

#

And I'm starting to think it's retrying that "Leaderboards Interface" thing to infinity.

coarse crag
#

This really looks like a bug, what you could do perhaps to help is identify which thread is stuck... now I'm not sure how you go down the individual threads if they don't spawn new processes in ps

#

but that + in process stats means the process is multi threaded. let me see...

#

try to add -T, probably ps Taux | grep 117923 should do. I'm not sure if the threads are gonna read after satis, let me see

#

just to be clear this would be just diagnosing the issue, not fixing it. You could of course work-around knowing which thread is sticking up. But basically I think this should turn into a bug report. :)

#

I don't have any processes here running multi-threaded stuff 😂
But the flag is ps auxww -T | egrep <...>, for some reason -T didn't work with the aux sugar

blazing halo
#

Okay... what am I looking for?

#

Spits out like 42 lines.

coarse crag
#

multi lines with the UE4 process... do they show anything extra in the command line, or all the same?

blazing halo
#

Same, some Sl+ and some SNl+

coarse crag
#

they -might- give hints to what the thread does, so you can individually kill the thread and find out which one was the stuck one

blazing halo
#

Still, I guess I don't really care which one is the stuck one, since I won't be able to tell what it was doing - and even if I did, I don't see how it helps me.

#

So, basically, a bug.

coarse crag
#

ah, then there's not much clue, unfortunately.

coarse crag
blazing halo
#

Don't have an account on the whole bug reporting system. Kinda pissed me off that I needed one, so I never made one.

coarse crag
#

you know, you can just run the process with strace and see lotsa fun output that would make no sense... haha especially a multi-threaded one

blazing halo
#

Why wouldn't it make sense?

coarse crag
#

now the funny thing is that not everyone is reporting this

blazing halo
#

It just prints syscalls, right?

coarse crag
blazing halo
#

It would be. There's a ton of calls to the OS it'd be making.

coarse crag
#

yes, it'd take a lot to make any sense of that, especially cause it's running in separate threads

blazing halo
#

I guess having the thread ID in there might help. :D

coarse crag
#

LogOnline: Error: OSS: Leaderboards Interface Requested
What about this error?

blazing halo
#

Let's see what it does. I completely forgot about strace.

coarse crag
#

Wouldn't it indicate it could be stuck waiting this leaderboards interface response?

blazing halo
#

Like I said, I suspect it's doing exactly that.

#

Kinda silly not to include code to just fail it after n retries, but... beta stage, right?

coarse crag
#

ya, and I believe, there are LOTS of code there to surround every little check, so until we fall in one failing like that, I believe lots of sanity already took place. Just an "edgy" case I guess. The thing is that it is always happening to you

#

Something I read in wiki is that the dedi server dies "gracefully" if you send it a SIGTERM signal (15), and says nothing about ctrl+c's signal, SIGINT (2).
So for one thing, ctrl+c works for you on linux at least, because it does do a "non-graceful" shutdown.

blazing halo
#

Actually, it does a graceful,

#

if you issue it again it kills it outright.

#

Unless it's already shut down, in which case it just terminates.

#

That bit I really like, and it shows foresight.

#

The fact that they added a signal to trigger a graceful is another thing that's really nice.

coarse crag
#

ya, I was talking just based on what I read in wiki. I gotta give this dedi server a try here

#

perhaps the wiki could use some updates :P

blazing halo
#

Yup, the last thing it gets stuck on is futex(0x7fbc5d9289d0, FUTEX_WAIT, 125546, NULL, so it seems it really is waiting on something. Possibly a deadlock.

#

Wouldn't be surprised if it's something silly, like it retrying on failure, and it failing due to shutdown.

#

(that's an inescapable situation: it will retry indefinitely, and it will fail indefinitely)

coarse crag
# 
Connecting anonymously to Steam Public...Retrying. . .

Oh, perhaps my server is VAC-banned from steam :S

coarse crag
#

most likely waiting that leaderboards interface it so eagerly requested. :P

#

wee my VAC ban is gone, connected and downloading...

blazing halo
#

> my VAC ban is gone

That doesn't exactly inspire confidence in your assistance, you know... ;]

coarse crag
#

Ya. I'm not even sure what I'm talking about with that vac ban thing... I think it was a thing back in the time I played some game on steam that was online. The cross-play aspect of satisfactory using epic games probably has some kind of VAC but should work differently from steam's

blazing halo
#

... the different nature of the ban still doesn't really help, here... :D

#

(unless you got blocked, not banned - a "block" is when you get put in a "time-out" for using too many resources, too many queries etc.; a "ban" is when some algo or a human blocks your access for specific offences, like cheating in online games, and generally doing stuff on purpose to the detriment of other players)

coarse crag
#

ok, I reproduced your error.
If I ^c once, it gets stuck the same as you reported

unknown.png
#

And this may have to do with the error/warnings issued along the initialization, about being unable to log in to steam.

# 
[2022.05.14-23.48.25:410][  0]LogSteamShared: Warning: Steam Dedicated Server API failed to initialize.
[2022.05.14-23.48.25:437][  0]LogOnline: Warning: STEAM: Failed to initialize Steam, this could be due to a Steam server and client running on the same machine. Try running with -NOSTEAM on the cmdline to disable.
[2022.05.14-23.48.26:243][ 20]LogOnline: Warning: OSS: EOSSDK-LogEOSEcom: Purchase flow is disabled due to overlay setup failure (EOS_NotConfigured).

These really look related to the leadeboards stuck one.

Specifying -nosteam (or -NOSTEAM as suggested by one of the warnings) did not change anything.

shadow verge
#

is there like server plugins yet

coarse crag
#

you mean, things to automate the daemon? or more like in-game mods?

#

I'm not sure by reading the wiki page, whether for dedicated servers -beta public and -beta experimental branches are merged and the same, or if for the dedicated server "bleeding edge", I should still stick with -beta experimental...

#

I found an issue already logged about this issue. @blazing halo it would probably help get this fixed if you logged in and upvoted the issue. Or else this could just become a feature, that is, if nobody cares about it. :P (well, and in case this doesn't bother one of the developers either)
https://questions.satisfactorygame.com/post/62516d2cca608e080350be94

#

the search in the site could be a bit better, it was a bit difficult to find the right topic.

shadow verge
#

or like items automatically disappearing after a certain amount of time

coarse crag
#

hmmm they should be "appliable" as much as game mods are, but I think not even game mods are really supported (but pushed by some craftier people), right?
I'm sorry, I'm not much acquainted to how modding is being used in the game

#

but if you look up some listings in #looking-for-group-old, you'll see some people "modded". Let's see if there are dedi ones that checked "modded" in

#

actually no, no dedicated server I see says "modded" at least in the few last ones I seen

vague geode
#

Sorry if this question has already been addressed many times, but does anybody else have an issue with your character not being able to move for a bit after logging onto your server? It happens to some people on my server, but no everybody. Also, the character does actually move, but on screen all you can do is look around then after 30secs or so(it's getting longer and longer) all of a sudden everything is fine and you show up wherever you moved to while you couldn't see that you were moving.

quick mist
vague geode
round glade
#

i blew up a rock above a caterium node earlier, and i came back and the rock returned?

#

i can move through the rock, but i cant build on it

#

is there a way to fix this?

blazing halo
# shadow verge I guess server side mods, stuff like a plugin that makes it so all lizard doggos...

At present, mods don't work with the dedicated server. As I understand it (as I did a little digging into it), it will attempt to load them just fine (meaning, the support is there), but since there originally was little modding support - the community had to come up with methods for stuff that wasn't available (like hooking new recipes into buildings). This ended up becoming a mod itself - one that essentially created and supplied all of these mechanisms - and most (if not all) mods currently available rely heavily on that particular one. And as far as I'm aware, there's currently no tooling to make it work under a dedicated server - meaning, all mods that depend on it (pretty much all of them) will also fail to work.

As for Linux availability (if that's your thing), it also doesn't help that all mods are currently distributed as Windows binary code (notably, the one I talked about is). No Linux binary in sight. So... that might take a bit longer still.

shadow verge
#

damn

#

well I use a host which I believe uses linux (as it should)

#

especially with low ram

blazing halo
#

Technically, if you could put together a mod that doesn't rely on the mod loader, and release a Linux binary for it (if it needs one) - it potentially should work.

#

That limits it to mods that modify already existing stuff, I suppose. But if you're gunning for day/night cycle changes and such, it's theoretically doable.

vague geode
# round glade i blew up a rock above a caterium node earlier, and i came back and the rock ret...

Same thing has happened to me on a caterium node up in the Northern Forest today. Well, it was a little different. I threw like 10 bombs on the rock and it didn't blow up. I logged out and back in and the rock was gone and I was able to put a miner on the node. Next time I logged in the rock was back, but the miner was still there working, so I guess my suggestion would be to log out and back in a couple times and if the rock ever disappears again, go ahead and put a miner there.

blazing halo
# round glade is there a way to fix this?

Per instance, yes. Go to a location, stand there, save game, load saved game. On load, the server will detect that there's a player there, and will remove obstacles. They will return on subsequent loads/restarts, so do your business there and leave. ;]

#

Hm, maybe saving/loading isn't necessary... But the point is, there has to be a player there, for the server to remove that stuff. It's a known bug.

native temple
#

I don't think there's a way to really unlock the node

#

it happened to me too but it was an iron node so I just left it

blazing halo
#

Well, the boulders will be gone when you return (after leaving a player there), so it works temporarily at least. Long enough to place whatever you want to place and leave.

#

There's a caterium node down south I had this happen to, but the boulders problem is general.

vague geode
shadow verge
#

I assume coffeestain will make a library in the future

#

or a hook

#

since its still experimental

blazing halo
#

I don't see a technical reason why they couldn't make an internal interface that a) exposes the exact same stuff the mod loader does, and b) allows its internal functions to be "overloaded" so that if you happen to load the mod loader (and it's a new version, providing more stuff - the community tends to be quicker with these things, and with their ear "more to the ground" as it were) then the internal functions end up "slaved" or outright blocked.

#

As for the economy of doing that... who knows.

#

There's also no technical reason for the community not to release a new version of the mod loader, with the appropriate functions removed, moved to some "no conflict" variant, or only exposed if the engine doesn't expose them for some reason (older/legacy version?). But again, I don't know how economical that would be.

#

One thing's for certain: rebuilding all of the existing mods is not economical in the slightest, and would likely turn a number of authors off modding Satisfactory. I mean, why bother if you're expected to rebuild your stuff every few weeks? Waaaaay too much work, when a lot of work goes into making a mod in the first place.

frosty gale
#

just a quick question, if i start a server on my linux machine can i just copy over an already existing singleplayer world to keep playing on the server

fervent carbon
#

Yes

#

The wiki has this documented

south sluice
#

is there away to make that when a player is on server auto save be after 10mins?

frosty gale
vague geode
mighty salmon
#

I'm having an issue where my dedicated server doesn't seem to want to run longer than 24h

shadow verge
#

like a hook or library for server addons

blazing halo
#

Same here.

#

The mods are the same for both server & client. They have to be, they must be loaded by both. There's really no distinction, apart from what needs to be enforced at the server vs what needs to be present at the client.

quick mist
coarse crag
finite kindle
#

Is there anyway to add mods to a dedicated server?

late coral
finite kindle
#

oh dam, feels bad

rich geode
#

So I just set up a server on a spare computer I had, got everything up and running, set up port forwarding on my router to go the right place and also a DDNS address so my friends could join, but for some reason when I try to connect to that DDNS address from within my LAN, it can't find it. It works fine if I put in my local internal IP, though.

#

I'm using steamcmd and nssm to run the server, btw

mighty salmon
#

man, vehicles are still not good multiplayer I see.

coarse crag
#

that's because you are inside the lan, you don't need to pass thru the router to get to the destination. Port is forwarded only from connections coming from outside.

#

There probably are means to make this work, but do you really need it?

rich geode
#

Ah ok, yeah I was thinking that might have been it. I was just trying to see if the DDNS I set up would work for my buddy so he could connect too