dedicated-servers

[opaque pine]

is that not what im doing

[slim spoke]

i dont know what ur doing

[opaque pine]

so, i made a save with the server so the server folder would appear in savedgames. i deleted the saves and imported a save from singleplayer into the server. and im trying to figure out how to load that save.

[slim spoke]

1. Find the desired save file and Session name in the Load tab in the main menu.

Session Name is of critical importance, save files only display it if they haven't been renamed while being saved, the only way to see the Session Name otherwise is from the Load Menu or a save editor

[opaque pine]

so i have to rename the session with satis calc?

[slim spoke]

4. Insert the Session Name from step 1 into the server settings tab in Server Manager, press ↵ Enter. The server should initiate a search for the save file with the matching Session Name.

thats what i know

[opaque pine]

i tried that. it wont let me change the session name at all

[slim spoke]

Loading a save file

[opaque pine]

thats what im trying to say. in server settings, i try changing the session name and it resets itself everytime i type something

[slim spoke]

u decide u wanted the cmd to change the session name

[opaque pine]

what

[slim spoke]

load in game and try changing it

or use scim or something to change it

[elfin idol]

Hi guys! So I have decided to move my server to another computer, how do you transfer the savefile from one server to another? I have looked around but I can't even find the save.

I read a little of bit of the discussion above, but I became more and more confused.

[slim spoke]

just copy the save file and put on the server

run server and start a save. then goto server manager and load ur session for the save u want

wiki has the details/cmds

[modern raft]

has anyone had this problem some days a go our server (Dedicated) did a automated restart and after that our 7 lizard doggos has since been walking away from us in there pen like they do in the wild
we cant retame them

[wild herald]

Hi I have an issue with floor wholes, sometimes after placing them pipes an conveyor around them get invisible. Is there an workaround ?

[snow belfry]

so anyone with more experience in linux than i have (i am just a beginner) i have the server running with systemd as a service.
I remember back in the day i used screen to open up running sessions, but can i do that with a service ? I would have to install screen and change the service so it runs with screen right ? That is my current assumption.

[long vale]

Hi all. After the update at the 4. nov I get following error. I tried to remove it and add it again. Anyone else has the same problem? Or a possible solution? Thanks

[snow belfry]

did you add the -multihome ?

[slim spoke]

if u search systemd u might find the pics of guys setup

dont need multihome now

[snow belfry]

oh 😄 okay

[slim spoke]

i said same today but bogdan told me diff

[long vale]

Dont think so, my friend build it.
But it seems the problem wasnt on Satisfactory side, it was just a DNS problem...

[snow belfry]

well i do have the systemd service running, i basically want to get the output from the service on the terminal

[slim spoke]

im not sure, so i wont guess and mess things up

[snow belfry]

im 80% sure i need to setup the service with screen^^ but i wanted to make sure

[slim spoke]

but if u search there is someone running that setup u could ask them

look for images in here

[snow belfry]

yeah looks like i need to change the service to run via screen, i wish linux would just frikkin make screen default 😄

[elfin idol]

Thank you for the help! Turns out I knew exactly what to do. I just looked in the wrong place for the savefile.

Though I ran into another problem with that the NAT detection failed. But I am troubleshooting it at the moment.

[slim spoke]

kool good luck

[cinder wind]

I've had problems retaming them after dying in single player, and had the same problem yesterday. Not sure what it is, I've pretty much given up on taming them all together.

[hot ivy]

Hi there,

I've got a dedicated server installed via steamcmd on Windows on a headless system. Is there an easy way to restart the server?

I've tried using services (both nssm and powershell) but neither of those seem to work with arguments like -log which is needed. Anyone happen to know a solution?

[spiral root]

How do you start your server?

[hot ivy]

Currently I just run it from terminal with the arguments

[slim spoke]

server manager console tell it to quit and then start it how u did b4

[hot ivy]

That's what I'm doing already which isn't ideal. Guess I need to figure out why the service isn't working properly, unless that's just how it is for now.

🤔

[slim spoke]

have to fix ur setup to work

[ornate pawn]

@echo off
title rusttimerestart
set restart_time=11:00
:start
start /WAIT C:\steamcmd\steamcmd +login anonymous +force_install_dir SatisfactoryDedicatedServer +app_update 1690800 +quit
start FactoryServer.exe -log -unattended

:loading
set current_time=%time:~00,5%
ping 127.0.0.1 -n 2 > NUL

if %current_time% == %restart_time% (
echo Please wait...
goto loading
)

:loop
set current_time=%time:~00,5%
cls
echo %TIME%
echo %current_time%
echo Next Restart at "%restart_time%".
ping 127.0.0.1 -n 2 > NUL
if %current_time% == %restart_time% (
taskkill /f /im UE4Server-Win64-Shipping.exe
cls
goto start
)

goto loop

[spiral root]

Why?

[lofty latch]

pastebin.. >_<

[ornate pawn]

Try renaming the Saves folder in Appdata, creating a new session with the same name and coping the autosaves back once its done. I had to relclaim mine.

[ornate pawn]

Auto update. Got a better way?

[foggy agate]

why do you ping localhost?

[ornate pawn]

Just as a pause otherwise the batch terminates

[foggy agate]

ahh yes, that makes sense 😄

[trail lichen]

Erm... what? - Is that real? XD

[lofty latch]

Yup:

only actually restarts if there's actually an update to install 😛

[ornate pawn]

What is that? Where do I get it?

[hot ivy]

Is that AMP?

[lofty latch]

correct

[hot ivy]

Aww poop

[lofty latch]

lol

[hot ivy]

Cubecoders >_>

[lofty latch]

I shouldn't really post the link myself - I use example(dot)org/page here

if someone would kindly furnish Mark with the link

[ornate pawn]

Found it. Thanks

[lofty latch]

👍

[hot ivy]

That seems pretty neat though, going to give it a go, thanks

[vocal jungle]

lol

[grand geode]

So after I shutdown a dedicated server even after 24 hours later it looks like CSS or somebody is constantly polling the server using about 8kbit/sec of data constantly trying to get information

13:15:20.302626 PPPoE [ses 0x49b7] IP x.x.x.x.45988 > x.x.x.x.15777: UDP, length 10

The above is captured from a router which the server was behind and these come out at a rate of about 50/second........

[vocal jungle]

could be a player ingame

[grand geode]

the server has not been running for 24 hours....

and thats constant like 24/7

and it has nothing to talk too......

It also called it self FGServerBeaconClient and not any of hte game users who were playing

[2021.11.02-13.12.54:881][984]LogNetVersion: FactoryGame ++FactoryGame+main-CL-170636, NetCL: 170636, EngineNetVer: 16, GameNetVer: 0 (Checksum: 1337622931)^M
[2021.11.02-13.12.54:948][986]LogBeacon: Client netspeed is 60000^M
[2021.11.02-13.12.54:948][986]LogBeacon: Beacon Join FGServerBeaconClient EOS:(EOS)b829e959e70d44f7a9f506dd3b95fb50|000268945cbd4c4ca2a46040cd0ef149^M
[2021.11.02-13.12.55:016][988]LogBeacon: Handshake complete for FGServerBeaconClient_2147455575!^

Does anyone else see the same client connecting to their server?

[foggy agate]

might wanna check the IP where it belongs to :)

[grand geode]

I can i think its a CSS server poller..... cause its not where any of the game client that were playing came from

The problem isn't that it does it. The problem is that it doesn't "backoff".......

Like its constant 50+ udp packets / second polling....

[bold atlas]

I haven't tried the dedicated server yet. Is it possible to entirely turn off STUN?

It keeps causing trouble; I'd much rather just open ports.

[solar remnant]

I don't think STUN is used in the dedicated server?

[elfin idol]

Ok, so quick question: If I had a server running on one PC, when I move the server to another PC on the same network. Do I need to close the ports or something on the old one? I have already changed the portforwarding to the new PC

[bold atlas]

Yes, but you do that by stopping the dedicated server.

[elfin idol]

Okay, so the server automatically closes them?

[bold atlas]

No, the OS does. It's only possible to have open ports if some application is actively listening.

[elfin idol]

I see, I'll continue troubleshooting then. It's really weird now cause if I start the server on the new PC the NAT detection failed happens and it stops doing anything. But if I start the server on the old one at the same time, then the new start to do something. I have no clue what's going on 😂

[bold atlas]

That probably means you haven't forwarded ports correctly.

There might be a firewall in the way as well.

[elfin idol]

I checked the firewall and even allowed the individual ports. I haven't restarted my router though when I think about it

[bold atlas]

Check with tcpdump which machine the traffic is arriving at. 😉

sudo tcpdump -ni any port 53 will show you all DNS traffic, for example.

[foggy agate]

at least, I dosn't get these messages....

[elfin idol]

Thank you for the tip, I'll try that! I'm pretty sure that I just missed something basic

[grand geode]

Anyway raised an issue for it.... https://questions.satisfactorygame.com/post/61853629831c85205235ab58

cause its either a broken polling server or a broken client polling without backoff....

Either way its a little "anti social" as far as internet protocol go to do that

[sinful nymph]

it sounds more like an attack than anything

[grand geode]

Well yes thats how I am treating it. Just putting together a ISP abuse report for a DOS style attacking and sending it to the authority above them

[tame solar]

assume it's a bug bug before malicious intent, I'd say

[grand geode]

Well yes. But since they won't "stop" or ther eis anyone to contact about such a bug. It will be treated as abuse because even if it is a bug. Its still abuse

[quick mist]

The log capture shown above shows a complete successful handshake. is the elevated traffic denying service or degrading system performance? If you shut down the service do you still see "successful" handshakes on packet captures or indeed elevated traffic?

[grand geode]

What date is in that log?

[quick mist]

in the past, hence the "if you do this, do you see that?" questions

[grand geode]

That's the problem with it.... it doesn't back off when the service is shutdown it continues at those rates. Which is 700mbytes/day that will push people who have run a dedicated server on a home fibre line into things like fair usage policy issues with their ISP

Hence how its actually a DOS attack 😉

[quick mist]

Is it continual or only when a client has the Server Manager server picker up? Because "real-time latency/"ping" display in the server manager" is a known bug that's been reported. (I don't have the QA site link handy unfortunately)

[grand geode]

constant for previous 24 hours (without server running)

This is what I mean by "won't backoff"

[quick mist]

and (as I asked) if you either kill the service experimentally or block the offending inbound IP, does the traffic back off after it no longer has successful handshakes? is your system or service performance actually degraded?

[grand geode]

block what? I can't stop their server sending me UDP packets... they are already dropped at this end because there is nothing to repond to them

its uni-directional traffic

[quick mist]

you can drop inbound UDP traffic at the ingress and still track it. that way you can see if the rate back off when the client starts seeing "oh, it's offline"

I'm not asking these questions to try to convince you there is not a problem, I am trying to define the boundaries of the scope of the problem

[grand geode]

Yeah so? The traffic has already been sent.... so its still used the bandwidth

I would need to ask somebody upstream to block it... like my isp for example

and they don't really do that unless they absolutly have to

So I am asking offending AS number to investigate / take action

[quick mist]

Fine. That's wonderful. It would still be nice to get accurate descriptions of how the traffic behaves, which is why I am trying to obtain detailed information.

[grand geode]

Well thats simple. It looks like its a query packet for the satisfactory server. When it was running it would get a packet and reply with a packet. So it behaves like its a valid client (hence the login) and its constantly polling or something but it isn't a normal player

So when you shut the server down. You just simply get a constant stream of these query packets..... and they don't backoff even after 24 hours. Just keep sending the query at a rate of about 100 per second

[quick mist]

That's all I was asking for, thank you for that

[grand geode]

So there no way to prevent the UDP traffic from using bandwidth except to stop it as source.... or upstream somewhere on the path

[quick mist]

I hypothesize this may be a client with the server manager left open, and your server added to the roster. If this is a case, that is a known bug as I said a screenful and a half ago. You can test this hypothesis by attempting to replicate.

[grand geode]

Yes. I though that as well but ran back over the client logs from all people who connected and looked at the IP ranges and none of them are even remotly in the same IP block.

[lofty latch]

or every single client that's ever added it trying to query it more to the point

[grand geode]

I also only seen it from one source constantly

[quick mist]

If every client that has the server on the roster is still polling even while the Server Manager is not open, that is a new bug.

and should absolutely be reported

but it is best to test whether or not this is the case before assuming so

[grand geode]

yeah but its not the "clients" that had the ip added. It also does not happen from my client like that

its also a client with a strange server login hence the user "FGServerBeaconClient"

and Beacon join event. Like its some kind a poller for the dedicated servers

[quick mist]

which is what the Server Manager is, which is why my first contextual relevancy point was that part of the game client

[grand geode]

That generally is common. Thats not a problem. The problem being is that it doesn't "backoff" when it should

[quick mist]

Like I have said thrice now: I agree, that is a problem

[grand geode]

a normal client if it cannot reach the server seems to backoff and just mark the server offline and doesn't ask again

let me just double check that (have to reboot to windows brb)

Yeah it is doing that actually a client is constantly polling......

but at a much slower rate and only after I enter a server manager page

but if i load a save it still continues to poll

[ornate pilot]

is it possible to set the pasword through the server files?

[quick mist]

No

See the FAQ section in the Wiki if you need to reset it because you lost the original password

[vapid ermine]

So I had my server running all night and nothing was crafted is this normal I thought the factory will still run with no players in the server

[lofty latch]

not if the option to sleep the server is enabled

[quick mist]

By default, time stops is no players are in the game

[vapid ermine]

where can I find that

[quick mist]

Server Settings

[vapid ermine]

the ini file?

[quick mist]

in the Server Manager

where you Manage the Server properties such as Settings

[vapid ermine]

ahh i see

dam

thanks so much

[quick mist]

sure thing!

[vapid ermine]

1 down

1 to go

my friend cant connect not sure if its my isp

on to the next one

🙂

[quick mist]

if your friend cannot connect, the order of things I would check is:

[vapid ermine]

the server isnt even giving its details

to him

[quick mist]

Is the server actually running?
Is the server listening on the correct ports?
Is the server bound to the correct interface?
Is the server firewall allowing ingress on the correct ports?
Are the port forwarding settings at the NAT pointing to the correct address?
Are the port forwarding settings at the NAT forwarding the correct ports?
Are you or anyone else able to connect to the server from a client on the LAN?
Are you or anyone else able to connect to the server via the internet?

If the answer to all of those questions is 'yes', then the problem is with your friend's setup.

[vapid ermine]

ok will check these thx for the info

[quick mist]

o/

[vapid ermine]

I am getting a Nat warning

so that maybe the issue

[quick mist]

NAT warnings in the server logs are expected and generally can be ignored

[vapid ermine]

ok

[quick mist]

Note to self: add that list of questions (and some means to answer them for Linux and Windows servers) to the Wiki

[sterile nacelle]

Having a new and interesting bug with the server on U5. I can build a belt, but it stays blue. I cannot use it, delete it or connect to it. If I leave the game and rejoin, it appears to be gone. But nothing can be built in that spot.

[grand geode]

restart server.. the belt is there just not rendered. Try to walk though it....

[sterile nacelle]

ok, server retstart it is

[grand geode]

eg there are belts (and functional) between the splitters / containers

[sonic eagle]

is it a bug or the merger can't take stuff from lift?

I built the same thing on another floor of my factory few days ago and it worked

[vocal jungle]

@lofty latch why you no auto populate the config and save file locations

[lofty latch]

I'm waiting on official settings to change the save location rather than hacking it in

[rose valley]

That's a reasonable complaint. I actually wanted to implement some sort of variable polling rate but didn't get to do it in time for release and now it's still in the backlog. What I think we should do is:
a) Clients should only poll while in the server manager.
b) Polling rate should vary. Servers that do not respond should be polled once a second at most or maybe twice to make it work well in congested networks and account for possible packet loss and way less in general (perhaps 10 polls per second or something like that) for servers that are responding to queries.

I didn't treat this with the utmost priority so far cause I wasn't aware of all the implications but it sounds like this should go to the top of the priority list and get sorted out asap.

[lofty latch]

double the backoff time each time it fails until reaching a sane upper limit

[slim spoke]

need option in server manager to restart/update the server

[rose valley]

There will be one eventually and it's going to be part of the API as well but this takes the cake now as it has real serious implications.

[slim spoke]

nice

[gaunt gorge]

@slim spoke Restarting absolutely, as for updating, it's a bit harder to do, because for example right now you use SteamCMD but later they may offer other ways to do it

[slim spoke]

thats all i need to know thanx bogdan

[nimble ibex]

I am trying to move a save file from one server to a new one. Does this work? I keep gettting You need to create or load a game...

[gaunt gorge]

so without having a logged in SteamCMD (even anonymous) it needs to have access, which may not be the best.

[rose valley]

Yeah, updating specifically would be tricky but depending on your setup a restart could also be an update.

[gaunt gorge]

@nimble ibex You can, save the session name, then once you go to the new server, you can put that session name in and it should load it.

[grand geode]

ahh Bogdan. Yeah I added a 2nd comment which is probably more accurate that the first. I assumed it was a CSS polling active servers or something. But nope its clients. I think the client side is way to aggressive in its polling. Its like 50 queries / second or something crazy..... Like if there has been 20 clients on a server currently. they are all going to poll at 20 * 8kbytes/sec the moment the server manager page is selected in the client's.

Would you be able to substantially slow down the poll rate so something more reasonable at least as a short term fix?

Standard rule I have normally used when making a protocol or some such.... Double the time each time you fail (aggressive backoff with a cap) eg start at 1 second. then go to 2 -> 4 -> 8 -> 16 then cap it out at about 30 seconds or so. Or even if you fail X times. Glue the refresh / retry to a button after the first 90 second sof failure.

[slim spoke]

what if people installed steamcmd to a folder in server

[gaunt gorge]

@slim spoke That's an option, but then it requires they hard force everyone to use SteamCMD when that may not be the long term plan.

Obviously right now they do force us to use SteamCMD, but they mentioned other options later.

[rose valley]

I'll see what we can do. It's Friday evening over here. Chances are it won't be fixed until Monday. Is that a really big problem?

[vocal jungle]

you win this round

[slim spoke]

not rushing anyone .. just throwing ideas up

[grand geode]

Yeah I get that I can probably survive until Monday 🙂 I will try to ask some of the non active client side to remove it from the server manager

[gaunt gorge]

@rose valley We appreciate the team listening to feedback and interacting as much as they do, dedicated servers have been pretty good for me personally so far (aside from the MultiHome).

[rose valley]

Yeah, removing the server from the save manager on the clients should help with that.

[quick mist]

Multihome is great! It's what allowed us to stay operational during the reign of the previous update! (:

[gaunt gorge]

@quick mist It is, now that we know about it 😛

[rose valley]

Yeah, multihome was the solution. The problem.. as nasty as it was.. was inevitable unfortunately

[vocal jungle]

the 6 hour slowmode on #looking-for-group-old seams essive

[grand geode]

You shall not bind ipv4 sockets twice to the same port 😛

[stiff chasm]

This polling issue you're struggling with, is it causing a bandwidth issue?

[grand geode]

It can do because its sustained..... and multipled by number of clients

[lofty latch]

Rest assured that you're not the first developers to ship a server making that exact mistake xD

[quick mist]

Knowing it's Friday evening: have a fantastic weekend, y'all at CSS 😄

[rose valley]

Well.. it's more like thou shalt not bind an IPv6 and IPv4 address on the same port in this case, cause IPv6 also implicitly and non transparently binds to IPv4 on some systems.

[gaunt gorge]

@grand geode @stiff chasm I had a client with this issue, and we didn't have any bandwidht issues, but they managed to ddos themselves until they completely removed all servers (we had two setup) and then re-add just the one.

[stiff chasm]

20kbps per client doesn't seem like a huge number to be worries about...

[lofty latch]

also in case anyone else would care to vote: https://questions.satisfactorygame.com/post/61850c4c831c85205235aa0b

[quick mist]

that or "Thou shalt make sure thy port is not already bound before thou triest to bind unto it"

[rose valley]

But that's how you check, by trying to bind it xD

[grand geode]

It is if you had 20 players.... its 1000/packets/sec or about 14GB of traffic per day 😉

[rose valley]

Yeah it's a bigger problem for the server but also as a client if you have 20 servers in your sever manager, a pretty big issue.

[grand geode]

Yup client side issue as well

[stiff chasm]

Thats really not that much for a game server, 400kbps is not huge.

[grand geode]

This is "non-connected" clients lmao

[stiff chasm]

I totally agree the polling shouldn't happen unless the client is on that UI screen.

Or at least connected to the sever.

[gaunt gorge]

@stiff chasm the issue happens if the client is connected or not, also because the packets are small, this is lots of processing per client, per connection, per packet these all add up hugely.

[quick mist]

if it's connected it doesn't need to poll, does it? (:

[grand geode]

@rose valley Should probably back off to a manual retry / refresh button or something after a few minutes imo

[lofty latch]

I just hope we get the +connect host:port syntax and can skip the server manager entirely 😉

[grand geode]

Actually it polls them all anyway after loading a single player game if you selected server manager prior to loading the single player

[gaunt gorge]

@grand geode Or a better solution is to only poll while the game (menu) is active (focused) any set it to throttle to every 1,5,10 seconds etc

[grand geode]

So a client will use bandiwdth even when not in multiplyer.....

[stiff chasm]

Fixing this part is what should probably be prioritized.

[grand geode]

but if you have like slower client ISP like 10mbit or so. the client can probably disable its net connect and the end user also won't understand why. Just simply by selecting "server manager" at some point

[rose valley]

Why skip it? We poured our hearts into it.

[rose valley]

It's not tho. That's supposed to be just a quick live test.

[prime gorge]

Are these a thing though? It feels odd to rent anything below 100 MBit. Gigabit is the norm, at least in Europe.

[lofty latch]

half joking, it'd just be useful in some cases

[quick mist]

As a client I actually prefer ingame server managers (like that implemented in the SF client) to extrinsic stuff. I'm pretty pleased with the first-pass on the Server Manager UX

[rose valley]

It's super early anyway. A lot of content will be added to it.

[quick mist]

as an admin who wants to monitor my server, that's when I start wanting APIs (:

[grand geode]

What on the end users line connection?

[rose valley]

I think it's rude of developers to assume high end connections. We obviously can and will do better.

[stiff chasm]

I think it's pretty safe to assume a dedicated server will have a decent connection. If it doesn't you'll have a poor experience

[lofty latch]

not really given how many people just want to host at home

[prime gorge]

Are we talking past each other? I have the feel you are talking about renting a server as a service. I am talking about renting a whole server/cloud instanc/vm/whatever.

[rose valley]

But this is a client side issue first and foremost.

[grand geode]

I was hosting on the back of an FTTC connection. ~80mbit or so to the house. I suspect a lot of people like me are testing it that way for now cause I have spare hardware (though possibly also under powered)

[grand geode]

Well yeah but it happens on both sides client and server 😉

[ashen pecan]

Yup I know many friends that just use old laptops for hosting whatever gameserver of the week they're hooked on

lots of home connections out there

[stiff chasm]

100% agree, the solution is to fix the client, not the server. As I mentioned earlier, the server should be expected to handle that amount of requests.

[rose valley]

That's the plan, indeed.

[grand geode]

and the problem for server admins is they have no real control over the traffic the client is sending because its udp / stateless....

so there is no natural "backoff"

[worn spoke]

Definitely a thing. Many parts of the US still have max available speeds well under 100Mb. (Mostly rural and remote areas, but still.)

[quick mist]

I think it was already clearly stated that it's an acknowledged issue which Bog is going to escalate in priority

(for which ❤️ by the way )

[prime gorge]

You are talking about home internet connections. Or actual datacenters?

[worn spoke]

Home internet connections. As explained above, a lot of people will just want to set up a dedicated server at home.

[grand geode]

Sometimes there isn't much difference... Like I am about to upgrade my home connection to fibre so......

[stiff chasm]

Fiber != a 'data center connection'. A lot of the time datacenters are hooked in directly to a T1 fiber provider, the backbones of the internet.

[quick mist]

A datacenter with a "T1" would not get the time of day from any clients

[grand geode]

its expiremental... so people won't want to go throwing £££'s at it until its a bit more stable. Like the group I setup was able to imploded the save within a week because of the build speed and the save time got way too big too quickly....

[quick mist]

10Mb/sec up/down would get laughed out of the room for even a residential broadband connection these days

[worn spoke]

Even once stable, not everyone has the money to rent a server for $20+ a month.

[grand geode]

T1 doesn't actually exist in the EU 😉 its an america stand. Its also way more complex than that. Mostly involving peering points ATM clouds and all sorts of backhaul

[alpine dawn]

if you replaced this with multiple 10gig+ connections then it would be true

[grand geode]

Some data centre hosts btw... have to pay for their bandwidth to their parent hosts as well. Which actually often is under powered too..... it just how the internet works. So someitimes being in a data centre isn't any better either if they have cut corners as well

[alpine dawn]

and actually now its more common to see 40gig

[worn spoke]

Tier 1 networks are inherently going to be faster and less latent than anything you can get as a consumer at home. Comparing home internet to a data center with a connection to a backbone is not fair.

[grand geode]

Yup worked on an embedded system which was a test rig for 40gbit connections. For testing routers, switches (layer 3) etc....

[quick mist]

We digress. (:

[worn spoke]

The peering alone would give the data center a huge advantage.

[grand geode]

We had a PC with 4*40gbit cards in it vomitting IP traffic at stupid speeds.....

[rose valley]

Guys, keep it focused, please!

[alpine dawn]

we have ADD

thats hard

[grand geode]

So? Trains next? 😄

[alpine dawn]

I like trains

[quick mist]

that or stable signage or less-intermittently-visible belts

[alpine dawn]

haha

[cedar notch]

like this? https://questions.satisfactorygame.com/post/61817b49831c8520523590ac

[prime gorge]

I guess I know what you tried to post 😄

[alpine dawn]

waiting for my damn explosives to work

[rose valley]

These discussions are very fun but take them to the offtopic channel.

[alpine dawn]

wasnt me that tried to post

[grand geode]

@rose valley Don't know if its of use or not but the trains seem to break in fright platforms and its specific to when the fright platform is created in a dedicated server. Like I could drop to single player delete and re-create the fright platforms set the load / unload options and it was "good to go" from that point. From my probably skewed view point there's probably something missing in the state initialization there.....

[prime gorge]

I humbly want to disagree. It is important from where a feature request comes from. I was just wondering why someone is worrying about 8 kb/s from user requests. It's unnecessary still, don't get me wrong. But we are talking about a private hosted service with a consumer grade internet connection. Not a datacenter application. Which renders such traffic completely irrelevant. Even when 20 people would constantly ping the server.

[grand geode]

ITs a problem because its per user who has the server added in their server manager. So 50 users = 50 * 8 100 prior users? 100 * 8

[alpine dawn]

drove a tractor out to go exploring last night got off of it to walk around went to get back on and it wouldnt...ended up deconstructing and rebuilding it to make it work again

[calm saffron]

My personal priority order would be something like:

1. Server crashes
2. Client crashes
3. Anything that cannot be worked around without taking the server offline (e.g., restarting it, or having to download the save to your local machine) - I think this is mostly trains at the moment?
4. Anything that can be worked around, but the work around is tedious. Signs would be this.
5. Major visual glitches. I think invisi-belts fall here
6. Minor visual glitches.

[foggy agate]

I had none of these problems....

[grand geode]

Have not actually had a lot of server crashes. Except like rebar gun and so others which I think have been fixed now. So that areas been good for me so far....

[prime gorge]

I do acknowledge it as a problem. But I come from a datacenter perspective. Even though I agree that this behavior needs to be addressed and fixed, it's way less irrelevant for me, than for you.

[calm saffron]

Yeah, that was a general priority list, rather than “I still have server crashes!”

[grand geode]

Client crashes seem an issues.... and disconnects. I think if you just sit in a truck / explorer currently the client gets booted when the server save occurs......

and that can be a problem cause yu can end up if the client connects too quickly (before server side times out the client) they start a new character so old charater ends up stuck in machine forever (requires save editor to fix)

[quick mist]

oh yes one common issue @rose valley - Zombie Conductors.

If a client gets dropped while in a train there does not seem to be any reasonable way to get the zombie out. When the client reconnects it is a fresh spawn and the driver in the train no longer has an associated player.
("Use a save editor" is not quantifiable to me as 'reasonable')

[prime gorge]

@rose valley Just were remembering Killing Floor 2 and their web based server manager. Looking over @calm saffrons priority list and save file up/download FR it came back to my mind. Would this be a use case to have a web based server browser? Or somthing like a companion web based feature for save file management and other stuff yet to come (e.g. mods if needed to upload manually).

[grand geode]

Did have an interesting train one. 2 players collided. did a re-rail at one of the trains disappeared... but was "still there on the track" had to fix that in a save game editor as well by deleting that section of track

[calm saffron]

I think local-to-remote saves is a planned feature for the in-game server browser? I’m sure I remember reading that somewhere.

[grand geode]

That can be done by hand currently already

[quick mist]

we know

[calm saffron]

It can, yeah, it’s just mildly annoying.

[grand geode]

I was looking at this from both perspectives. From a server admin point of view which if you host a game and have 200 people go though it your going to have a problem with the. From a SW dev perspective... your going to have problems overflowing / processing too many udp packets from the recv queue. From a networking engineering perspective. Your going to have 1000's of packets per second additional to process. From a client / user perspective you put ever one of these issues on the end users network as well

[rose valley]

We know about it and are discussing alternative solutions internally. It's annoying. I run into it a lot when I play too.

[grand geode]

Would it be possible as a short term improvement to just simply force the player out of the vehicle they are in until they reconnect? Or you discussing uch better alternatives?

[rose valley]

It's a little early to tell. We're leaning on having the community come up with something to fill this gap. It's not the kind of game that lends itself well to public servers.

[prime gorge]

Humbly disagreeing. Even though 200 people would ping with 8 kbit/s it means a total traffic of 1.6 MB/s up and down. It simply doesn't affct me to loose roughly 1% of my data bandwidth. It could affect performance though. Yet again, how likely is it that 200 people have one server in common in their server browser?
Yet again it comes down to one or two icmp requests as Bogdan already confirmed 🙂

[rose valley]

This is one of the solutions we discussed. There are other possibilities too. We're still discussing 😁

[tame solar]

The zombie driver bug affects singleplayer too, whenever offline mode creates a new body

[grand geode]

Ok go write a program that has to read 1000's of udp packet per second and see how you get on 😉

[calm saffron]

Can the solution be a YEET GUN that yeets players out of vehicles?

[quick mist]

My proposal: anyone connecting spawns in the WC in the HUB if it is present.

[worn spoke]

They’re self-imposing a cap off for users being supported officially, so I don’t think the issue you’re describing is something they should concern themselves with compared to other issues.

[prime gorge]

Works for me, thanks for answering it anyways 🙂

[tame solar]

Or just, allow vehicles to be dismantled when a player is riding it

[worn spoke]

Newly connecting?

[quick mist]

or reconnecting

[lofty latch]

I want a literal ban hammer.

[worn spoke]

No

Definitely not

[quick mist]

it was not a totally serious proposal

[worn spoke]

Phew lol

[quick mist]

but it would incentivice a good transit network (:

[rose valley]

Guys, relax. We will improve bandwidth usage. It was acknowledged from the start. There is no debate here.

[quick mist]

quietly beholds the ranch populated with dead horses

[grand geode]

@rose valley One of the thing things that was kinda discussed int he group i was playing in was the accuracy of the items on the belts. That when things end up with a lot in a specific area. I guess it just overflows network bandwidth / netcode and there is probably no easy way to deal with it. One of the suggestion that came out of that was to only show items on belts that you selected marked when in a busy area and a client on multiplayer

[alpine dawn]

can you spare one for me to beat

[prime gorge]

Arma comes to my mind on this one. When someone is unconscious in a vehicle and you want to get in, you are putting the other avatar out of the vehicle. So if there is someone in the train, but not connected (aka in this rest stance), someone can push this avatar out by sitting in the train.
You would need to check that someone is safe on the spot put on, but maybe this could be a valiable solution for your solution pool?

[quick mist]

Not my ranch, not my horsemeat.

[alpine dawn]

:/

damn glue factory

[alpine dawn]

make it GTA style and just yank the body out of the car

[dapper blaze]

whats the solution for loosing inventory and respawn point where i logged out, each time my server restarts?

[prime gorge]

Or Saints Row Style and jump into the train through the front window, kicking out the driving person and take over the vehicle 🤣

[alpine dawn]

lmao

[grand geode]

@rose valley Another thing that was discussed (currently single player solution). During an autosave it displayed a warning just prior to the server. Is is possible to have something like that relayed to the clients so they know a lag spike is coming?

[alpine dawn]

haha count down timer to autosave

[grand geode]

well did have a few people teleport off cliff's and stuff lol

[rose valley]

https://questions.satisfactorygame.com
Everything is possible, given enough interest. This is the best way for us to gauge interest.

There is a lot we need to do about servers and many of the things you guys are asking for are ideas we already have and are just waiting for their turn.

However, if you request things that we already have in mind that only helps reinforce those ideas so that is not a waste of time at all. Bring it all on!

Server API brainstorming

[tropic leaf]

Just curious and putting a feeler out there for this. I've finally gotten some time outside of my work schedule to spin up an internal dedicated server to test with. Seems to be going good so far.

My single test instance appears to use about 3.8GB of RAM with CPU usage being nominal. Factoring in some overhead I'd be able to run about 20, maybe 25 instances concurrently depending on CPU load when there multiple players connected concurrently. Internet is 1Gig/1Gig fiber, so ping times should be good too.

My question would be is there any interest in the community in having a dedicated server hosted for them? I wouldn't charge anything, but I would ask that if it's not being used to let me know so I can give the resources to someone else who would use them. Another caveat would be that there would be no SLA and the servers could go down at any moment for maintenance like updates and such.

If there is enough interest, I'd need a few people to actively test an externally available instance or two once I get them setup. If you're interested send me a PM.

[alpine dawn]

well...bring it all on except golf...that one got closed(wont implement)

[foggy agate]

#looking-for-group-old would post it there too. Might be a higher chance people will look there.

[tropic leaf]

Already did, this is actually a cross post here.

[orchid onyx]

can i join someones dedicated server?,

[prisma void]

FYI for those running windows 10. An update came through today or yesterday. I could not figure out what the issue was right away. I even shut down and started the computer. Fallowed https://satisfactory.fandom.com/wiki/Dedicated_servers made sure i was getting the right info to run the server. I am running a box pc with win10 pro on it. Then connect from my laptop. I run headless steamcmd. I then desktop share my way in and start the server manually. Eventually I will run this as a service. But tell then. You need to elevate permissions. Have administrator in the command window title. So that the software can actually access and write to its directories/files. My folder is closer to the the root of the drive rather than on the desktop. Seem kind common sense, except if you use a batch file to start the dedicated server. It does not necessarily go into elevated administrative mode like it should. I keep looking for the administrator in the windows command title. If you do not see it. Most likely the update and running of the server will fail to update and you will get a dedicated server mismatch. I do not believe you need to have the server running in admin mode. but I do believe you need to have the updates in admin mode. Just something to chew on. Thank you for reading thus far.

[foggy agate]

#looking-for-group-old

[prisma void]

@foggy agate your privet messaging is disabled

[ornate pilot]

i changed from coop to multiplayer so more than 4 players could play together (with the same world)
but now still only 4 players can join even after i changed it to 10 or lower/higher in the files
is there a way to make it possible so more players can join?

[empty axle]

lol

[foggy agate]

I know, why do you ask?

[prisma void]

@foggy agate you wanted to join someone's server. I was going to offer.

[worn spoke]

They were pointing someone to #looking-for-group-old

[foggy agate]

um no I didn't wanted to 😄 I have my own servers 😄

[empty axle]

Do you guys also get Segmentation fault errors when you close your server?

[foggy agate]

no 😮

[empty axle]

weird

Also weird that no commands work in the console

[rocky sentinel]

I do, happens on linux CLI and in my docker container

[empty axle]

Ah, okay

[clever anvil]

commands not working on stdin is normal

It's only implemented on the game console, which is more akin to rcon

[rocky sentinel]

You can ignore that, most warnings are ignorable right now

[empty axle]

aight thanks

[gentle thorn]

The server doesn't read stdin, you're just typing into the void

[empty axle]

On the wiki it says there is a stop command

but oh well its in very much alpha so i don't judge

[gentle thorn]

That's if you connect from the in-game console in the server manager

[empty axle]

ctrl+c works aswell

oooh

oh fuck i just noticed i can't put our current save on the server since it uses Refined Power, guess there is no way to install these yet, right?

[quick mist]

commands are not entered on the stdin console; go to the Server Manager; there is a Console tab there for sending commands to the server

and you are correct, Xir, servers do not currently support any mods

[empty axle]

hmm, alright, thanks!

[prisma void]

um, batteries aka power storage. does not charge with biofuel generators does it. I have like eleven of them running and the battery ~ power storage is not charging. But it works for coal generators.... scratching my head. it used to work before update 5.... put now on dedicated server. wondering if anyone else can produce this....

[empty axle]

who uses biofuel generators to charge batteries lol

[prisma void]

i do...

[quick mist]

Bio burners only make as much power as is required on the network

[prisma void]

great for timing. when i am on the run

[quick mist]

batteries only intake surplus

[foggy agate]

thats expected, because I think, they only charge when power is left over and bio burnes turn down...

[quick mist]

so they do not make any power demand for the bio burners to respond to

[charred shell]

What are server requirements btw?

[quick mist]

they be on the Wiki

[empty axle]

on the wiki

[charred shell]

cheers

[foggy agate]

https://satisfactory.fandom.com/wiki/Dedicated_servers

[prisma void]

I had like 11 bois running.

you would think it would charge

[foggy agate]

no bc they turn down when not needed, but thats something for #satisfactory

[quick mist]

again: batteries do not put demand on the network, so the burners don't make excess energy for them to siphon

[foggy agate]

dosn't belong to the server side.

[empty axle]

Sometimes the power logic in the game annoys me, as an electrical engineer

[gentle thorn]

Am I the only one in this history of Satisfactory that gets Invalid authentication token errors with the dedicated server? If I search for that in the questions site or discord the only results are me.

[prisma void]

@gentle thorn firewall rules I am asuming?

[quick mist]

sounds like someone has a bad password.

[gentle thorn]

No, I can actually connect in rare circumstances.

[quick mist]

huh

[gentle thorn]

It also works perfectly when running in docker locally vs. in the cloud (also docker)

Same image

It also worked perfectly with the previous build that had the 0.0.0.0 binding bug

[gentle thorn]

Anyway, I posted a full writeup with logs to the questions site. But I would love to know how the Entry ticket mechanism works so I can take some sort of diagnostic steps.

[rose valley]

Can you link the QnA entry please? I missed the start of the discussion and scrolling up fails me now.

[empty axle]

A bummer that the dedicated server isn't open source

[gentle thorn]

https://questions.satisfactorygame.com/post/618561c5831c85205235ac56

[rose valley]

Can you delete the server from your server manager and try again?

[gentle thorn]

Sure, one sec

[rose valley]

The entry ticket has a validity of 60 seconds. I'm not sure if there are any networks that slow for that to be a problem.

It's issued when you press the Join button through the beacon channel and then immediately used to enter the game with.

Could the process have been suspended for some reason? For example, clicking on the log window if you have one and selecting characters there will make the process freeze until the selection is cleared. Stupid windows thing that we can't control it seems.

[bold atlas]

Is there a github repository for the dedicated server?

The launch script could do with some love.

[empty axle]

sadly no

[gentle thorn]

Naw, I was doing this with tail in a separate shell

[rose valley]

The launch script is generated by unreal's build system. What would you change in it? I can take a look.

[quick mist]

FactoryServer.sh seems pretty straightforward.. what's broken in it out of curiosity?

[bold atlas]

So, first off it starts with #!/bin/sh but I don't think this will work with sh. It looks like it should be #!/usr/bin/env bash.

(Yes, sh is usually bash, but on e.g. Debian it's dash)

[quick mist]

..why? not all servers have the bourne again shell

/bin/sh is practically guarnateed to be there

[rose valley]

What's the non portable bit?

[bold atlas]

It could also be rewritten to work with posix sh. 🙂

[empty axle]

If it doesn't work for you just edit the file

[gentle thorn]

Honestly who cares about posix anymore. Who's running Satisfactory on a mainframe?

[empty axle]

it should work on every linux system tho

[bold atlas]

Won't it get overwritten on update, though?

[quick mist]

I don't see anything nonportable in it at first glance

[rose valley]

No that's not how it goes. Personally I care for standards and portability. If we really have a problem I want to know about it.

[empty axle]

If it does try only giving r access to the file

[rose valley]

Unfortunately I'm also a newb so I need to rely on you to bring it to my attention.

[bold atlas]

Anyway. Let's set that aside for now; I'd want to rewrite a lot of what's in it anyway.

Here's what I'd propose doing instead:

[quick mist]

anyhow: please respond: what is the nonportable bit in it?

I do not see one

[bold atlas]

/bin/sh is not guaranteed to exist

[quick mist]

relying on a shell not guaranteed to be there is suboptimal

/bin/sh is guaranteed to exist

[bold atlas]

It doesn't on my system. XD

[quick mist]

then your system is not conformant to the standard

[bold atlas]

That's a nit. I know my own sysstem is weird.

[empty axle]

Any Unix OS has sh

[rose valley]

but isn't /bin/sh a way to invoke the system's posix shell?

[quick mist]

yes, it is

that's why it's the standard "guaranteed to exist" shell

[bold atlas]

I'm sorry I even mentioned it. Can I get to the bit I wanted to suggest?

[rose valley]

lol please yeah but don't be sorry. I just want to understand what and why we need to fix.

[quick mist]

Please.

[empty axle]

From a reddit comment

There's a /bin/sh on every Linux system but it may not be the original sh - in many cases it's a link to bash.

That leads to the complication that many people write bashisms even in scripts for /bin/sh. This will work on RedHat and Arch but won't work on Debian and derivatives where the bashisms will fail.

so yeah on some distros its a symlink to bash

[quick mist]

let's for the moment agree that the shebang should stay as #!/bin/sh- what is the change you are looking for, Baughn?

[quick mist]

but it still exists so as to conform to the standard. (:

[empty axle]

it would break so many things if it didn't exist

[bold atlas]

UE4_TRUE_SCRIPT_NAME=$(echo "$0" | xargs readlink -f)
The $0 is incorrectly quoted. I only know the bashism here, but if this is called using a path where a path element contains multiple space (e.g. a directory named foo bar) then the extra space will disappear.

Erm. And the backslashes disappeared when I pasted.

[quick mist]

no it will not

that's what the "s do

[bold atlas]

tsugumi :: ~ » cat foo\ \ bar/test.sh 
#!/usr/bin/env bash

echo \"$0\"
tsugumi :: ~ » foo\ \ bar/test.sh 
"foo bar/test.sh"

[quick mist]

however, if the results of readlink -f has a space, that needs to be fixed, yes

@rose valley this would fix the complaint about spaces:

#!/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"
"$UE4_PROJECT_ROOT/Engine/Binaries/Linux/UE4Server-Linux-Shipping" FactoryGame "$@"

[bold atlas]

(Obviously though, this is a real edge case.)

That's the one I was going to suggest. But I'm not sure nested quotes like this work with posix sh.

[rose valley]

Can someone, just for reference, paste the original script here please?

[quick mist]

they do

sure, here's the original

#!/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"
"$UE4_PROJECT_ROOT/Engine/Binaries/Linux/UE4Server-Linux-Shipping" FactoryGame "$@"

[bold atlas]

What I usually use in my own scripts is this:

#!/usr/bin/env bash

set -eu -o pipefail
cd "$(dirname "$(readlink -f "$0")")"

(YMMV. I don't think it fits here.)

[gentle thorn]

Why target POSIX when Linux isn't POSIX. There's no OSX or UNIX version of the dedicated server

[rose valley]

are xargs and readlink guaranteed to exist on any posix system?

[quick mist]

yes

[rose valley]

no but some day we might want it so why make the life of our future selves harder now?

[bold atlas]

Let me run this through dash, then, to see what happens. 🙂

[empty axle]

Linux is POSIX compliant

[quick mist]

or tries to be

until some one says "I don't like /bin so I moved it to /executables" and then complains that no shell scripts work anymore (:

[gentle thorn]

This was the key. While I've been waiting around for my server to deploy I've had the game open and sitting on the Server Manager screen. If it idles there for 60 seconds I can't connect. If I restart the game and don't idle, I can connect. It seems like the game needs to refresh the tokens before a join attempt.

[rose valley]

Interesting. We do generate the token immediately before a join attempt but maybe we use the wrong timestamp for it.

[gentle thorn]

It's classified as "mostly" POSIX compliant.

[rose valley]

Would you mind posting this to https://questions.satisfactory.com ? Currently off the clock with nowhere to write it down but this is a very good find.

[bold atlas]

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

This works on bash, zsh and dash.

Haven't tested ksh and csh, but no-one uses those, right?

[rose valley]

But not on fish gdamnit!

[bold atlas]

Nothing works on fish, you degenerate. 🤣

[quick mist]

I don't even use fish, and look at my name

[rose valley]

I really like fish. I wish it had a posix compliant mode. Can't use it as a default shell anywhere.

[bold atlas]

They're working on it. The lastest version does actually support sh-style assignments.

[solar remnant]

It has a plugin for that I think or something like that...

[rose valley]

note to self: google that soon

[bold atlas]

I really want to throw in a set -eu at the top of that script. But I'm not sure those work in dash, and testing this was annoying as heck.

[rose valley]

Can we get a qna post on the portability issue though, to make sure we get something out of this exchange?

[bold atlas]

Writing one

[solar remnant]

https://github.com/oh-my-fish/oh-my-fish

[bold atlas]

'tis a pity the server doesn't come with a shell.nix

[alpine dawn]

just checked on one of my ubuntu servers

[lofty latch]

That's alarming if dash has any differences in behaviour

[bold atlas]

https://questions.satisfactorygame.com/post/618570b9831c85205235aca0

[bold atlas]

Tons. But I checked to make sure this works on dash.

[quick mist]

much like vim when invoked as vi, dash will when invoked as sh run in Bourne Shell Compatibility Mode

[bold atlas]

(That is not how I checked. But good to know.)

[rose valley]

bash too behaves differently depending on how it's invoked.

[alpine dawn]

seems Redhat likes bash

[bold atlas]

Most systems use bash. NixOS is just weird in not having anything except /bin/sh and /usr/bin/env.

(And mine doesn't have /bin/sh, because I'm trying to make sure none of the OS scripts depend on it.)

((Tip: They do.))

[quick mist]

Most distos have bash and use it extensively, but they all also have something responding to sh

[rose valley]

I'm honestly pleased the server runs on all these exotic distros. (Exotic to me, I'm a fedora newb).

[alpine dawn]

dont know if I have anything else I can test

[rose valley]

Alpine/busybox?

[quick mist]

Gentoo, LFS, Arch, Sourceror...

[alpine dawn]

Gentoo must work...I saw someone wrote up a startup howto in the wiki

[rose valley]

Arch does bash. (I used Arch once, btw)

[alpine dawn]

but gentoo will likely use whatever you ask it to

haha

[quick mist]

KNOPPIX 😄

[rose valley]

Arch has an AUR package. I'm assuming it must work.

[quick mist]

I wonder if there's a package in BSD's linux ports system yet/already? (:

[vocal jungle]

but will it run on a TI-84 plus

[bold atlas]

It does not. I have to use steam-run, which provides an Ubuntu-like chroot.

But that's not your fault. NixOS is weird. Super nice, especially for servers, but not Linux-like at all. 😛

(Though if you want me to officially package it, I'd be happy to do so. Heh.)

[lofty latch]

Does it run on Musl based distros?

If so that implies glibc is statically linked.

[bold atlas]

ldd says no.

libc.so.6 => /nix/store/mij848h2x5wiqkwhg027byvmf9x3gx7y-glibc-2.33-50/lib/libc.so.6 (0x00007f8a8f6bb000)

...wait. How on earth did that work.

...wow that's a lot of patches to ldd.

[elfin idol]

Also Update, I managed to make it work, it was the firewall for some reason. I reinstalled everything multiple times, and all of a sudden it just worked.

Thanks to everyone who helped!

[alpine dawn]

seems to be a small handful of non-statically linked or provided libs

[bold atlas]

All of which should be available everywhere.

Using steam-run for this is massive overkill. I wonder if anyone else is on nix, though.

[alpine dawn]

libUE4Server-HTTP-Linux-Shipping.so interesting...

likely just default required for UE4 server though

but the capability is there

[rose valley]

It is. But https is not. Should exclude that or something.

[gentle thorn]

Really? I swear it was complaining about missing ca-certificates in my container

[alpine dawn]

cut down on a few KB of ram

XD

[reef gull]

That, and a MB or two of storage

[gentle thorn]

I purge everything in the debug file manifest to save myself 2.5gb of image size

[alpine dawn]

4.3MB actually

just checked

haha

[rose valley]

That's the Eos plugin. Which is also unused ATM...

[reef gull]

oooo.... 3-4 floppy disks...

[alpine dawn]

yeahhhh

haha

[reef gull]

big savings

[alpine dawn]

its free real estate

use that 4.3MB on something better

[bold atlas]

Does the dedicated server fork before saving?

[alpine dawn]

its more of a spoon kinda server

😛

[reef gull]

🤦‍♂️

[bold atlas]

Sporks are plainly superior. No, but if it did that would mostly eliminate the save stutter.

[alpine dawn]

right...so spork the save file

[ashen sentinel]

yes

[quick mist]

Would it though? The cycles still need to get spent to move the data to spinning platters or toasty chips

[alpine dawn]

if you got both you are halfway to nachos

[bold atlas]

Evidence says yes. Most computers these days have spare memory and CPU cores.

It reduces the cost of a save to just that of copying the process image, which is less than a second even in the worst case.

How long the clone process takes to write the save file afterwards is largely irrelevant.

(Reasons not to do this: You still need to support people who don’t have that memory. Also, all mutexes need to be unlocked, and the game quiescent, before the fork. This may or may not be feasible.)

(Also it’s impossible on windows.)

[alpine dawn]

damn...I just fired up htop for the first time while running the dedicated server...I didnt realize how many threads it runs

30 threads on an idle server

haha

[bold atlas]

In honour of the dedicated server, here’s one that’s about to be attacked by spiders that I drew.

Kinda-sorta drew.

[alpine dawn]

interesting that there are varying levels of niceness

[rose valley]

That's... Pretty darn clever.

[bold atlas]

Works for Factorio. Watch out for thread and mutex behaviour, though!

[lofty latch]

Ooh that reminds me, we definitely need (de) quiesce commands for the server via its API to allow for safe backups.

[bold atlas]

Not really. ZFS snapshots do the trick.

Would be nice to have, yes!

[lofty latch]

Cross platform and regardless of host file system

[bold atlas]

An easier fix might be to rename the save as the last step.

Use a .tmp suffix before.

[alpine dawn]

lmao I didnt even remember what filesystem I have on my server

had to go look

[bold atlas]

ZFS and Btrfs are the only two I’d consider storing anything valuable on. YMMV about Btrfs.

(Are Satisfactory save games valuable? I’ll leave that up to you.)

[alpine dawn]

ext4 for my / which is running on a thumb drive and XFS on my storage which is hardware raid5

[bold atlas]

So that will save you from a disk failure, but you might lose all your data in case of controller failure. mdraid is generally higher performance, too.

It won’t prevent file corruption, though!

[lofty latch]

Store your save game in a mongodb database on a machine with no power redundancy!

[bold atlas]

Feeds @lofty latch to a fluffy-tailed hog.

[alpine dawn]

I inherited the box from my friend who was upgrading, its over 10 years old I think

[quick mist]

save integrity is a solved issue though.

• save to temp file
• did save process exit clean? yes?
• rotate old saves
• move temp save into place

[alpine dawn]

its gone through at least 4 different hard drives since I have owned it

but somehow keeps going

[bold atlas]

It’s obviously your choice. But I’d keep backups if I were you.

[lofty latch]

I'm worried more about other processes wanting to be able to hold a file handle open for a while

[bold atlas]

With a few fsyncs in the middle, but yeah.

[quick mist]

Linux already handles file handle contention perfectly well

Windows OTOH..

:shudder:

[bold atlas]

File handle contention is only a thing on windows.

[lofty latch]

Not when you're me and I hold exclusive locks on files xD

Because I'm mean

[quick mist]

oh it's a thing on windows; on Linux the contention is "Kernel, give me a handle. Thanks!"

[bold atlas]

Sooo I’m pretty sure they’re tearing that code out of the kernel. LWN had an article about it recently.

[quick mist]

I once saved a couple thousand dollars of labor by telling everyone in the office to take their hands off the keyboards while I dived into /proc to recover "deleted" files that still had open handles

[bold atlas]

How would you like a new job? 😄

[quick mist]

...how's it pay?

[alpine dawn]

this is a perfectly optimal raid setup right?

[bold atlas]

I hear we pay pretty well.

<- Google SRE. If you’re familiar with that level of detail, you’d be better than 80% of the people I interview.

[reef gull]

cringes in sysadmin Never, never mix drives... it causes bad things to happen

[alpine dawn]

haha

dont worry, I also have write cache enabled with no BBU

[quick mist]

..I failed at the second or third interview for a Google SRE job a few years ago

[reef gull]

falls out of chair unconscious

[bold atlas]

Aww. Well, It’s a luck-based mission.

[calm saffron]

That's pretty common - a lot of people have to make two or three attempts to pass FAANG interviews 😕

[lofty latch]

If you know what REISUB is about, you're pure OG.

[bold atlas]

Mm. The general rule seems to be you can try once per year.

It sounds like you passed the screening anyway.

[sick stratus]

Trains on dedicated bugged load/unload?

[quick mist]

(I loved the pre-interview "rate yourself" stage. "out of ten, how good are you at ___? here is what each score means. [...] 10/10: "I wrote the book on this. (please provide ISBN number)")

my first interviewer commented that I was the first person to use fractional values for that questionnaire they'd seen

[bold atlas]

Yup. It’s how we select interviewers who speak your language.

[alpine dawn]

I believe someone mentioned issues with trains

I dont know personally as I havent gotten to them yet

[quick mist]

actually now that I think on it, IIRC "I wrote the book on this" was 9/10; 10/10 was "I created this"

[calm saffron]

Vehicles in general and trains in particular seem to be the buggiest bit of multiplayer/dedicated servers right now

[lofty latch]

@quick mist π/10

[quick mist]

I was tempted but restrained myself to rational self-assigned scores

[lofty latch]

Lol

[calm saffron]

And 0/10 is "I had a lobotomy and don't even recognise this concept anymore" 🤔

[bold atlas]

I mean there’s a guy working here who’d answer 10/10 for Python so…

[quick mist]

I would not quantify myself as transcendant

[calm saffron]

π+e/10 😏

[sick stratus]

welp for me not working load and unload on trains... and if I get on the train the autopilot stop working...

[gentle thorn]

Tech interview process is a garbage fire of unchecked egos and dumb traditions.

Source: did interviewers for a major tech company

[bold atlas]

Which one?

…actually, we should move this conversation. And I’m about to dive into a game, so I’m moving to /dev/null for a while.

(Not much of a a view, but so spacious!)

[gentle thorn]

The one run by a billionaire and has a market cap >$1T

[quick mist]

There are two of those

[gentle thorn]

I think there are more than just two

But I supposed that depends on the stock market

[calm saffron]

I think the leader of any company worth over $1TN is probably a billionaire. It's definitely true for Apple, Amazon, and Tesla.

[stray bone]

Is it normal for the server to suddenly "restart" itself? Got a message on the client: Server will be restarting in 60 seconds

[gentle thorn]

I had heard the server includes a 24hr automatic restart as a precaution to recover from hypothetical memory leaks

[quiet fossil]

Normal. Game has always performed a restart if you run 24 hours. On the dedicated server, it'll kick off the clients, but they can log back in in short time.

[sour yoke]

How goes the headless servers for everyone?

[quiet fossil]

Stable after a few patches went in. We're starting to see some good uptime. Still some quirks to iron out, but overall super happy with it. 🙂

[sour yoke]

That's been something I wanted. I host a lot of game server for friends

Got back into this game recently with a new save and another group to hear of update 5 stuff

I did a double take when I heard them making an offical Docker image

[gentle thorn]

If you're interested in running the server in Docker, a few people have accomplished that already

[sour yoke]

Yup it is how I run all my game servers

if they didn't I would have made my own

[gentle thorn]

I couldn't wait 😄

[sour yoke]

Just wasn't expecting Coffee Stain to offer their own image.

[alpine dawn]

I likely should have fired it up in some kinda VM

but meh

the rest of my server is a clusterfuck as it is

[sour yoke]

It's what I used to do mostly due to running multiple MC servers modded, creative, and a vanialla survival usually.

[reef gull]

Yeah, guess I'll have to put more effort into the Remote Monitoring mod, XD

[alpine dawn]

lmao it would be great if someone created some web status thing that could be run on any webserver that showed the map and all the buildings in real time

[sour yoke]

I monitor my stuff through Portianer

works for me for now.

[gentle thorn]

I'm running it in Azure

[reef gull]

Working on getting more stuff added, but that's what Remote Monitoring does... XD puts it in JSON and Geo-JSON

[rose valley]

At times I feel a little doubt about that seeing how easy it is for the community to create high quality images on it's own. OTOH, I think it would be nice if studios would embrace alternative such as Docker more often so here's to that.

So our image won't bring anything new to the party. It will just have our name on it.

[sour yoke]

Oh for sure. I think it's been gaining more traction as of late despite Docker itself kind of getting the boot as of late within the grater Deve Ops Container conversation

[reef gull]

Still allows for people to know that the Docker they are getting is legitimate and safe... as much as I don't like Docker, that is one thing you guys are doing right.

[alpine dawn]

next up fully bootable satisfactory server OS

XD

literally just a debian server running satisfactory dedi

lmao

[bold atlas]

NixOS is definitely better suited for that. I could throw one together in about twenty minutes. 😛

[alpine dawn]

give it a cool ass Grub menu and boot screen

[sour yoke]

@reef gull This big time. Official images have a lot more peace of mind for me.

[rose valley]

I'd like to have a docker image without the steamCMD dependency but I'm worried about download times. It would not lend itself well to layering I fear. What do y'all think about it? Also, dockerhub is what we are going for. Remember it can be used from podman as well.

[alpine dawn]

you guys going to self host updates?

[reef gull]

Oh yeah, I would prefer an official image to something someone made up... regardless of my opinions on Docker, it is still the right move. Sorry if I said it in a way that was confusing.

[gentle thorn]

Use a multi-step build. Use steamcmd to install, then use a new clean base and do COPY --from=steamcmd /server /server

[rose valley]

No. We'd push them to dockerhub.

[alpine dawn]

ah

[sour yoke]

Oh I get it. Containers have kind of become this divine all-in-one solution for everything despite the many issues it has and you probably should just run a VM instead.

I am just a dude who hosts game servers on a repourposed laptop. So those downsides don't have the same effect on me as something in live production making people money.

[reef gull]

I prefer VMs personally, I have a homelab at home for my "college"

I'm just wondering how the mods will work with the DediServer, I know the servers aren't fully stable yet to be throwing a new monkey wrench into things...

[alpine dawn]

so....are each laptop its own container in this scenario? I am just picturing a pile of old thinkpads each running a different dedi

[sour yoke]

It's just one laptop running a bunch of Docker containers

[alpine dawn]

oh

haha

I liked my visual better

like 20 old thinkpads stacked

[sour yoke]

I think your senario is called a cluster.

[alpine dawn]

in like the corner of your livingroom

[sour yoke]

could do that at some point if I get more broken laptops to turn into servers. Run Swarm or Kubernates on them for no reason. I have no need for scalability but would be fun to mess around with.

[alpine dawn]

haha

ubuntu MAAS was interesting and that would be like the perfect use case

metal as a service

[reef gull]

Hey, just make sure to have neat cabling, don't need no wire spagetti

[sour yoke]

There is a reason I never post pictures of my factories.

cough

[alpine dawn]

satispagettiry

[worn spoke]

I imagine mods will need a command line installer or something; I wonder if the modding community is already working on that?

[reef gull]

satisbuggory, satispagettiry.. what else we got? 🤣

[gentle thorn]

Random poll: How long is the Total Save Time for a basic dozen-or-so machine save on your dedicate servers? I'm trying to gauge my file storage speed

[alpine dawn]

I dont even notice saves

lmao

[gentle thorn]

Sure, but its in the log

[alpine dawn]

oh...

let me look

shit...remind me again where that is...is it in the home dir or in the installed dir?

[gentle thorn]

Install path /FactoryGame/Saved/Logs/

[alpine dawn]

is it World Serialization (save)

[gentle thorn]

Yeah, starts there, then it has compression, write to disk, backup, then total

Compression: 0.188 seconds
Write To Disk: 0.001 seconds
Write Backup to Disk and Cleanup time: 0.001 seconds
Total Save Time took 0.345 seconds```

[alpine dawn]

k I just greped for save

[2021.11.05-08.17.05:985][640]LogGame: World Serialization (save): 0.800 seconds [2021.11.05-08.17.07:264][640]LogGame: Compression: 0.499 seconds [2021.11.05-08.17.07:264][640]LogGame: Write To Disk: 0.727 seconds [2021.11.05-08.17.07:265][640]LogGame: Write Backup to Disk and Cleanup time: 0.727 seconds [2021.11.05-08.17.07:273][640]LogGame: Total Save Time took 2.089 seconds

thats the last one

I can go look for more

[gentle thorn]

How big is your world?

[alpine dawn]

is there like a node count or something in here? we are just starting oil production so tier5 I think?

[gentle thorn]

Ah tier5, probably got a lot to save 😄

[alpine dawn]

well...that and its saving to a shitty flash drive

so there is that

[valid geyser]

[2021.11.05-20.03.22:004][908]LogGame: World Serialization (save): 1.870 seconds
[2021.11.05-20.03.23:230][908]LogGame: Compression: 1.222 seconds
[2021.11.05-20.03.23:230][908]LogGame: Write To Disk: 0.002 seconds
[2021.11.05-20.03.23:230][908]LogGame: Write Backup to Disk and Cleanup time: 0.003 seconds
[2021.11.05-20.03.23:259][908]LogGame: Total Save Time took 3.125 seconds

all tiers done now i have to wait

[alpine dawn]

I should symlink the .config dir in home to somewhere on the raid

so the saves are stored on that

[gentle thorn]

Probably a good plan

[burnt blaze]

ok. Someone recommend a linux distro for me to cut my teeth on.

[alpine dawn]

ubuntu server is super easy...just going to throw that out there

[gentle thorn]

How deep do you want the cuts to be?

[alpine dawn]

if you want to go off the rails then go with something really hard like gentoo

haha

[burnt blaze]

Not quite hurt me, daddy levels but I do sysadmin work on Windows

[quick mist]

Just install WSL and you can try out a dozen distroa

If WSL is not an option, then depending on what you want to learn, in no particular order:
Ubuntu, Ubuntu Server, Mint, Fedora, Genoo if you really want to learn

[calm saffron]

NixOS 😏 😂

[gentle thorn]

I use Debian as it tends to be a popular option for many of the docker images I tend to run (if you ignore everything running Alpine these days)

[alpine dawn]

yeah nobody but you runs Nix

lol

you are like the lead dev and single user

XD

[calm saffron]

Oh I don't run NixOS, I'm just suggesting it as an even more masochistic option than Gentoo 😛

[alpine dawn]

well yeah

[quick mist]

masochistic? LFS or Sourceror

[burnt blaze]

ubuntu server it is

[quick mist]

not a bad choice. Lightweight, relatively current repos, well-documented.

[sour yoke]

I am a Ubuntu boy it's all I ever used.

[alpine dawn]

documentation and community support are what you want when starting out

[calm saffron]

I'd start with Ubuntu, as it's quite popular and it will be easy to search for any issues you have. Once you're familiar with Linux, you can look at other distributions, and see if any of them tickle your fancy.

[alpine dawn]

and ubuntu is widely used

so you should be fine

[quick mist]

"Which distro should I start with?" is the kind of question that you can ask 30 folks and get 75 answers hehe

[alpine dawn]

yeah

[burnt blaze]

Oh I figured. I just wanted a direction. Not going to think too hard about it for now

[sour yoke]

Ubuntu isn't a bad one

and for the meme, Arch

[quick mist]

if you have the hard drive space and want to shop around, get VirtualBox for zero dollars, and just go ham

[burnt blaze]

It's more time that I'm short on

[quick mist]

fair

[bold atlas]

NixOS is really nice for anything you want to keep working over time. Unlike every other distro, it doesn't degrade over time / with reconfigurations.

[quick mist]

and mood.

[gentle thorn]

Ubuntu/Debian are the right choices if you want documentation. Most how-to-install guides are written for those two distros.

[bold atlas]

If https://github.com/Baughn/machine-config/blob/master/tsugumi/configuration.nix#L106 looks like heaven, then you should try nixos. 😛

[burnt blaze]

that does not look like heaven

[alpine dawn]

I get the feeling that @bold atlas prefers servers that you have to plug a serial cable into to get a console as well

[burnt blaze]

Maybe ones where you need to switch those little pins on storage and such

[alpine dawn]

haha

[quick mist]

I'd rather fiddle with DIP switches than jumpers any day

[alpine dawn]

at least dip switches you can use a small screwdriver to quickly change the settings

jumpers just get lost

haha

[dim grail]

I got crashed out of my server, and when I rejoined I'm on a new character

How do I fix this?

[alpine dawn]

you mean you dont have any of your previous inventory or your game is new?

[dim grail]

No inventory, and my hot bar is reset

[sour yoke]

Out of all these updates, I am probably the most excited about a hole in the floor

Soon I shall make an update 5 world and contibue to the many bug reports.

[vast mango]

Hi, I don't know what I entered as my admin password, is there somewhere I can retrieve it or see what I entered ?

[royal halo]

can someone help me with the update for the server on linux pls (i've forgot the command line)

[desert canyon]

@vast mango

How do I reset the Administrator password?

In the location where you can find the server's saved-game files, look for a file called ServerSettings.port and delete it. This will delete all server settings relating to passwords, server name, and session name, and you will be asked for a new Administrator password when connecting as when the server was first set up.

[desert canyon]

steamcmd +login anonymous +force_install_dir ~/SatisfactoryDedicatedServer +app_update 1690800 validate +quit

[royal halo]

thx

[desert canyon]

echo "alias satisfactory_install_update='steamcmd +login anonymous +force_install_dir ~/SatisfactoryDedicatedServer +app_update 1690800 validate +quit'" >> ~/.bashrc

if you trust random commands from the internet, this will add it to your ~/.bashrc

then you can just type satisfactory_install_update

[vast mango]

Thank you very much 🙂

[alpine dawn]

this is why I just followed the wiki to make it update on restart

systemctl restart satisfactory.service

done

[desert canyon]

yeah I have yet to get it setup as a service.

[reef gull]

Wiki has a nice bit for doing that, so you don't have to from complete scratch

[desert canyon]

I spent 3 hours trying to get docker to work, then gave up and just spun it up CLI

[alpine dawn]

lol

[desert canyon]

no you must alias everything

it's important

[alpine dawn]

yeah I just install to the bare os

I should have spun up a VM of some kind

[gentle thorn]

I have it working in Docker if you have any questions

[desert canyon]

well it was running, but it wasn't showing up as online. So I'm assuming traffic wasn't getting to Docker when it was hitting my VPS.

and Docker is way outside my wheelhouse, so I moved on lol. If you have any idea why wolveix's instructions wouldn't account for traffic getting to it, I'm all ears.

[alpine dawn]

good ol virtual networking

[gentle thorn]

Yeah, that does seem like a networking issue. It should be as easy as exposing the three UDP ports, and if the host is routable then you're golden

[desert canyon]

[2021.11.05-21.48.18:166][540]LogExit: Exiting.
Shutdown handler: cleanup.
Assertion failed: !HasCommands() [File:Runtime/RHI/Public/RHICommandList.h] [Line: 3816]

Pure virtual function called!
Signal 6 caught.
Segmentation fault

thank you for playing satisfactory?

[alpine dawn]

speed run

[wraith pebble]

what is the default setting for the multihome argument?

[quick mist]

The default setting is to bind to all interfaces on all IPv4 and IPv6 addresses available

[edgy crag]

Hrm. Everything was working so well. Now it seems my tractor testrun is borked now

[sharp pagoda]

Does anybody experienced lags since the update yesterday?

[quick mist]

only in other games in an Experimental state

[dusty urchin]

do tractor record menu on dedicated servers work? Or is it a known issue?

[edgy crag]

I just got mine fixed

seems to be working alright now

[lucid cairn]

Has anyone had luck setting up multiple servers on a single host yet? I followed what would be the normal rules but its not detecting the server

[quick mist]

Because the beacon port cannot be changed, the only way to run multiple instance on one host is with virtual network interfaces and extensive use of the -multihome switch, or containerization/virtualization with bridged networking

[burnt blaze]

I thought the ports could be changed, just not redirected?

[lucid cairn]

The beaconport just steps up when it detects its already in use.

But I did get it working. Was weird

In reality, I forgot to set my firewall range to allow the new ports, so thats why it didn't work. But in unraid I set it up as norm.

[sour yoke]

unraid is best raid

[quick mist]

what about JBOD? (:

[sour yoke]

JBOD is the way I roll

[quick mist]

same. JBOD + SyncThing for the the important stuff, which gets pulled to three places

[sour yoke]

I love Sync Thing

I done away with third party cloud storage solutions

[sharp pagoda]

Got this issue too

[quick mist]

yeah, I was really happy to find SyncThing after BitTorrentSync got bought and killed

[sour yoke]

That's the thing about open source, you can always fork it from somewhere.

[dusty urchin]

thank god,

not only me

[sour yoke]

though you do have to find someone else to maintian it now for free.

[sterile nacelle]

Ok - quick question - shouldn't the server prevent windows from sleeping? I keep having game drops and find my server snoozing

[lofty latch]

I'd doubt it since sleep on client OSs is determined by HID activity

so either disable sleep or use Powertoys to temporarily disable it

[sterile nacelle]

On both windows and macos running software can request an exception to prevent sleep.

[lofty latch]

sure but servers shouldn't have sleep at all

[sterile nacelle]

I'll just start the epic client. It never allows a machine to sleep :p

[split zenith]

is anyone else getting this error from time to time while running the linux dedicated game client from steamcmd? [2021.11.05-23.21.21:144][804]LogOnline: OSS: Creating online subsystem instance for: STEAM [2021.11.05-23.21.21:144][804]LogGenericPlatformMisc: Error: SetEnvironmentVar not implemented for this platform: SteamAppId = 526870 dlopen failed trying to load: steamclient.so with error: steamclient.so: cannot open shared object file: No such file or directory dlopen failed trying to load: /home/steam/.steam/sdk64/steamclient.so with error: /home/steam/.steam/sdk64/steamclient.so: cannot open shared object file: No such file or directory [S_API] SteamAPI_Init(): Sys_LoadModule failed to load: /home/steam/.steam/sdk64/steamclient.so

[spiral root]

That error is explained on Wiki how to fix.

[split zenith]

ah right. sorry about that

[spiral root]

No need to appologsise.

[reef summit]

Anyone Else getting this issue? Items are disapearing or not building in place on a dedicated server, Even when i am the only one on it.

[quick mist]

belts going invisible but still functional are a known issue

[austere mural]

Anyone seen a server crash where its logging FGfoliage removal instances failed and after a bit crashing out>?

has a run of those instances at different numbers then just dies

[summer ravine]

so I'm not able to "Cancel Docking" for my train and its stick
i am on my own dedicated server that have I restarted a few times
any ideas?

[half shuttle]

i see those blue unbuilt items, i'm having a similar issue, but my docker log is spamming "container x not found" if your using a docker too, then restart the container. it'll work for a little while

[summer ravine]

i'm not even able to remove the train

[marsh thorn]

are trains still fucked on dedicated ?

i couldnt get them run probably

[keen vortex]

I can't seem to be able to connect to the server. I'm running Pterodactyl. When attempting to connect, the console is spammed with this:

[2021.11.06-00.31.23:924][888]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.23:924][888]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.23:957][889]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.23:957][889]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.23:990][890]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.23:990][890]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:024][891]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:024][891]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:057][892]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:057][892]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:091][893]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:091][893]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:124][894]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:124][894]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:157][895]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:157][895]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:191][896]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761
[2021.11.06-00.31.24:191][896]LogNet: NotifyAcceptingConnection accepted from: my.ip.add.ress:25761```

And this shows up:

Anyone had this issue?

[molten leaf]

I'm getting an old server, I don't know the exact specs but it's got two physical i5-2k's in it and 16 gigs of ram. If the OS can see both cpu's, can Satisfactory make use of them?

[half shuttle]

as far as i know, the server only runs on one thread, so it's doubtful. cant see into the future though

[rocky sentinel]

Check your ports, my experience with that is it's a port issue. One more more of them can't communicate.

[glad pasture]

Im having some trouble with my server not opening after the update 0.5.0.7 on windows

i tried adding -multihome but that didnt help

any ideas?

[feral dawn]

@glad pasture show whole startup line plz

[glad pasture]

FactoryServer.exe -multihome=0.0.0.0 -log -unattended

this?

or when I actually click it

[feral dawn]

ye that, one mo

try

FactoryServer.exe ?listen -multihome=0.0.0.0 -log -unattended

[glad pasture]

ok one sec

nope still crashes

a second after I start it

steamcmd +login anonymous +force_install_dir C:\Users\ejohn\Documents\steamcmd\satisfactorydedicatedserver +app_update 171579 +quit

this is what my updator looks like

[feral dawn]

when u say it crashes, is it just throwing an error to the console or proper crashing out?

[glad pasture]

like the console terminates half a second after it starts

I cant even see it

[feral dawn]

Ah righto soo

Do u have like a .bat file or something?

[glad pasture]

yes

start.bat and update.bat

[feral dawn]

okay after the FactorioGame.exe line, type pause on a line by itself

[glad pasture]

are the 2 i was changin

[feral dawn]

and that will stop the server from closing the batch window

so u can see the errors

(if any)

[keen vortex]

Thanks, looks like there's something up with the startup convars. Specifying them doesn't let me connect.

[glad pasture]

should it be -pause

[feral dawn]

on a new line

like that as an example

[glad pasture]

still crashes

[feral dawn]

yah but now it should stay open, and show u the text

perhaps shows why it crashes?

[rocky sentinel]

What's your startup command?

[glad pasture]

wanna join a call so I can stream it?

[keen vortex]

Out of curiosity, has anything happened to you? I generally consider mixing drives a good thing (mostly within the same spec). Helps avoid failures where an entire batch is faulty. Interestingly I had the exact same thing happen a few months back when a bunch of Samsung 512GB NVMes failed in quick succession.

[keen vortex]

It was ./FactoryServer.sh ?listen -Port=7777 -ServerQueryPort=15777 -BeaconPort=15000 -log -unattended -multihome=0.0.0.0

Tried without the multihome, didn't work either.

[rocky sentinel]

This is in pterodactyl?

[feral dawn]

looks like pter to me

[keen vortex]

Ptero and I ran it on the host directly too, didn't work.

[feral dawn]

(as that's what im running too lol)

[rocky sentinel]

you don't need -log and -unattended, that's included in the FactoryServer.sh

[keen vortex]

Ah okay, my bad I didn't see.

[feral dawn]

ngl bit odd that ur starting a .sh script at all for pter

[rocky sentinel]

here's my ptero startup line ./FactoryServer.sh -multihome=0.0.0.0 -ServerQueryPort={{SERVER_PORT}} -BeaconPort={{BEACON_PORT}} -Port={{GAME_PORT}} which translates to ./FactoryServer.sh -multihome=0.0.0.0 -ServerQueryPort=62007 -BeaconPort=62008 -Port=62009

works perfectly, assuming your allocations and variables are setup for the instances properly

[feral dawn]

Also for what it's worth

it seems that if you're in a buggy (at least for me!) you'll get timed out from the dedi server if you were the first player to join it.

which causes ur body and buggy to be inaccessable.

[dawn crown]

Anyone else having issues with truck stations not loading or unloading on their server?

[feral dawn]

Sorry bud, i'm in a call already.

[glad pasture]

ok np

i appreciate your help thus far

i still dont know why it wont pause though

[feral dawn]

You're gonna have to share a screenshot or something of the error that prints to the console, if any

[glad pasture]

is there a log somewhere?

[feral dawn]

So, u start the batch file - it opens command prompt, the server process crashes/and or terminates and the command prompt window dies, yes?

[glad pasture]

yes

like within 0.5 seconds

[feral dawn]

It's odd for it to erm

kill the command processor tooooo

U can try something else

go into the folder where FactoryGame.exe is.

[keen vortex]

Thanks, that works. Was most likely the duplicated -log and -unattended 🙂

[feral dawn]

hold Shift and Right Click a empty space within that folder (so, right click on no files/folders)

hopefully in the menu will be "Open In Command Prompt" or "Open In PowerShell"

[rocky sentinel]

Probably don't need -multihome=0.0.0.0 either, but I haven't tested that myself yet

I think that was fixed in the last update

[keen vortex]

Yeah, seems so ^

[rocky sentinel]

What shutdown command are you using to stop the server in ptero? I'm having problems getting it to stop gracefully with the ^C command for SIGINT, basically ptero is just killing the container which doesn't write the state info on a graceful shutdown.

[feral dawn]

are u using the semi-official egg?

[alpine dawn]

lmao thats like pulling the plug from the wall

[rocky sentinel]

I wrote the egg myself on day 1

[alpine dawn]

it was

[glad pasture]

okay i have powershell open

[feral dawn]

According to the semi-official egg, it's this...

[keen vortex]

Haven't gotten it working 😔

[feral dawn]

which seems a bit odd lol

https://github.com/parkervcp/eggs

u can find it here.

[rocky sentinel]

That's not the ptero SIGINT command, that just kills the container

[feral dawn]

Okay great

Oh Carus, i see - perhaps the server doesnt have any official stop command i guess.

@glad pasture paste this into the box

[keen vortex]

🤔

[feral dawn]

yeah but that isnt from STDIN i think?

[rocky sentinel]

the problem is the console doesn't accept input from STDIN, so none of those work

[keen vortex]

Oh oof

[feral dawn]

@glad pasture .\FactoryServer.exe ?listen -log -unattended

i expect it to crash still Eric, but maybe it'll stay open to say why

[burnt blaze]

Can someone tell me why this isn't working?

[feral dawn]

that looks rather wonky for sure.

how are u opening steamcmd?

[burnt blaze]

it's on linux

[feral dawn]

cause +app_update is a command line

[keen vortex]

Yeah app_update is a separate command

[burnt blaze]

So you can't do the + to add on a command?

[keen vortex]

no it's for adding convars

[feral dawn]

u use + on the command line

like ./steamcmd +app_update bla bla

if ur typing it in the cli, take off the +

[burnt blaze]

I see.

[feral dawn]

u sure u opened powershell where FactoryServer.exe is?

[burnt blaze]

Thank you!

[glad pasture]

oh factory server

you said factory game, I got confused lol

[burnt blaze]

Can you list the contents of that folder and make sure factoryserver.exe is in there?

[feral dawn]

Sorry Eric, it's late 😄 lol

[glad pasture]

all good one sec haha

it opened

stayed open longer and then terminated a few seconds later

[feral dawn]

did powershell stay open?

[glad pasture]

yeah powershell is open still

but it went to steam cmd

and didnt log anything

heres the folder contents @burnt blaze

[feral dawn]

Have u done a validate at all?

[burnt blaze]

Can you screenshot it with the folder path?

[glad pasture]

no i have not validated anything

it has been running since the most recent update

[feral dawn]

Might be worth getting steamcmd to validate the contents, just to be sure

[reef gull]

SSDs might have less issues, but spindle discs are more temperamental. There was a disaster recovery company I used to work support for that, we used to beat the crap out of the HDDs in iOps for the deduplication... I saw a lot of issues with mixed disks. After that, I refrain as much as possible.

[glad pasture]

how would I go about doing that?

[feral dawn]

Open SteamCMD to get the SteamCMD command line

[glad pasture]

got that

[feral dawn]

login anonymous

force_install_dir .\satisfactorydedicatedserver

app_update 1690800 validate

should do it

[glad pasture]

validating now