#dedicated-servers
1 messages · Page 16 of 1
Okay... So I did some more testing on a server with a relatively low ping (like 40 ms)... and... I can connect, it just takes several tries to do it. And the same happens with a minimalist configuration.
So at this point, I'm thinking DTLS.
severe packetloss? using wireless internet?
No packet loss, at least as far as I'm aware of, and not on wireless in any way, shape or form.
try doing a lot of pings?
What am I looking for?
How many would I need to launch before you'd expect I see it happen?
depends on how stable your internet connection is
i have a friend that's constantly pinging like 10 IPs because his wireless internet is trash
but, i'd do like 100 and try to connect to a server while it is running
... a 100?
yes?
Let's clarify: you mean, 100 instances of ping?
:/ i mean send a packet out like every 4 seconds and listen for response 100 times over
What ping sends a packet every 4 seconds instead of 1 by default?
If you need me to do a specific test, explain what I need to do. Provide specific commands if the command you need me to run isn't default.
/n 100
the rate doesn't matter, just the number
So... you basically want me to run ping /n 100 host.
???
My connection is rock-solid.
did it run and get 0% packetloss?
To Google? Yeah.
I thought you wanted me to ping the game host I'm trying to connect to.
if you want to but most IPs don't respond to pings
That particular one does.
oh? better then
(and the results are about in-line with what the game says - about 35~45 ms)
no loss?
1 packet lost out of 100.
👀 bad
Nope.
that 1 packet could be trying to connect 🤣
Bullshit, Satisfactory sends out 30 at a time.
And keeps spamming them until it gets a connection.
seems unlikely but i haven't sniffed it
The problem isn't that it doesn't get through,
the problem is that the game is ignoring the replies.
your sniffer detects them coming in?
I have them in my log.
🤔 are you using a VPN or proxy?
No, and no.
This is my game log (client).
(I didn't want to paste them, since Discord would hate me, and I wanted to show you the scale)
Server-side, it would look something like this:
[2022.07.28-19.11.16:612][504]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.16:612][504]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.16:612][504]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.16:612][504]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.16:612][504]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.24:602][981]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.24:602][981]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.24:602][981]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.24:602][981]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
[2022.07.28-19.11.24:603][981]LogDTLSHandler: Warning: Ignoring handshake packet received after completion.
👀 that looks like a bug
if it sent 30, only 1 would complete and respond, the rest or ignoring
It kinda looks like the server accepts the handshake, sends it back,
the client receives it... and ignores it for some reason.
Maybe the timestamp doesn't match. Maybe the packet has the wrong sequence number. I don't know.
But at that point, a state desynchronisation is introduced.
The server thinks that the handshake is complete, and ignores further attempts.
The client thinks that the handshake is still going on, because it ignored the completion packet.
And neither will re-evaluate their situation.
when was the last time you time sync'd?
At 5:20 AM, which was about 4 hours ago.
maybe try syncing time then joining? see if that results in instant connection?
Nada.
I don't understand the question.
player server or dedicated server
I don't have anyone I can connect with using mediated sessions.
i can try running it if you need someone to test against
I'm not sure I'm comfortable sharing my main STEAM account info like that.
(by which I mean I'm really not)
you'd join via GUID...i have the game on EGS
Then I'm not sure I'm comfortable sharing my EGS account ID. ;]
Besides, the mediated sessions are mediated.
(and also, I'm trying to debug a DS issue)
it might be a DS specific issue
I think it is.
Hi there, I set up my dedicated server on docker. My issue is the following: when a player joins, the server game is un-paused and directly the CPU usage bumps in the sky and its temperature too. I'm reaching 90°C+ instantly. Is there any settings or tweaks I can try ?
Server config:
- AMD Ryzen 5 3400G 4,2 GHz
- RAM 3000MHz DDR4 32GO
I clean the server (dust) every month since it's running 24/7, but that's all
blow out the case? 🤣
Not really, no - since you're pausing the game, the first time someone joins (after a restart), it will load the save game. That happens at full tilt, so that it takes the least amount of time. During that time, the server essentially rebuilds the game world and restores the entire state. You should also be seeing a RAM spike.
Still, with that CPU, this should happen very quickly (to the tune of maybe 10-20 seconds), and it should go back to "mostly idle" afterwards.
Satisfactory DS isn't particularly hard on CPU so if it's hitting that hot that fast, it's something physically wrong with the machine like no thermal interface material, dead CPU fan, or heatsink full of compacted dust
This. ^^^
If it's spiking to 90+ degrees within 20 seconds, I would first look into the hardware cooling solution, as that's not normal.
Mayme I'm missing something. But theses usage seem insane and don't reduce after any kind of delay
50C idle is bad
my server isn't idle at all, having 23 docker running 24/7
my 5800X runs 55C 100% load
What sort of cooling?
Because mine goes between 45 and 55 at idle (well, Windows 10 "idle" with Chrome open).
Water cooling + fan
To clarify, this is for @calm jacinth
Mugen 5 rev C w/ 65w Eco (87w TDP)
oh boy, pump died?
How in the name of fuck do you get 55 °C at 100% load on a Ryzen with a Mugen 5?
because eco mode 😉
Do you live in the Arctic? xD
it averages about 4.2~4.3 GHz
yup
So, you're not a good comparison.
well, 4.7 GHz doesn't really make sense at double the power consumption 🤣
Does it need the Ryzen Master running in the BG/starting with the OS, or is it a flag it sets somewhere?
Eco Mode? BIOS
if the blue line is the spot temp then it's probably fine
Ryzen will pull back on the clocks when it hits one of the limits
Still, it'll run 100% until it finishes loading the game (and possibly - loading the player in). Once that's done, it should go back down.
yeah
but if that 90C is the package temp, check the cooling system, make sure it's working
Actually, I'm sporting a Noctua NH-D15, and it does go up to those sorts of temps at 100% load.
From what I've seen, modern CPUs just kinda do this nowadays -- they run up to the temp limit, and then throttle. How good your cooler is, dictates how much less they'll throttle. Same with GPUs.
3400G should hit power limit before thermal limit
On 100 percent load on my 5950x is around 65C just stock settings but I’m on a Castle 360EX AIO
Yes, well. I'm not doing liquid cooling, I can't afford good liquid cooling.
I have a noctua NH-D15 also just as backup and temps when I used that one on the 5950x temps were around the same not much big difference really just with and AIO you get a cleaner look right now is on my 3950x the D15
How the hell do you get "not much difference".
Okay, I officially don't get it. Am I just uniformly shit at doing cooler installation, or something?
No, I mean with the NH-D15 at 100% around 65 °C. How. The. Fuck.
Yeah I can send a pic of it to prove it when I get home but… but on full load the 3950x and 5950x run pretty cool compare on light loads when is trying to boost the single core performance as fast as it can then temps get around 77c max
That's insane. From the first run, and after the last time I reapplied the TC, I've been running around 87-88 °C at 100% load, 3900X.
Ryzen on light load pumps a lot of voltage to the fastest cores to get the highes frequency it can around 1.45-1.50v when at 100 percent load it only pumps 1.1-1.2v to all cores and temps are lower then at lighter loads
It also has to do a lot with case airflow
Like you D15 could be sucking up a lot of hot air from the GPU, Motherboards etc or you probably have a hot air pocket and the hot air just keeps circulating around the case causing temps to be high
Yeah, running caseless.
Seems that opinions on the 'net are still split between "bruh you need a case" and "meh go for it it's fine".
Well if your running caseless you probably get a lot of dust built up unless you clean it every two months or so… but then room temp also play a big role..
I clean it, and the place isn't that dusty. Room temp is avg 20~25 °C.
Yeah I mean is hard to know really the temps you should get as there is a lot of variables likes how is the motherboard mounted if you have like a nice metal caseless test bench then you won’t see much flex but if you have it on a cheap plastic test bench then your getting some flex around the cooler causing the pressure to be uneven… what I would do is take the cooler off and check how the spread of the TC is if you see it pretty thin your getting good contact with the IHS but if you see it kinda thick or thin on one side and thick in other areas your probably not getting good contact increasing temps
Good example on the right they have good contact in the middle but the edges have less contact with the cooler
Ah, well. I suppose I'll be taking it apart this coming weekend. Joy.
I'll keep it in mind. We'll see.
my bad, it's sitting at 64C and 4.18 GHz
thats with BOINC eating 100% CPU and Maneater eating 100% GPU
using decade old Arctic Silver 5 too 🤣
Good morning
Really need help if anyone has it, I have a world which was created on a dedicated server that I am trying to move to a new machine but the session ID is "hidden"? How can I find the session name so I can load the world save?
Is there a way with a save editor?
@blazing halo I can confirm that not sorting the inventory whilst handling radioactive material worked. Thank you for the tip there.
Typical cpu and memory usage for my server.
Tick rate whilst I'm logged into the game is 11 currently.
Looks like it's been 15 days since the last reboot.
Save file size has now topped 17MB.
Is there any way of determining the current UObject count?
I found my missing body, killed her and got my 29 power shards and boombox back w00t.
I don't think so (might be a command for it), but the game will crash on load if you ever exceed it.
Hello, does anyone know how to make the night in the game darker?
Turn down gamma?
I don’t know if there’s a mod for it (which you can’t do on dedicated server anyway), but there’s not really much you can do about it.
I'm past the Uobject crash limit and have ini file edits in place to allow me to exceed it.
What I'd like to do is keep an eye on it.
The only thing I've seen in this area so far is an object count when loading the save into SCIM.
One object can have more than one UObject.
I think it depends per "object type" and whatnot.
I know
iirc no good way to check it 😦
No worries, someone else might :0)
A few screenies. I've decommissioned all the production in the tower (Nuclear Plaza), it's now empty. In the process of modularising nuclear waste handling and built a new concrete and ingot factory finally tidying up the space in front of the tower.
View from the roof of Nuclear Plaza towards Grass Fields. I can see my house from up here.
Yeah, I felt it. xD
If you look to the lil' island on the right there, you can see me there probably. xD
Those train stations are damn impressive, too. Needs decorations tho'.
Did you jump in?
Thank you. I'm trying to keep the object count down, I'm even stripping unnecessary objects from the factories as I visit them.
is the limit only on miltiplayer or singleplayer too?
It's an overall object limit for the Unreal 4 Engine.
Applies to all game modes.
My server started crashing a few weeks ago.
Usually when it misbehaves I move the save over to single player, deal with the issue, resave it and move it back.
Single player has less issues.
In this case single player wouldn't even load it.
I noted the UObject error and that was my smoking gun.
I was aware of BLAndrew575 and the fact he is running 4x over the object limit.
But he is playing single player and has 64GB ram.
Never-the-less I figured I should be able to get the game going again.
I made edits to the ini files and in between crashes and edits decomissioned redundant production.
It took at least a week of decomissioning activity and ini file edit trial and error to stabilise the game and the server.
The game went belly up at 16.6mb save file size.
Through decommissioning I got the save file down to 16.2mb and the game stabilised.
I've since been rebuilding and expanding again, the save is over 17mb now and still stable.
I have an extensive list of factories to build including the new Nobelisk factories.
To facilitate that I will need encased uranium cells.
Presently they're in closed loop production in Nuclear facilities A and B at the Islands.
The Swamp will be new modular Nuclear.
That will enable the export of components for other processes.
The Desert is earmarked for new modular Ficsmas.
I operate a model of kai zen, continuous improvement.
Almost every item in the game has its own distinct factory and my logistics network is based on trains.
I got the game for Christmas and in 10 months and 2800 hours game time exceeded the limits of the engine.
This is my first playthrough.
Screenshots of my build were featured on the weekly dev stream a few weeks ago.
If you want to collaborate visit the server (you'll need my ini files).
If it works for you and want to collaborate you're welcome.
Just visiting here to see how I built it so I can build another around the rail spiral north of the Islands.
Example of where collaboration needed.
This steamed copper sheet factory is 50% operational and supersedes two existing factories.
It needs a second water source plumbed in and some cosmetic work.
Just not enough hours in the day...
can someone help me to setup the server ?
i understand these things but this server really difficult
to setup...
https://satisfactory.fandom.com/wiki/Dedicated_servers just follow the instructions
i tried thats where i need help
are there plans to port it to UE5? Will that fix/delay the object limit?
Hello, I don't know if this is the right channel for questions of this nature but I'm gonna ask anyways:
tonight I felt like getting my Satisfactory dedicated Server back running (after a few months) so I ran Steamcmd to update my files and basicly, updated my ubuntu and so on. But when I came to the point where I wanted to actually start the gameserver, I got this error:
"chmod: changing permissions of './Engine/Binaries/Linux/UE4Server-Linux-Shipping': Operation not permitted
./Engine/Binaries/Linux/UE4Server-Linux-Shipping: error while loading shared libraries: libUE4Server-Core-Linux-Shipping.so: cannot open shared object file: No such file or directory"
so i gave my self permission for the file and tried again but came to the same result.
Then I reinstalled the Satisfactory Server as well as Steamcmd but when I attempted to start the server the same error appeared again. I've got no Idea what's causing it and therefore hope, that somebody here can help me . Thanks in advance
@chrome hare can you paste your FactoryServer.sh (or whatever you call the shell script you use to run the server)?
this one?
mine looks like this, where the setting of UE4_PROJECT_ROOT may be a new thing:
#!/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"
"$UE4_PROJECT_ROOT/Engine/Binaries/Linux/UE4Server-Linux-Shipping" FactoryGame "$@"
hmm yeah that looks like mine
no idea 😄
i'd try a fresh install from steamcmd in a new directory
I already tried to give my user sudo rights but didnt help eighter
i'd try a full reinstall first, should be really quick
alrdy working on it
now that i have failed to help with another person's problem i will ask for help with mine: is there any way to automate loading the most recent autosave when the server starts?
Well you need to explain what part you need help in…
Initially Snutt indicated they would not port to UE5.
More recently that stance has softened.
However, I believe the object limit in UE5 may be the same or not much different to UE4.
So that would not solve this problem.
What we have is a game which appears to offer limitless possibilities but is limited by the engine (ish).
I've not built a factory for every item yet and have already exceeded the engine capacity.
That speaks volumes or at least does on game scope.
could this issue b4 causesd because I - in my sheer stupidity - installed the server in the root folder?
i've always thought it must take serious balls to try and make a game like this because it's very hard to know when you start whether you will run into intractable engine-related scaling issues
you mean you installed the game files in /?
yes
yes, that would likely cause problems
I had it in the home folder on my old vps and seemingly just mooved into /
changin vps every jyear cuz 1st year is just 1€/month
i run mine on an old pc under my desk
H O M E C L O U D ™️
but yeah, perms in / aren't amenable to doing normal things and since you had a permissions error that's almost certainly your problem
perfect
does actually make sense, that in order to have a file belong to a user, it has to be in the useres folder ._.
that's actually not necessarily true
but it's a good convention to aim for under most circumstances 🙂
I'm a noob in ubuntu ...
The game idea is sound.
Some of the recipes though require some insane and out of the box thinking to simplify.
I realised very early on that logistics in Satisfactory was the hardest problem to solve.
It's incredibly hard to move items from A to B.
I already had 3200 hours in dyson sphere program when I got Satisfactory.
I brought a number of strategies from my playthrough of that game (64 stars / 64 dyson spheres) to Satisfactory and decided that the best strategy was probably going to be to modularise every component construction.
I've found I've hit an object limit with that strategy but it is a perfectly reasonable approach given the issues I encountered with attempting to build a vertical megafactory.
I gave up at floor 16 and put a roof on it.
It's known as Nuclear Plaza in my build.
ok thanks a lot it's working now
Hello, It's me again, I've got my Server running now but as soon as I unpause the world , the entire consol turn orange and 2 secs later the server crashes... From what i understand reading the crash log it's not managing to load the dirty areas and overloading its entire memory with loading request for those dirty areas before crashing
How much RAM would a dedicated server realistically need if there are two people playing, planning on going late game. Assuming that the server is running on linux and the ram is ddr4 3200?
I'm only asking because the wiki recommends 16gb, but running it on my machine, windows was using no where near that much.
You need something like 12 GiB for the most part, but if you build a lot, you will exceed it. And please don't look at how much the game is using - the server is a separate program and has its own requirements.
@chrome hare In general, you have chmod and chown. chmod changes the object's permissions in terms of what can be done with it by Owner (U, user), Group (G) and Everyone (O, Others). chown changes the owner and group of an object. They both have -R to apply these recursively, but that's usually a bad idea for chmod (you'll want to at least differentiate directories, which need slightly different permissions). Linux also differentiates between x and X, which is useful for the aforementioned.
There's a few gotchas for newbies in terms of permissions. In order to read a file, you need to have r on the file. In order to read a directory (and therefore read the path), you need r on the directory - but in order to traverse that directory, you need x on it as well. If you want to write to a file, you need w on that file. If you want to create a new file in a directory, you need w on that directory.
Restart the server and check again.
There's a reason the server restarts every so often...
You need 12 GiB of RAM to run it. If you're looking at the console and it's ending on a Killed, that's likely your problem. Note, this will not be in the log - only in the active console.
You must have loaded a save (or created a new world) successfully, and then you must configure the server to not pause when empty. As long as it has the correct session (save) name configured, it will then load it on startup and proceed to run the world.
okay, so essentially if i have a server running in a state where people can join and play, and pause when empty is turned off, it will autoload that save when restarted?
It should, yes.
guessing that means there's no way to do it with pause when empty enabled
thanks for the info
No.
It will then load the save when the first player (since restart) connects.
Yes, I'm aware it causes a timeout for the first attempt.
hmmm
i don't think that's the behavior i've been seeing
as in, a join attempt does not appear to cause the server to load a game
It's the behaviour you should be seeing.
Did you mess around with the session name?
i think i set it once and never changed it since
nor have i messed with any other settings as far as i can remember except the network quality setting
i can test whether it's doing what you described and report back, it's possible i'm wrong since that's not a behavior i'd expect
I'm not exactly sure how the session name thing works for the server. Apparently I have budding dyslexia, and I spectacularly failed to correctly name my last session - but was unable to manipulate the server to change it in any way, so I'm stuck with it, it seems.
i'm not even sure what the point of a session name is tbh
So... describe the behaviour you are actually seeing?
server comes up, doesn't automatically load a save, doesn't automatically load a save when somebody tries to connect either
but as i said i wouldn't expect the behavior where the first connection times out and then the world loads
i should confirm that's not happening
The expected behaviour is: player tries to connect, server unpauses, finds it has no world to load the player into, and then proceeds to load the world.
Since saving/loading takes up 100% of the server's attention, the client is left hanging in this situation, and times out.
Subsequent connection attempts (once the server has actually finished loading) should be fine,
with one caveat: if you have "save on player quit" enabled, the timeout will also trigger a game save, which will extend the time you need to wait before trying to rejoin.
I'm really not sure why the server doesn't just load the game on startup and then pause. :/ But that's an executive decision, I guess.
i expect the person(s) working on the dedicated server has bigger bugs on their plate 🥲
Maybe, but in terms of issue vs exposure, this is what would be known as a "low-hanging fruit".
A relatively simple fix (or, it seems that way), that would improve the experience for a lot of people.
sure, but on the other hand it doesn't affect gameplay
however it is definitely quite annoying
when i have to reboot my server and all my factory buddies ping me because i forgot to go in and manually load the save
You want to follow the 90/10 rule. A 10% effort translating into 90% impact. AKA picking low-hanging fruits first.
i'll test that behavior and report back if it's not working at all
I was always running it on a 8 GB Vps and it workedx totaly fine a year ago
The key words being a year ago.
I don't really understand why people have an issue understanding and accepting the words Early Access.
From Wikipedia:
"Early access, also known as alpha access, alpha founding, paid alpha, or game preview, is a funding model in the video game industry by which consumers can purchase and play a game in the various pre-release development cycles, such as pre-alpha, alpha, and/or beta, while the developer is able to use those funds to continue further development on the game."
(emphasis mine)
Hey my friends and I have been having an issue that started sometime around update 5. Whoever is not the host gets bad visuals for items on belts and lifts. On belts, things disappear and reappear and move in weird ways. On lifts, items are often misaligned vertically with the lift, appearing above or below it. When playing on a dedicated server, all of us see this behavior. Does anyone know if there is a way to fix or improve this? Maybe a setting or something?
Some examples of what i am seeing.
Yes, that's expected behaviour.
Do you mind if I re-use this video?
When you say expected, do you mean there is nothing we can do about it?
That is pretty much what I mean, yes.
😦
It's not a bug.
Well,
the fact that those elevators have things appear above/below them... that could stand fixing, yes. ;D
not being able to watch belts work properly takes away some of the joy of factory building
I'm sorry you feel that way. Unfortunately, this is, at its core, the difference between "local" and "remote".
It used to not be a problem back before update 5 times
On your local PC, this information can travel at GB per second, a million times a second.
Over the network - not so much.
That's the thing: if it wasn't a problem, it wouldn't have been changed.
the game barely uses any bandwidth at all
Have you tried it with the map covered with belts, with 4 players at the same time? ;]
Besides, it's not just the bandwidth.
I mentioned two parameters - bandwidth (which is the lesser problem), and update frequency.
the client knows how many items are on a belt segment and how fast the belt moves. it could just evenly spread them and keep them moving. it doesn't need to be some perfect match of what the server is doing, just an approximation that is not so lousy.
That's the thing: if you want to be able to pick the items up from the belts, it must.
Otherwise you will have desynchronisation between clients.
it is already desynced, so i dont know what you are getting at
So, you mean that if I pick up an item, you can pick up the same item and nothing bad happens?
two people looking at the same belt dont see the same thing
What he is trying to say that is not as easy as you think to fix something like this if you know how to fix it why not contact the devs with your fix
if you pick up an item, it removes one from the belt. what does that have to do with any of this?
The fact that other players must be made aware that the item is no longer available. If two players pick the same item up, it's a conflict.
Right now the client does a poor job of simulating items on a belt. it could do a better job. Picking up items is a separate thing. two players cannot pick up the same item because one will reach the server first and win and the other will either get nothing or a different item from the belt. That is unrelated to the visual simulation of items moving
It is related.
Mainly through the fact that the items aren't updated all the time - and this is why you see a simulation.
Just because the server knows an item has been picked up, doesn't mean all of the clients know this. And since one update theoretically takes 15ms, all it takes is a lag of > 15ms (which is tiny) to cause desynchronisation.
Right now the simulation is bad. No client sees what is really happening on the server. No one sees the items where they really are. Changing where they appear is not going to affect that
Sure, why not. I suppose the simulation could be a little better.
It's still going to be fake, though.
i dont care if it is fake. i just want it to look like items are moving along belts rather than disappearing and reappearing
If all you want is the visuals fixed, I suppose it could be done.
a simple fix just play it on single player and is fixed if it bother you that much but the point is that it is what it is and there is no fix for it just ignore it
I mean, it'll make it much more problematic in terms of picking items up, but I suppose there's already something in place (otherwise there'd be conflicts). So I guess it can be shifted from "ugly belts" to "items disappearing from inventories" and "rates not matching what belts show" (which, I suppose, is already a thing somewhat).
The problem is, I'm not sure how precise the accounting is on item instance IDs.
I mean, this is all set up so that it can account for billions of items on hundreds of thousands of kilometers of belts, for 4 players at the same time.
But then, it's pretty clear that there's a LOD calculation of some sort on everything too... so I don't know.
i somewhat remember some explanation of belts from the devs at some point when they were optimizing them. each belt segment is treated as a single entity, and that entity has data about how many items are on it and what type they are. it simulates moving them and handing them off to the next segment. But for some reason, the host simulation looks nice and smooth while client simulations don't. That is the part that makes no sense, they should be getting the data from the server and simulating the same thing with that data.
Yeah, except the simulation is more complex than you think.
Remember, it can't be just a single item "type" on a belt.
You can send many different items on a belt, and you need to be able to expect them to come up in the order they were sent.
it doesnt matter. data is data. an array of items doesnt have to have the same item in every array position
No, it doesn't,
but then it's not just 2 parameters any more.
It's not type + length, but the whole array.
i think they said they simplify it when belts arent mixed
I suppose that makes sense.
but either way, data gets sent, the client should be able to simualte the belt using the data it receives
Sure.
and not do weird things like items floating outside lifts or disappearing
And the fact that they're disappearing has to do with getting updates. Like: that item is no longer on the belt.
But I do agree that the simulation could be better.
Though you will notice, this gets worse the faster the belts you're using are.
yeah it seems time based so the faster speeds spread out the issues
Indeed.
And it goes back to the timing I mentioned.
The game sends out updates every 1/30th of a second by default.
If i build a new belt and stand there watching it (as a client), it will simulate perfectly. it is only after i move away then return that it looks broken. that tells me that it can work but something is busted somewhere
That... is something to ponder, yes.
(although I can't be sure how much of that is actually correct, because I don't really pay much attention to it right away)
when i started a new save for update 6, my entire starter factory simulated beautifully, but once i left the area to get a hard drive then return, all the belts were messed up
You could mention this in a QA posting. If this is actually correct, it could be pointing to the culprit.
when i made that video i shared, i built that lift loop and watched it run for a few minutes with no issue. i walked about 200m away then came back and it looked like it does in the video
Can you make one that mixes items?
I'm curious to see if the same thing happens when they're mixed.
yeah i could set that up
Because it could be that it's a difference between local-driven simulation and server-driven updates.
The fact that these items appear and disappear like that would indicate that the item accounting is absolute,
that is to say: every instance of an item is remembered and reproduced, no matter the situation.
I don't have hard evidence on that, but it seems to point that way.
if you look at my video, you will see there is a spot where multiple coal are stacked on top of each other stationary, releasing one every so often. not sure what that implies
The same thing.
In theory, the "next" item should be "sitting" there before it gets sent out, so it would immediately be visible in mixed-item belts.
(especially if the items you send are visually distinct)
It would also be good if you could film both cases - when it works, and when it's broken.
(just as a sort of "proof" that it is possible to do correctly)
What I expect is that as soon as the items sent are mixed, it should immediately start doing the "glitchy" routine regardless of distance.
(this would indicate that there is some sort of "simplification" that gets done by the client at first)
It would also help if you had a second player confirm what you're seeing right then and there,
so we would know if it's a local thing or not.
Because chances are, things are "glitchy" for everyone except the player that sets it up, at first (and this is also what I would expect, due to the networking nature of the thing).
Also, @merry bolt I assume you're the host in a listenserver/session play? As in, MP and not DS. That probably makes a difference, since there is host simulation to consider (which is not available on a DS).
If that's the case, it would limit the effectiveness of this "fix" to MP play only, and only for the host.
have been in both scenarios, but current on a DS
Could I trouble you to set the tickrate lower (say, to 15?) on the server and repeat the test once you have the results from this one? I'm curious.
Alternatively, you could load the server machine down to cause the tickrate to drop unevenly.
currently trying to test mixed lifts, but struggling to get enough items to fill the loops. one thing i noticed is that when there is a gap in the line, it fixes the broken lifts as the gap passes through
Another point towards absolute accounting.
I wouldn't be surprised if they used some sort of RLE for the belts and elevators, actually.
(RLE = run-length encoding)
is there a shortcut to remove one item from a stack quickly?
Not as far as I'm aware. You can transfer stacks, split them (shift), and transfer all of type (ctrl).
yeah those are the ones i know
The easiest way would probably be to make 3 producers that produce at the same rate.
oh i dont even have to move far away i can just turn around for 10 seconds then turn back and it is busted, then as soon as a gap comes through it fixes itself
not seeing any difference between mixed vs not mixed lines
So it actually might be a difference between local simulation and server simulation.
Or it might be a difference in the feed to the visuals.
Turning around invokes culling (i.e. hiding what you're not looking at from view), and I have a feeling that it's not limited to GPU frames.
Same thing happens with corpses, and I see no reason to consider these unrelated.
Here you can see it working, then break when i turn away, then fix itself as a gap arrives at each lift.
the horizontal belt at the top never fixes itself though. it works at first, but once it breaks it stays broken (items disappearing)
Although it did eventually fix itself later after i stopped recording once i stared long enough that the gap came around a couple more times
well not really fix, but the point where things appear/disappear shifts to a different spot
i am thinking it is a visual only bug
which seems like it should be fixable
anyway, knowing these issues seem to be universal and not just something weird my group is seeing is helpful information.
Hiiii! Are there any dedicated servers to explore?
#looking-for-group-old Hatake-san. ;]
thank you
is there a optimal save time right now we have it set every hour.
Interesting... so the lifts seem to be shifted instead, and it corresponds to the belts. Could it be a simple "buffer" issue?
(by which I mean: the conveyor - be it belt or lift - "buffers" the items before it sends them; for lifts, it happens above/below them; for belts, these items get "stuck" at the start of the belt... and however much it buffers, that's the amount that's "sticking out" and also the amount that's later missing...)
is there a auto kick option for servers?
There're no access controls apart from the passwords.
For future reference, if you want to blank something out - just use a solid-color bar. I can see parts of that address coming through. There's no need for frills - just blank it out.
It's funny how people always blank out their IP, but fail to blank out their STEAM and EPIC user IDs... xD
It depends on a number of things. Where is your server located? Where are you connecting from?
How would dedicated servers work bc everyone will have the same research
It's co-op. Not competitive
how do I setup a dedicated server? I did it a while ago on experimental but I want to make a new one on early access
I do have port forwarding setup and everything already
If you look to the channel's topic, there's the wiki link with instructions. Since you've done this before, it should be familiar.
yeah thanks I just found it and am reading it now
I would like to mention that, while it is possible to set it up via STEAM or Epic, doing it through commandline is more versatile and a good thing to learn overall.
Unless you don't really care about the how, and just want a working server. In that case... I guess it doesn't matter?
should I go as far into the directory as I can go?
in command prompt for setting it up
Do you intend to run multiple servers at the same time?
I really should organize my D drive but I have some things directly in it
I ask because there's a server already running in the background in that screenshot.
Hm... what exactly are you trying to set up? If you have it installed from EPIC, and it's working?
well I installed it on epic and I can run it
I never have run a server on early access
how can I upload a save file to it?
As long as the server's executable is started under the same user,
it will find the saves from every other instance of the server running as that user.
(since they're kept in the user's home directory)
In order to load a different save into the server, you must launch the game and connect to the server (without joining the game) in the Server Manager.
From there, as long as you log in as admin (the game offers two password options - one for administration, and one for regular users as a means of access control), you will have access to the various panels. One of them is for loading saves.
my Debian 11 server is located in a datacenter
colo
Do you have any idea what this could be? @blazing halo
What does the log say?
@sage hound
FUnixPlatformMemory::OnOutOfMemory(unsigned long long, unsigned int)
It's running out of RAM.
How much RAM does your server have available for the executable?
Ah, here we go: Physical RAM available (not considering process quota): 8 GB
That's not enough.
You'll want to allocate at least 12.
thanks it work now
It depends if you're running it inside a virtual machine or not, and on what sort of virtualisation it is if you are.
In a physical setup - what you see is what you get. But in a virtual one, it most usually isn't.
And especially if you're renting a server somewhere, it's most likely (overwhelmingly likely, in fact) to be virtual.
how much memory does this game need?? Ran out of memory allocating 313733120 bytes
Need around 12gb minimum
just realized that was talked about. yea apparently the host isnt giving enough. AvailablePhysical 3614052352
AvailableVirtual 273920000
UsedPhysical 12447764480
PeakUsedPhysical 12608974848
UsedVirtual 13209726976
PeakUsedVirtual 13545353216
If those are bytes that’s a little less than 4GB available
yea idk why they arent giving more lol.. i have been so mad. server keeps crashing every hour on the auto save time. and the file says this my butt keeps logging in thinking it will save this time. building a bunch and nope... buddy doesnt understand swears its something we changed last night with connection times but its not
Yeah I don’t pay for hosting so contact them I self host so I don’t have those issues
i would if i had the equipment.
my pc is good and would hold it just dont want to do it that way.
Ohh yeah I remember that
does this mean anything to anyone Warning: EnumRedirect for enum ECachedNatType maps 'ECachedNATType::CNT_MAX' to invalid value 'ECachedNatType::CNT_MAX'!
I get the general idea behind the message, but can't tell you anything about the context (or, what the results might be). It likely means that NAT detection won't work... but judging by the value reported, it probably would've said either "strict" or "TBD" anyway.
Did it crash on that? If not, ignore.
every crash for last couple hours says same. i thought it was a memory issue but down where it says what i posted it says error so maybe that?
Nope - this Error: Ran out of memory allocating 426401792 bytes with alignment 0 is explicit enough.
It runs out of RAM.
thats what i thought it was. the host says they do not limit the ram on these servers tho
Well, the crash info says they do.
thats what i thought to. just for a buddy's piece of mind we changed the connection time out crap none of those changes would have effected this right. he doesnt believe me lol
Either that, or their configuration is in some way unsupported...
I mean, the log clearly states that there's 128 GB of RAM reported by the system.
idk. we have had the server for roughly a month and been fine. well couple issues here and there but after i changed the time out last night it ran smooth all night. log in today and its not saving every hour and crashes when it tries to save. resetting all the work done
[2022.10.31-18.00.35:512][ 0]LogMemory: Physical Memory: 115536.14 MB used, 15378.52 MB free, 130914.66 MB total
[2022.10.31-18.00.35:512][ 0]LogMemory: Virtual Memory: 134204840.00 MB used, 12891.86 MB free, 134217728.00 MB total
These two are also kind of interesting...
Those values are quite silly.
How in the name of fuck are they using 127.987 TB of address space?
idk what that is lmao..
Still, if you're running a big save, then the clue may be in the middle values: physical 15378.52 MB free (= 15.01 GB), virtual 12891.86 MB free (= 12.59 GB).
The virtual address space is essentially the highest number the CPU can address. In this particular instance, it's set to 128 TB - or 131 072 GB. In this space, everything addressable must be kept, including things like VRAM and virtual devices.
It's the reason 32-bit machines with 4 GB of RAM can't actually use all 4 GB of RAM. The maximum value a 32-bit system can address is 4GB, and device memory must also live in that virtual space... so if your GPU had 512 MB of video RAM, that got subtracted from your virtual... meaning you could only use up to 3.5 GB of RAM regardless of how much you actually had.
64 bits was suppose to solve this problem for pretty much "forever".
So I ask:
they have 128 GB of RAM, and 128 TB of address space.
HOW IS IT THAT THEIR FREE VIRTUAL MEMORY IS ACTUALLY LOWER THAN THEIR FREE PHYSICAL MEMORY?!
What they fuck do they have in there, 5000 mining GPUs with 24 GB VRAM each?
It could be possible I have a 3090 and it has 24GB vram lol but yeah there is something going on with their set up
Not enough PCIe lanes in the world, man.
EVEN IF we assume each is using a grand total of 1.
Well for mining you don’t need pci-e lanes they connect via usb
... really?
Yeah they have like PCI-E 16 slot with like 20 usb out stuff like that and there is also mining mother board with a bunch of PCIe 1x slots
Saw one, yeah... there was like 8 or 10 of those.
Not 300. xD
Ah, no. I retract my previous statements.
This seems to only get logged on Windows, and it takes the page file into account.
[2022.10.29-04.31.15:705][ 0]LogMemory: Physical Memory: 16032.98 MB used, 16649.08 MB free, 32682.05 MB total
[2022.10.29-04.31.15:705][ 0]LogMemory: Virtual Memory: 134188920.00 MB used, 28807.89 MB free, 134217728.00 MB total
This is on mine. The free virtual is still much higher, but then my paging file is set to a multiple of RAM.
Request that they increase the Page File's size.
Not sure how it could be coming up short, but it seems that it is.
Anyone with a better understanding here?
Because I'm a bit confused.
Is there any sort of health check tool for the dedicated servers? I see a reference to someone developing one about a year ago, but wasn't able to find any other info.
le sigh DS crashed again
failed to save dmp, and log is full of warnings associated with holo conveyors
whoa, lost probably 5 minutes of work
nope, more than that, a lot more
back to the last autosave
yeah, had obvious problems with conveyor placement for a while
like 3 lengths of conveyors were just holograms, never actually built
Restart the server, should fix it.
Hi! Can I use config file instead of command line options?
Instead of what?
Some things you can probably put in the commandline
From looking at the wiki not possible to put the ports in the config file.
You'll need to make a shortcut that contains the port arguments in that case. Or make an alias if you're used to those and use the commandline.
there is very little information in wiki about install and using DS such as service. NSSM, for example, has info about launch NSSM GUI for installation service, but what are the parameters?
Cause, I want to use SC CREATE command in CMD
I created service user, add Local Group Policy permissions at server. Used SC command and created the Service for DS, but service cannot start. NTFS permissions accept Full rights to the Epic folder
After manual launch FActoryServer.exe -log -unattended I see
next issue in screen log:
[2022.11.01-09.30.34:826][ 0]LogGenericPlatformMisc: Error: SetEnvironmentVar not implemented for this platform: SteamAppId = <digits>
Next issue
May be I need to install the game client to winserver yet?
And what shall I do for using MODs on DS?
Non-issue
Non-issue
Mods are not compatible with the DS.
In general, I have not tried putting the ports into the config files. In other UE games, they usually end up there... but not in Satisfactory, it seems.
Or, rather... it depends on the game, really.
In ARK, the RCON port at least ends up in the config...
Okay, so cursory testing confirms that -Port=X is equivalent to placing
[URL]
Port=X
into Engine.ini
Searching the original game files (update 5) reveals you can also set the Beacon port (-BeaconPort=Y) by adding
[/Script/OnlineSubsystemUtils.OnlineBeaconHost]
ListenPort=Y
into Engine.ini
Unfortunately, I can't find any mention of 15777.
What's the log output?
If any?
Otherwise, what does Windows' log say?
Quick questions for anyone that has bought one of those 10$ a month dedicated servers ; do they simply give you a preconfigured server's IP, to which you can they connect and "Load" your save in, or is it more of a "we give you an ssh key and you fend for yourself to install everything that's needed" type of deal ?
Well, there you go... Added the info on .ini files to the Wiki.
Are you talking about the various game hosts like IndifferentBroccoli and *spits* Nitrado?
First link on good ol' Google is "g-portal", I assume it's the same ilk ?
I thought of grabbing an AWS instance and doing the whole setup myself but the bill would probably be more than $10
Note that G-Portal, as far as I'm aware, is either incompetent or actively malicious, because it's been a year and they still haven't gotten the clue that the user's home directory needs to be mapped to FTP access, because that is where the save files live. (so you can upload your save there, but you can't download it afterwards)
AWS specifically is a no-starter, from what I know - attempting to connect there doesn't seem to want to work.
I'm a zoomer and I never use FTP, can't you just cd/mv your shit around with this protocol ?
Not sure what being a zoomer has to do with anything, but if you're talking SSH - as far as I know, they don't give SSH access at all. Because why?
And yes, you can, but you have a "home" directory you can't exit.
And the user's home directory is usually nowhere near the server's install path (which is typically the home directory they lock you into, as far as I can tell).
Hilarious
Not sure what they're playing at, here, but I do not recommend G-Portal.
I also do not recommend Nitrado, and in fact advise against giving them any money.
I already do IT shit as a day job so I just wanted to relax and buy some decently priced key-in-hand solution : 'p
Guess that ain't happening. Thanks for the feedback man, much appreciated.
I don't really know what to tell you... I've seen different hosts discussed here,
but there's always a catch.
I don't remember which one, but one that isn't scammy seems to not get the hint that the hardware requirements have changed,
and offers an 8 GB host by default.
That was fine for U5, but U6 requires 12.
Cue lots of people coming in asking why they get on-connect timeouts.
(answer: because the game crashes OOM as soon as they try to load a save/start a new game :D)
Which one would that be ? If it isn't scammy I guess I can try shooting a dm over and tell em the memory requirements have changed
Yeah, that's the thing - I don't recall. 😂
It may've been IB, but I don't really know.
Hey @loud minnow do you recall?
@mental locust you, maybe?
I self-host, so I honestly don't care. :D
(especially since people come in here to get tech support for a service they're paying real money for, which rubs me all kinds of wrong ways)
the service did not respond to the request in a timely manner
Have you tried a different service handler? Like srvany, if that's still a thing? Interesting...
I do self-host an eSXI on a small micro server in my living room, which wouldn't have the hardware requirements for dedicated servs.
All of it's RAM and cores are already taken by vms, too.
I'm using only manual configuration
Not use any managers
I'm asking because it's an error reported by the Service Manager if the service in question doesn't respond to it in a timely manner... which is to say - it says "Service, start!" and service says "..." when it should say something.
Windows uses a Service Manager. The program you're using to run the game server as a service hooks into this manager.
Wait a second...
How did you configure Satisfactory to run as a Windows service?
Did you just straight-up hook the DS executable into the Service Manager?!
Dude, that ain't gonna work.
Windows service executables are special, in that they must respond to a number of queries and commands issued by the Service Manager.
That's why you need to use a handler/wrapper, something like nssm or srvany.
Any plain ol' executable doesn't have that capability built into it.
It doesn't respond in a timely fashion, because it doesn't even know it's being queried (much less how to respond).
Exactly. This information not placed in wiki. But at first of all I try to start with nssm. This pair don't want to start anything
Wiki not say what params we need to use when we create Windows service. Probably its not "-log -unantended" and not factoryserver.exe... I don't know at all
> this information not placed in wiki
Have you tried https://satisfactory.fandom.com/wiki/Dedicated_servers/Running_as_a_Service#Windows ?
Yes, I used these instructions
How did you determine that "this pair don't want to start anything"?
What happened, vs what did you expect to happen?
In 30 minutes I can answer this question)
So... it's been 30 minutes and then some.
Well, if ya need me - ping me.
I'm in some pretty sh*t now.
That's me on that train or rather my body.
The game crashed as I was boarding the train.
On reload I respawned back at the hub.
I can't delete the train, can't access my inventory and from experience unlikely I'll ever spawn back in that body.
Ideas?
Satisfactory.valebridge.net
Hello I am looking for info on how to increase the tickrate of my dedication server
Why do you want to increase it? - it should be 30 I believe, if it's below then you need a better CPU ideally :)
@mint junco I'm in, let me see..
@blazing halo my users have a save that uses 12 gigs of ram
The tick rate is more a "ceiling" than anything else, like FPS in graphics. You can't make it go faster by increasing the limit. Or, well, you can - if you set a higher cap, and the hardware can support it, it will go faster... but I'm not so sure that you want to do that. Going faster means it'll have less time per tick to do things (1 second is still 1 second, and 1 Hz is still 1 Hz no matter how you flip it around).
dell rack server, windows virtual server manage with steam Intel Xeon 2.40Ghz 32gb ram
E5620... yeah, that's not going to cut it at all. Not only is it nearly 13 years old, its top speed is only 2.66 GHz.
If your tick rate dips as low as 3 TPS, you've got problems.
Especially if it stays there.
A target rate of 30 TPS means that the game has, on average 33.4 ms to do everything, in order to do it 30 times every second.
A TPS of 3 means that it takes the game 333.4 ms to do what it should do in 33.4 ms (or, 1/10th of that time).
I can't change it? the server does not use all its power it is barely running at 40%
The scheduler runs on 1 core, really.
4 core but I can add more on the virtual machine
This game is much more dependent on single-thread performance than it is on multi-threaded performance (which is more the thing those Xeon CPUs are actually very good at).
They're great for interop and multithreaded applications... but Hz for Hz, they're below par in today's standards.
Like I said - the game needs single-thread performance more than multithreading. Adding more cores will do absolutely nothing.
Your only option is to replace the entire thing, at this point.
There is no CPU that you can use that will offer acceptable performance on that socket. There just isn't.
It's probably great for hosting HTTP, FTP, Samba... stuff like that. But for game servers, you'll need something much more modern.
ok
It doesn't need to be top of the line, mind. A used Core i7 will work nicely.
(unless you want to host big factories, or many players - in which case I'd advise you to go contemporary - so i9-11k/12k or Ryzen 7/9)
(or Threadripper, if you need more oomph in general)
SCIM?
I doubt you'll be able to recover your character (or the inventory contents), but you migh be able to load the save up and delete the train.
@frosty gale need a host?
no thanks I'll manage
At 3 TPS... I'm pressing [X] to doubt as hard as it'll go. ;>
The guy has some serious hardware at his disposal, and he's offering it for free. Work out an agreement so that both sides know exactly what they're allowed and/or required to do. He seems eager to help out... so why not let him?
to each their own
True.
@novel yarrow it was very a good idea but it appears that trains are nuke proof :0(
rip
place the nukes on a reactor lol
Radiation didn't kill orphaned player ?
Sadly not.
Ok, so the last thing You can do is load up save on Calculator website, remove orphaned player, save, upload to the server.
Bearing in mind I was wearing a hazmat suit and had a full stack of iodine filters so that might be a factor.
But I have noted previously that you're immune to radiation when on a train.
In update 5 I used to ride the uranium train and noted the iodine filters were not depleted on the journey.
Cause there is no "offline" mark over there.
Hey, hmm... what happens when you derail a train with a player in it?
They may be nuke-proof, as is most everything else, but they're not derailment-proof.
If I end up editing the save that may solve the problem getting rid of me but was hoping to recover my inventory.
Probably are
Try to ram a second train into it.
You can derail it but it won't throw the player in the train
I've been in one before and plummeted from the viaduct in a collision.
Oh. Well, that kinda blows.
@blazing halo The player is stuck inside and is not marked offline..
If you re-rail the train, does it do anything to the player?
We'll find out, I'll try it
Probably not
Probably not, but it's worth a try.
Wish we had a voice chat
Can SCIM edit inventories of players that're in trains?
You're on Discord.
I think that to get it work (kill orphaned player) he needs to be offline, have You restarted server to kick off connections ? Save before ofc.
It happened after a crash,
so I think it's safe to say the game is kinda stuck with that player.
Restarting might be an idea, but I'm not holding my breath.
@blazing halo if I let him use a thing on my server could it be fixed
@mint junco check out my player's avantar, is if marked as "offline" ?
Avantar?
*avatar
I seriously doubt it. It looks like it's a game logic issue.
Rip
From the looks of it, the game essentially switches modes when the player is in a vehicle (or, more to the point, in a train).
Now it seems that we have proof that it also seems to affect networking bookkeeping as well.
crash
People need to stop using weak hardware
re-railed and the old me is still in there.
How did it rerail
I re-railed it
Oh ok
I'll try a server restart next
It's not that this guy's hardware is weak - it isn't. It's just that the factory is huge. I visited. It was stuttery. :D
I cant connect to my friend's dedicated server. I am able to connect to others
@blazing halo really
hes looked at the troubleshooting channel and checked everything
It's always stuttery until the area has loaded, then I get 30+fps
Dude, the train was jerking around even close-up. xD
Your friend hosting the server inside his home network?
And another friend of mine could connect to his server. So I cant connect to his server, but I can connect to others. And others can connect to him
idk ill ask
Can your friend come over here? It'll be awkward having to go through a middle-man.
True, they've always done that on the dedicated server. Got used to it.
They don't do that on mine... xD
But then, my save looks like an abandoned uranium mine town from the '60s, in comparison... xD
@honest harness
woops
wrong person
mb sorry
@lilac pilot
lol mb
anyways, kao they asked if you are using your home network
Yes
Can other people connect?
yes
another person connected to his server, and I can connect to other dedicated servers
I work, someone outside the network works. Port forwarding is good, firewall is good, and I've confirmed they are seeing the right IP from the address I gave them.
When you say you can't connect, do you mean in the Manager, or into the game proper?
ill get the message specifically one sec
Thomas is the one that cant work
So not "they" but Thomas.
^ here specifically, this confused me
Yeah, I expected a "He" or "Thomas" there.
"the address does not seem to be valid"
Okay, that's not a "can't connect".
Both kao and a third person can connect to his server, I can not
That's a "the game doesn't understand the address, so it won't even try".
The third person is not on my internal network
They are using the same address information
exactly
Then either Thomas is putting it in wrong, or you're using IPv6 exclusively.
I'm not using v6 at all
Domain name or IP?
Can you double check kao if Im copy and pasting something slightly off
DNS address
Public?
Password protected, but Thomas never got the prompt. The third party that could connect did.
No, I meant if the domain was visible/resolvable on the Internet.
...yes
The third party is external
This isnt my first rodeo on setting up a dedicated game server
I ask because it's not immediately obvious, and you could be sharing a config with the third party.
Not something I can read from tea leaves. :)
Third party is completely external to me and has not static routing into my internal network
using the same info I was to connnect
So we're limited to two options: either Thomas is putting the wrong thing in (maybe you provided him with address:port notation, and he's trying to paste the entire thing in?), or his DNS resolution isn't resolving the address.
Try doing a ping.
The other server I joined used an ip
I've seen Thomas's screenshot and the address and port is valid
@pure dragon try pinging the address, see what the system spits out
Cant use that as a valid test when most routers will ignore external IP pings
I had to disable that feature already to test that
also didnt we test that?
Can test DNS resolution at least.
We did
The test succeeded. He is resolving to the correct IP address, and we successfully pingged when I enabled external network ping response.
If it resolves and pings, the only thing we're left with is Thomas putting it in wrong.
Do a screenshare and see exactly what he's doing.
Thomas did you copy and paste or type it in?
copy and paste that line
Try typing it
Copy/paste works fine with Satisfactory.
ill dm you what I copy and pasted
In theory 🙂
In practice as well.
ill type it in manually
While he's doing that, can you tell me what the server log says around the time he's trying to connect?
We're at edge case. Things like "did discord translate this into a character that isnt properly being copied and pasted" since I've already confirmed that it looks correct
Killer
lmao
So now that's in theory 😎
@blazing halo Thank you for the assistance as well
Considering that it hasn't failed me yet, I'm still more likely to consider user error first.
Sure thing.
For instance, I don't know if SF trims whitespace, so copying some.domain.name might not work.
The reason I reiterated it wasnt a problem was I saw the screenshots of what they were putting in and I was the one who gave them the information. I dont believe there was whitespace, but there are any number of weird things that could be happening with character translation that I needed to eliminate.
It's a domain name, there should be no issues with character translation.
There could have been an invisible newline character as well
That's more what I mean
Not unless you're using a third-party Discord client.
(otherwise I don't believe it's possible to send such characters easily)
well I know that I typed in the verbatim thing that I copy and pasted
But yeah: if in doubt, copy & paste into a text file first, to be sure, and then copy & paste that.
I've seen the weird things that come out of a copy and paste between messaging clients. I dont make assumptions anymore if it's brought to my attention.
Discord, from my experience on Windows and macOS, is very civil about these things.
I had an idea to load an earlier save.
That didn't work.
I also loaded the current save into SCIM.
It would appear that players in trains are not shown on the map.
What about the trains?
Located the train and no player icon is shown.
Well, I'd expect you won't be getting the inventory back, then.
You should be able to delete the train, at least?
Or it could be possible that once the train is deleted, the player will be forced to reappear.
It has to do something with the entry, I guess.
You can't delete a train with a player in it.
Either that, or it'll get removed and that'll be that.
Even through SCIM?
Ah... good point
Even despite the fact that it should be an illegal configuration?
(I mean, how can a player be in the train, if the save is being loaded offline?)
In SCIM I found you at paradise islands. So the player locations are known.
(yes, I know what I said; in this instance, "online" and "offline" takes the meaning it has with mounting images; "online" meaning inside a running server, and "offline" - in a tool, or anything that is not a running game)
Which makes sense, because you can't be "inside a train" while offline - that's not a valid configuration.
It's theoretically an impossible situation.
Keep the save file, just so you can provide it if asked by a dev.
If it can't be helpful in tracking down the cause of the issue, it might be helpful for testing fixes.
Export a save from SCIM and see if this "ghost player" still exists in the exported save.
It might fix it by simply reconstructing the save according to the rules it understands.
Good idea.
I've resaved the file from SCIM and am loading it now.
I noted in SCIM there is a player page.
I am shown as host, which is interesting as it's on a dedicated server.
But the save did originally come from my single player game.
I looked at the guest inventories and none of them is the old me.
Still in the train in the edited and resaved file.
This is a very interesting situation.
So although SCIM doesn't display the player in the train it is still in there and is re-exported with the save data.
I've got a copy of the save should it be of use.
@mint junco how do you set the ip as the game name
I created a public facing DNS entry (A record) as satisfactory.valebridge.net pointing to my public IP address and set the session name as valebridge.net.
Awsome!
How can i wipe the save and password as i need to reset both. I forgot the password and the world is extremely broken (laggy like crzy and keeps crashing me)
press windows + r
then type this in
%LocalAppData%\FactoryGame\Saved\SaveGames
delete server settings.port
^
yo anyone know if the server auto saves? or whats the command to save it etc
yes it auto saves and there's no command to force it
there is a command to change the interval
@blazing halo send private message
Good morning everyone
anybody that can do me some help with the server start script? im getting most likely a network error, and it closes the cmd windows in 10 sec
i can run last oasis and rust, using the same server, so just need directions on the script itself
tysm
now to fix the server
is this accurate or is screen rant full of it
WIll i be able to run the game sevrer on a nuc (8gb ram 8 core i5)
so correct me if i am wrong. the satisfactory dedicated server (downloaded from steacmd) is the stuff to make the server work, each time ill need to load a save file and it will host that
thinking of setting up a paid dedi server but have no clue what im doing would someone have time to talk me through some things how to set it up and upload saves ect please DM me thank you in advance
Hi ! Can someone help me with Satisfactory Server Setup. I Set it up on my local server from pterodactyl. Server is showing ```NOT AUTHENTICATED
@trail kayak no
FG.AutosaveInterval xxx set the xxx to 600 and the server will save every 10 min
server.SaveGame saveName
Is the command for a manual save, the save will be called saveName
No to the nuc or screenrant
bump up the RAM of your NUC to 16 Gig and it'l work. Use Linux and something like AMP to reduce overhead tho
Those are the game client requirements. Look to the Wiki for server requirements.
Server crashed after i placed a conveyor belt
i am trying creat my own server,but always show offline, how can i fix this (have watched the tutorial video and done the actions mentioned)
Basically ,i know nothing about codes, what does this log mean , did it work or not ?
@calm jacinth
That's a client log, isn't it.
?
How can i know if dedicated server is sucessfully started or not
could somebody help me check my server IP ?
47.98.150.148
I just want to know whether its port is available
gimme a sec
really thanks bro
unreachable
oh, goosh
are You able connect locally ?
yeah ,i can connect my local
seems something's wrong with my ECS
How do i check this ECS ip is opend for 15777 or not ?
netstat -an | find "15777" to see if it's listening
it's listening, so now the NAT rules: at least these 15777,15000,7777 - UDP
on Your router
I just check my ECS controller, all 3 ports are permitted
Am I allowed to do UDP port probe on Your IP ?
sure ,go for it ,bro
does my location affect ? cz i am in China
Looks like geo-lock.
means i can never use dedicatedsever to play with my friend ?
BTW, TooMeek could i try to link your server just to check if its any other prbs affect
Actually i dont even konw my dedicatedserver is successfully installed or not
All my servers are behind VPNs. But I may open one for U, let me check...
thanks a lot for your patient heip ,anyway
I need help with setting up the server.
I want it to run on my computer so that my friends can play in our world while I'm doing other things.
I installed it from the steam library, launched it and allowed the ports mentioned in the wiki, but now I'm lost with what to do.
got server up and set up with save @ teir 4 ran fine for first few hours left it running while no one was on joined today and the lag was brutal any help would be appreciated
You should probably talk to Tier 4 about it.
We don't know what to do either. We don't have enough information. When does it happen? Does the server continue working, or does it crash? What hardware is the server running on?
Checklist for running the server inside your LAN, with Internet availability:
- is your WAN IP public?
- is your LAN server machine's IP a static assignment?
- did you forward the correct ports? (at the moment they need to be identical mappings)
It's the PC that I normally use for gaming. I don't think there's a static IP
I used the Powershell command from the wiki, so 7777 15000 and 15777.
You should set the DHCP assignment for that PC as a static assignment, ideally outside the normal DHCP pool.
It is running on my computer and I just started the server, the error occurred when I created the world
I did not understand a single thing of that message. I know what DHCP is, but no idea how any of it works
Did the server continue to run, or did it crash when this happened?
Perhaps buying a server from one of the hosting companies would be more to your taste?
I need every bit of money I can get rn
So no, not an option
Okay. So let's start with determining if your WAN IP is publicly reachable on the Internet.
Please log into your router's admin panel and find the WAN properties page. Note down the IP (it may be called "WAN IP", "Internet IP" or something like that).
Next, go to whatismyip.com (or some other similar service) and note down what it displays there.
If these two values match, your WAN IP is public.
If they don't - it is not, and you may not host public services.
looks like they don't match
I only got IPv6 value from my router
Can I fix/change that somehow?
... that's weird.
Is your ISP exclusive with IPv6?
To be honest, it's much more likely that you've found the wrong page.
Ah, I know what that is.
Your ISP can't allocate you IPv4 because they don't have them, so you're on an internal IPv4-over-IPv6 tunnel.
This can only be changed by your ISP, but since they're using it - it's likely they don't have IPv4 available.
So, no. Your only likely solution is to change ISPs.
In theory, your machine should be reachable over IPv6, but not many ISPs support IPv6.
god dammit, I knew there was a catch with an offer that cheap
I assumed it crashed
Can you not assume but actually know?
Actually, many ISPs no longer offer IPv4 addresses, because the IPv4 pool ran out a few years ago. IPv6 was supposed to have replaced it by now (it was drafted back in 1995, proposed as the new standard in 1998, but only ratified in 2017), but IPv6 adoption is slow.
idk about the situation in Germany but speaking from experience, it's going very slowly
I heard you like CGNAT
- so I put a router over your router, so you can route while you route with a private IP on your private IP.
The funny thing is - they reserved a completely separate pool for the purpose (100.64/10), so they wouldn't need to care about any private pools already set up in people's houses.
yet some of them just ignore this
They do?
Also:
- so we broke all of your services by putting you in it, and never told you about it.
So, did your assumption pan out?
Because if it did, I likely have your answer.
ballpark, what's a dedicated server cost in american money?
im, at tier 4 sorry sto give an idea of size of my save/factory .... not wumbo size
im actually at the same issue as his right now but we had the dedicated server running not long ago and we formatted our server rack forgetting to pull the whole server folder and now that were trying to get it online now all sudden it doesn't wanna work
Is there any way to talk to the servers to get info like status amount of online players etc?
It went offline but it becomes online when I reboot satisfactory
Okay. How much RAM is available for the server?
32 gb
Okay. I'm going to have to see the server's log.
Sure. Ping me.
Im getting ready to start my first server with some friends. they want to add mods and I know that some serverhosting doesnt allow it. does anyone know a serverhosting service that does allow mods?
Mods are currently not compatible with the dedicated server.
If any service providers claim to offer mod support for Satisfactory DS - they're lying. Example: Nitrado. They heavily imply that they support them, give you a quick tutorial on using them in single-player, all the while telling you to buy a server - and when you do, after you've paid, then you see a "Coming soon" in the "Mods" section of the server configuration.
thank you. you are a gentlemen and a scholar!
Happy to help.
Note that I've said that the mods are not compatible, not the server. :-)
It's funny, actually... A central dependency that all mods rely on, currently only exists for the single-player and mp-session game modes (which is essentially single-player "with friends"), because it was conceived and built before the DS came out.
that makes sense. you saved me a ton of potential headaches... im a purist kind of gamer but some of the modes look like they have a great quality of life when building big.
As far as I can see, there is development happening, and fairly regularly as well.
So... it'll eventually happen. Can't say when, but it looks like it's not dead.
Not sure how much of that pertains to the server support though... Would need to dig a little deeper.
For now all I can say that there is some sort of development going on, regularly. So... it's not dead, as I said.
And not in maintenance mode, either. Which bodes well, actually.
Fun fact: I can actually be quite rude! ;) But I usually don't unleash that side of myself on people that genuinely appreciate help and can work for their reward.
(that is to say: work with me, and I won't be an ass to you)
I just got a server from a third party vendor, what - if anything - do I need to use the server to play the game?
Aside from the Satisfactory dedicated server tool
Hey, I have a physical Linux server, running Debian, Dell PowerEdge 2950, 8 core Intel(R) Xeon(R) CPU X5470 @ 3.33GHz w/ 32G of ram, and a swap file with another 32G. My save file 27M. I'm at the end of the game, more or less, working on the final objective. When I load up the save, things get laggy. Seems like the simulation isn't keeping up with the production of the objects etc. Is there anything additional I can give this server instance in order to make it work so I can finish it? I can't put more actual RAM in the box because its maxed out.
Unfortunately, the only thing you can do is upgrade the CPU. And it would have to be a massive upgrade, too - as much as that CPU kicks ass, it's now 14 years old - and it's showing.
And even then, I don't think it would help all that much...
i appreciate your assessment. thanks!
Take a look at your TPS (ticks). If it's dropping below 30, or - even worse - consistently under 30 - that's your issue.
If, however, it's somehow keeping a solid 30... then it might mean your bottleneck is elsewhere, and you could experiment with increasing the tick rate.
(keeping in mind that this would likely affect the server's latency sensitivity, and might introduce even more lag due to shortening the time slices)
(in general, time per slice is 1/tickrate, so at 30 it's 33.3 ms; at 60 it would be 16.7 ms)
so.. new server w/ new CPU then huh?
Check your tick rate first. When you see a lot of stuttering or lagging, check to see if it's dropping.
Also, there are some things that aren't technically "lag", like the belts and elevators with "disappearing" items or things showing out of alignment.
Further, when the game is being saved, the server is non-responsive. By default, this takes place every 5 minutes (configurable by setting).
If the time it takes to save the game exceeds the connection timeout (default: 30 seconds), you will experience disconnects.
That is also not a "lag".
Save time is directly affected by single-thread CPU performance.
It is possible to increase the timeouts. This would have to be done on both the server and all clients that connect to it.
ya, ive felt the save ""lag", the over all game being wierd is different and its on a NVME so idk that this is related to the save activity
ya, this seems like its happening all the time
What it does, is collate all of its RAM state and compress it into a sparse save file.
That is what takes the most time, and by far.
That 27 MB save file likely translates to something like 25 GB of RAM use, would that guess be accurate?
(I would expect it rising up to about 28-29 during load, before settling on about 25)
wow that's amazing that 27MB can eat 25G
That's because the 27 MB holds a compressed and sparse version of your world.
I fully expect it to be using some sort of RLE on top of everything else.
RLE?
Run-Length Encoding
(for example, "aaaaaaaaaaawwwwwwwwwwwwwwwwweeeeeeeeeeeeqeeeeeeeeeeee" could be encoded RLE as "11a17w12e1q12e")
(as you can see, it lends itself best to saving repeating information - very useful for sound data and bitmaps, for one... or perhaps long stretches of foundations in a game, for example)
But each item in the save, and this must be calculated in 3D (X/Y on the map + elevation), must be reconstructed in RAM to become a fully-operational Unreal Engine Object.
Those Objects have their own state, they have their own connections to other objects, and they all must be kept track of.
Thus... the RAM use.
The more area you have explored and modified, the more changes vis-a-vis the default map must be tracked.
So yes... long story short - the game spends most of the save time actually noting down the frozen world state from RAM.
i started the server and its only using 128M of ram.. i think i need to get on it with the client to get the level to actually load
If you have it configured to pause when empty, yes. After start, it will first wait for an initial connection before loading the save.
ya, thats how it is configured
When your save is this big, I would recommend you disable the pause option.