Struggling with ssh on my arch linux server

[novel hearth]

I run a linux server myself, with its ssh working fine, after sometime (let's just say 1-2 days after its being turned on) it just says connection refused: no route to host
Temporary fix: to fix this I just restart the server and everythings back working good. But obviously looking for a permanent fix to prevent this from happening, all kind of help is welcome.

[novel hearth]

if anyone knows why this might be happening, it is welcome

maybe something to do with nmcli (network manager config)?

[proven forum]

Is the server still connected to the network at the time?

[novel hearth]

Yeah

[proven forum]

How did you verify that

[novel hearth]

like

its at the same network

via lan

but

the issue might be cuz of the ip

[proven forum]

But is it actually still connected when ssh fails?

[novel hearth]

yeah, but for some reason it just breaks

i need to reboot for it to get fixed

[proven forum]

No route to host usually means that either the host is offline or your own routing is messed up. The former is more common.

[novel hearth]

so then what could be the issue=

my router?

[proven forum]

Your client machine

[novel hearth]

the server itself u mean?

[proven forum]

Wdym

[novel hearth]

idk

[proven forum]

No, your computer you're using to ssh into it.

[proven forum]

Just was a wild guess?

[novel hearth]

na the computer caan't be the issue

cuz its outside of the server's internet network

[proven forum]

Wdym

[novel hearth]

like

i have port forwarding

im using a computer from another place

to ssh

[proven forum]

Are you sure the port is open

[novel hearth]

yeah pfc

ofc

[proven forum]

It's not scheduled to close at certian times or after inactivity?

[novel hearth]

cuz i have a rpi4 with the same stuff (network connection and that on the same wifi network)

and it works perfectly fine

[novel hearth]

um

i don't really know that

maybe it has to do with router updates?

[proven forum]

I would assume that has a diffierent rule through?

[novel hearth]

idk

[proven forum]

Unlikely

[novel hearth]

ill try doing a fresh instal

[proven forum]

Are you able to access ssh from within the network when this issue occurs?

[proven forum]

That probably isn't neccessary.

[novel hearth]

you mean ssh to other devices when the thing is down?

[proven forum]

No, I mean, can you connect to your internal network with the ssh client machine and ssh into the server when your are unable to ssh from the internet?

[novel hearth]

nah because

the ip seems to change too

u mean locally right?

[proven forum]

Yes

[novel hearth]

the ip gets changed too

[proven forum]

That's your issue

[novel hearth]

idk why

the ip gets changed but when i reboot it gets back to normal

[proven forum]

Are you using DHCP to get an IP on the server?

[novel hearth]

nah

im using nmcli

nmcli con mod (attributes=

[proven forum]

Using DHCP or not

[novel hearth]

not using it

[proven forum]

You've manually set an IP and default gateway?

[novel hearth]

yeah

same stuff with the raspberry pi and it works without an issue

[proven forum]

What's the static IP set to, what is the IP changing to, and what is the IP when it boots?

[novel hearth]

192.168.1.15 --> static
192.168.1.135 --> when it changes
when it boots --> 192.168.1.15

but for a little period of time it says 0.0.0.0

but when i refresh the routers page it says 192.168.1.15

weird issue also is that the hostname isn't displaying

[proven forum]

Do you have another DHCP daemon running?

[novel hearth]

nope

seems to be the weirdest bug ever

do you know what could be causing this? @proven forum

[proven forum]

Something is breaking and causing the server to request an IP from DHCP, or something else is setting the IP. Try setting the IP for that device on your DHCP sever, which is probably your router.

[novel hearth]

weird thing is that

i think dhcp is set to range for some ips range

but on the rpi

i've not set any dhcp

and it works well

how could i try a workaround this?

[proven forum]

Not sure why this issue is specific to only one of your devices, just stating what I know

[novel hearth]

maybe if i try fresh install?

maybe i fucked something that has to do with ip

or whatever

[proven forum]

I guess you can give it a try if it's not a massive hassle

[novel hearth]

nah since i just use it for mining and shit

[novel hearth]

What could be the issue @proven forum

[proven forum]

.

That's all I can say with the info we have

[lament egret]

Nah

[novel hearth]

Why

[lament egret]

Have you actually checked the server logs after that issue? SSH, Journalctl, can you ping from the server to your PC? Did you checked ip a or any logs from your DHCP client? (dhcpcd or dhclient (deprecated))

You can also try applying this @novel hearth https://wiki.archlinux.org/title/Network_configuration#Revert_to_traditional_interface_names

Which I dont think is going to change anything but it will make Network device names static

For example, if your network interface is enp1s0, after you connect a new PCI-E device and reboot the server, the new name of the interface could be enp2s0

Bit it's probably DHCP so check the logs of your installed DHCP client with journalctl

And also arp -a to check the ARP table when your PC cannot reach your server

Or viceversa

[proven forum]

It can help with some poorly made drivers

It's basically required for a few realtek ones to work reliably

[novel hearth]

cant ping or anything

only restarting fixes it

which is too annoying cuz a server is meant to stay alive

and just physical access for updates/etc

i have my raspberry with the same config and its been 18 days up nwo

now

[lament egret]

@novel hearth but did you even tried to Search something about the internet in the logs?

You have local access to the machine, just connect it to a screen, put a keyboard and run sudo dmesg -w and sudo journalctl -f

At the same time, you can do that using tmux

(If you are only on a TTY)

[novel hearth]

ill try that

[novel hearth]

this is everything

what could be the issue @lament egret

[lament egret]

What about the Journal lot?

Log?

[novel hearth]

im gonna go on linux and send it rq

Jun 05 19:53:53 forcaworks1 NetworkManager[453]: <info> [1717610033.1482] device (wlo1): supplicant interface state: inactive -> disconnected
Jun 05 19:53:53 forcaworks1 NetworkManager[453]: <info> [1717610033.1482] device (p2p-dev-wlo1): supplicant management interface state: inactive -> disconnected
Jun 05 19:53:53 forcaworks1 NetworkManager[453]: <info> [1717610033.1540] device (wlo1): supplicant interface state: disconnected -> inactive
Jun 05 19:53:53 forcaworks1 NetworkManager[453]: <info> [1717610033.1540] device (p2p-dev-wlo1): supplicant management interface state: disconnected -> inactive
Jun 05 19:59:07 forcaworks1 sshd[11801]: Accepted publickey for forcadeitor from 100.104.158.87 port 55680 ssh2: RSA SHA256:h+Pp3HMDbG5BTqGpG83zBeRStg+NSELNibimocYlK4g
Jun 05 19:59:07 forcaworks1 sshd[11801]: pam_unix(sshd:session): session opened for user forcadeitor(uid=1000) by forcadeitor(uid=0)
Jun 05 19:59:07 forcaworks1 systemd-logind[454]: New session 5 of user forcadeitor.
Jun 05 19:59:07 forcaworks1 systemd[1]: Started Session 5 of User forcadeitor.
Jun 05 20:00:38 forcaworks1 sudo[11811]: forcadeitor : TTY=pts/0 ; PWD=/home/forcadeitor ; USER=root ; COMMAND=/usr/bin/journalctl -f
Jun 05 20:00:38 forcaworks1 sudo[11811]: pam_unix(sudo:session): session opened for user root(uid=0) by forcadeitor(uid=1000)

@lament egret

[lament egret]

That does not say anything

[novel hearth]

that's all it appears

[lament egret]

You have the server on right now?

[novel hearth]

yeah

[lament egret]

Without internet access?

[novel hearth]

nope

i just restarted it

i have ssh access

@lament egret

[lament egret]

You just restarted it one time right?

[novel hearth]

yeah

[lament egret]

Then do sudo journalctl -b -1 --no-tail > file.txt and upload the file to https://pbin.nadeko.net

Select burn after reading and upload the file.txt using attach a file

You can transfer the file from your server to your PC using SCP or rsync

[novel hearth]

how to do that?

transfer file i mean+

[lament egret]

Using scp?

[novel hearth]

or what if i just cat the file

and copy paste it

[lament egret]

You can do that, but it will be a big large output

[novel hearth]

https://pbin.nadeko.net/?6ef8200505d6c6e8#GvJ1mHwKoe57wxPcVmomayfnEZd9Xs4V1euocvfXxXKf

[lament egret]

@novel hearth is getting connected via wifi?

[novel hearth]

ethernet

why do you ask?

[lament egret]

becuase there is a lot of logs about the wifi device from network manager

if it's a server, why are you using network manager?

[novel hearth]

i use that on the raspberry too

[lament egret]

I suggest to uninstall network manager and only use dhcpcd

you don't need network manager on servers

Or any VPS provider would have network manager built-in on their linux ISOs

[novel hearth]

ok so

my idea is

i turned it on today

let's see how many days it takes for this to go off again

and we could see what to do

[lament egret]

The last restart was becaue you didn't have internet nor ssh access on the server right?

[novel hearth]

eys

yes

[lament egret]

I would suggest to uninstall network manager if this happens again then

and only use dhcpcd

because there a lot of weird logs from network manager with the wifi interface

[novel hearth]

ok

thanks so much

but+

if this happens again

and we need to use

dhcpcd

can you show me how to

because i never used thatç

[lament egret]

https://wiki.archlinux.org/title/dhcpcd @novel hearth

[novel hearth]

It's been 10 days without touching it and it's been working well

since today

so my guess is