I’ve had this issue since my first installation in September, and it’s followed me through a new installation. My internet is occasionally disconnecting for 10-30 seconds before coming back. It’s completely inconsistent when it’ll happen, sometimes it’ll go hours without doing it and sometimes it’ll do it three times within a 10 minute span.
These lines show up in the NetworkManager logs as it happens:
Jan 20 20:26:03 EndeavourOS NetworkManager[1024]: <info> [1705807563.1549] device (enp42s0): state change: activated -> unavailable (reason 'carrier-changed', sys-iface-state: 'managed')
Jan 20 20:26:03 EndeavourOS NetworkManager[1024]: <info> [1705807563.1550] dhcp4 (enp42s0): canceled DHCP transaction
Jan 20 20:26:03 EndeavourOS NetworkManager[1024]: <info> [1705807563.1550] dhcp4 (enp42s0): activation: beginning transaction (timeout in 45 seconds)
Jan 20 20:26:03 EndeavourOS NetworkManager[1024]: <info> [1705807563.1550] dhcp4 (enp42s0): state changed no lease
Jan 20 20:26:03 EndeavourOS NetworkManager[1024]: <info> [1705807563.1668] manager: NetworkManager state is now CONNECTED_LOCAL
These are in the dmesg logs:
[Sat Jan 20 20:25:54 2024] r8169 0000:2a:00.0 enp42s0: Link is Down
[Sat Jan 20 20:25:55 2024] userif-3: sent link up event.
[Sat Jan 20 20:25:55 2024] userif-6: sent link up event.
[Sat Jan 20 20:26:01 2024] userif-6: sent link up event.
[Sat Jan 20 20:26:01 2024] userif-3: sent link up event.
[Sat Jan 20 20:26:10 2024] Generic FE-GE Realtek PHY r8169-0-2a00:00: Downshift occurred from negotiated speed 1Gbps to actual speed 100Mbps, check cabling!
[Sat Jan 20 20:26:10 2024] r8169 0000:2a:00.0 enp42s0: Link is Up - 100Mbps/Full (downshifted) - flow control rx/tx
Only thing understandable by myself is to check the cabling, but I doubt that’s it because I’ve tried multiple cables with the same result, and my connection never drops when I’m booted into Windows.
This happens on both wifi and ethernet, and I’ve tried both ethernet drivers with the same result.
I was going through your older post to get to get your system information:
Looks like that system is different from the system your troubleshooting today(this one looks like it has newer Network card). Can you run:
inxi -Faz | eos-sendlog
It sounds like you are dual booting your system with Windows (which has no network outages)? Have you applied all the updated firmware to your desktop?
Do you also have the the Arch and LTS Kernels installed? If yes, and you boot the LTS kernel, is Ethernet and Wifi cards still operational. Or do we need to install additional drivers ( like you did in the testing you mentioned at the beginning of the post?
The reason I ask, I have a Realtek wifi (usb) devices in some of my home machines. In some cases the regular kernel built in driver was unstable for connections. The solution was switch to lts kernel. Then i install the recommended driver in the Arch wiki.
Yeah this is a completely different system from that old one, here’s the inxi link. I use primarily the Zen kernel, but I’ve tried both the regular and LTS ones with no change.
I’ve updated the BIOS if that’s what you mean, not sure what other firmware needs to be installed. The ethernet port I’ve been using is just the one integrated into the motherboard which I did all the firmware installation for when I built the machine.
From when I worked at IBM and with Lenovo usually there were also firmware for devices (e.g. cameras, harddrives, network cards). But based on the below link I am only seeing Bioses for your Motherboard (B550M-PRO-VDH-WIFI).
Sorry this response is a billion years late, kept procrastinating on doing some actual testing.
The motherboard option didn’t seem to affect anything. Keeping the two ports free seemed to help somewhat? It still happens often but seems to be much less frequent than it is when I have things plugged into them.
Bumping this, did a full reinstall today and the issue persists. Was going to post on the Arch forums by the suggestion of @kagetora13, but seems their number one rule is to not post if you’re running anything but vanilla Arch, so that was out of the question.
No dice. I’ll provide some logs that aren’t two months outdated lol, here it is after two disconnects in a row:
[Thu Mar 14 17:21:21 2024] r8169 0000:2a:00.0 enp42s0: Link is Down
[Thu Mar 14 17:21:48 2024] Generic FE-GE Realtek PHY r8169-0-2a00:00: Downshift occurred from negotiated speed 1Gbps to actual speed 10Mbps, check cabling!
[Thu Mar 14 17:21:48 2024] r8169 0000:2a:00.0 enp42s0: Link is Up - 10Mbps/Full (downshifted) - flow control rx/tx
[Thu Mar 14 17:21:51 2024] r8169 0000:2a:00.0 enp42s0: Link is Down
[Thu Mar 14 17:22:18 2024] Generic FE-GE Realtek PHY r8169-0-2a00:00: Downshift occurred from negotiated speed 1Gbps to actual speed 10Mbps, check cabling!
[Thu Mar 14 17:22:18 2024] r8169 0000:2a:00.0 enp42s0: Link is Up - 10Mbps/Full (downshifted) - flow control rx/tx
Nice timing on the reply, was just typing something up lol. I had tried multiple cables so that wasn’t the issue.
Went out and bought a cable tester to test my wall ports since that was the last remaining thing I could think of, and sure enough they were completely screwed up. Cut both ends and did them from scratch, lo and behold the issue seems to be gone AND my connection speeds are 10x what they used to be. It’s only been a few hours so I’m not TOTALLY sure it’s gone but I’m pretty confident in saying the issue is solved.