Problem with Intel NIC

Morta · 2023-10-16 19:06:11

Oct 16 21:01:08 5erver kernel: bond0: (slave enp7s0f0): speed changed to 0 on port 1
Oct 16 21:01:08 5erver kernel: bond0: (slave enp7s0f0): link status definitely down, disabling slave
Oct 16 21:01:08 5erver kernel: bond0: active interface up!
Oct 16 21:01:08 5erver jellyfin[1259]: [21:01:08] [DBG] [8] Jellyfin.Networking.Manager.NetworkManager: Network availability changed.
Oct 16 21:01:08 5erver jellyfin[1259]: [21:01:08] [DBG] [8] Jellyfin.Networking.Manager.NetworkManager: Network Address Change Event.
Oct 16 21:01:08 5erver kernel: ixgbe 0000:07:00.1 enp7s0f1: NIC Link is Down
Oct 16 21:01:08 5erver kernel: bond0: (slave enp7s0f1): speed changed to 0 on port 2
Oct 16 21:01:08 5erver kernel: bond0: (slave enp7s0f1): link status definitely down, disabling slave
Oct 16 21:01:08 5erver kernel: bond0: now running without any active interface!

http://0x0.st/HJi2.txt

with nomodeset

seth · 2023-10-16 19:09:22

"nomodeset" and what? Did you re-enable the ASPM?

To be sure: w/ pcie_aspm=off the NICs remained stable, the "only" problem was that your VGA device also dropped out?
In that case the immediate goal would be to get the AST VGA to work w/o ASPM (though again: super-weird relation tbw)

Morta · 2023-10-16 19:27:11

No with pcie_aspm=off isn't stable at all. video works well.

Morta · 2023-10-16 19:43:14

8: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 4e:b7:bc:18:c0:bf brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.100/24 brd 192.168.1.255 scope global noprefixroute bond0
       valid_lft forever preferred_lft forever
    inet 192.168.1.156/24 brd 192.168.1.255 scope global secondary dynamic noprefixroute bond0
       valid_lft 6466sec preferred_lft 6466sec
    inet6 2a02:168:a774::2222/48 scope global noprefixroute
       valid_lft forever preferred_lft forever
    inet6 fe80::39fe:6366:30c:f19/64 scope link noprefixroute
       valid_lft forever preferred_lft forever

Why my bond0 have two IPv4?

seth · 2023-10-16 20:50:53

No with pcie_aspm=off isn't stable at all.

Sorry, I missed your edit of #19

A typical cause for multiple leases on any NIC would be multiple dhcp services, that would also explain the frequent network loss but isn't reflected in your journal.

How does the network behave if you stop NM and run dhcpcd on only one of the HW NICs?
https://wiki.archlinux.org/title/Dhcpcd#Running

Arch Linux

#26 2023-10-16 19:06:11

Re: Problem with Intel NIC

#27 2023-10-16 19:09:22

Re: Problem with Intel NIC

#28 2023-10-16 19:27:11

Re: Problem with Intel NIC

#29 2023-10-16 19:43:14

Re: Problem with Intel NIC

#30 2023-10-16 20:50:53

Re: Problem with Intel NIC

Board footer