Network connectivity broken in 5.11.21-1?

[victorhooi]

Hi

I have just setup a new Proxmox system that I've recently set up on Proxmox 6.4.

Network card is a Mellanox ConnectX-5 with 100Gbe ports.

I've also installed the latest 5.11 kernel to test this out. I know that 5.11.x was working on previous installs on this system.

However, when I did an update yesterday, the current 5.11 kernel (5.11.21-1) seems to have broken networking - when I run with this kernel, the vmbr0 interface is in state UNKNOWN and I'm not able to ping the upstream router.

If I then boot back in the 5.4.119-1 kernel, the interface is back in state "UP", and I have network connectivity again.

Has anybody else seen this issue? Or know what's going on?

Thanks,
Victor

 

[mira]

Could it be the interface name changed with kernel 5.11?

 

[victorhooi]

Hmm, this is the output of ip addr, with the 5.4 kernel:

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: eno1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
    link/ether e0:d5:5e:96:9a:18 brd ff:ff:ff:ff:ff:ff
3: eno2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
    link/ether e0:d5:5e:96:9a:19 brd ff:ff:ff:ff:ff:ff
4: enp1s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master vmbr0 state UP group default qlen 1000
    link/ether 0c:42:a1:02:08:dc brd ff:ff:ff:ff:ff:ff
6: vmbr0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 0c:42:a1:02:08:dc brd ff:ff:ff:ff:ff:ff
    inet 10.7.12.3/23 scope global vmbr0
       valid_lft forever preferred_lft forever
    inet6 fe80::e42:a1ff:fe02:8dc/64 scope link
       valid_lft forever preferred_lft forever
7: enp3s0f3u1u3c2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
    link/ether c6:32:12:74:58:a9 brd ff:ff:ff:ff:ff:ff

And with the 5.11 kernel (I had to take a screenshot, as I can't SSH in to the box):



If I check /etc/network/interfaces, I see that vmbr0 is bridging the port enp1s0.

In 5.11 - are you thinking it's been renamed to enp1s0np0?

Will simply editing the /etc/network/interfaces file directly fix this? Is that a good practice?

 

[mira]

Yes, sometimes the naming changes because of the kernel or BIOS updates and so on.
It seens yours now contains 'np0' at the end.
Just change enp1s0 in /etc/network/interfaces to enp1s0np0 and you should be able to get it working with `ifreload`.

 

[victorhooi]

Thanks!

I can confirm that adding the suffix "np0" to the original device name "enp1s0" resolved the issue.

Should I also rename the iface enp1s0 near the top fo the files as well?

auto lo
iface lo inet loopback

iface enp1s0 inet manual

iface enp3s0f3u1u3c2 inet manual

iface eno1 inet manual

iface eno2 inet manual

auto vmbr0
iface vmbr0 inet static
    address 10.7.12.3/23
    gateway 10.7.12.1
    bridge-ports enp1s0np0
    bridge-stp off
    bridge-fd 0

 

[mira]

Yes, as it references a device that's not available.