Proxmox page unreachable

[cyrus104]

I installed a proxmox on a machine that I was recently using for ESXi (Intel NUC 9 Xeon w/ Mellanox Connectx-5). After installing 8.1-1 (whatever the latest is as of yesterday) and then gave it a static address in the native vlan. Once it was operational and updated, I added a vlan vmbr0.20 on the right interface. I gave it an IP and applied, now both IPs worked and I removed the native vlan IP and added a gateway to the mangement vlan interface. Everything was still working.

I restored 5 VMs from a regular (non-PBS) CIFS share drive. All of the VMs transferred and looked good. I came back a bit later and now I am getting a site can't be reached and the interface cannot be pinged for some reason. IP Link shows the interface up, the switch shows the link speed selected and is allowed for vlans. I have a unifi switch and the port is configured the exact same as the port of another Proxmox node.

I have checked the network interfaces via a PiKVM I have on the computer.

Here is what my /etc/network/interfaces looks like:
auto vmbr0.20
iface vmbr0.20 inet static address 10.10.10.2/24 gateway 10.10.10.1 auto vmbr0 iface vmbr0 inet manual bridge-ports eno1 bridge-stp off bridge-fd 0 bridge-vlan-aware yes bridge-vids 2-4094

Thanks

 

[aaron]

Have you checked in /etc/hosts if the IP still matches the hostname? If the IP set in the /etc/hosts does not exist anymore, it would cause the pveproxy and potentially pvedaemon services (both needed for the API & WebUI to not start properly AFAIR

You could check that by querying these services if they are up and running and what they logged:
systemctl status pveproxy and journalctl -u pveproxy for example.

 

[cyrus104]

@aaron I checked the hosts file and it looks good.

I just rebooted and everything is working now, which isn't ideal.

I update if it happens again.

 

[cyrus104]

@aaron well, I'm back. After an 4 hours of working fine.... I'm back.

It's a ping issues too, not a UI issue.

 

[cyrus104]

ok, so this is a weird issues with the Mellanox ConnectX-4 card falling offline... I'm struggling to identify a fix I didn't need to do anything for ESXi, XCP-ng, or Ubuntu.

 

[aaron]

ok, so this is a weird issues with the Mellanox ConnectX-4 card falling offline... I'm struggling to identify a fix I didn't need to do anything for ESXi, XCP-ng, or Ubuntu.

How exactly? Because this doesn't seem to be a simple misconfiguration resulting from changing the IP address configured for the host itself.

Do you have the latest firmware on the NICs? With Mellanox ones, you can easily use the mlxup utility. https://network.nvidia.com/support/firmware/mlxup-mft/

It is possible that with newer kernels you will see some issues. Updating the firmware is usually the first step to see if if fixes the issue.

 

[cyrus104]

I updated the firmware and have check dmesg. The cards stays up for about 30 seconds after boot and then both interfaces go into cleanup and disappears.

[ 26.907682] mlx5_core 0000:01:00.0: E-Switch: Unload vfs: mode(LEGACY), nvfs(0), necvfs(0), active vports(0) [ 26.949676] mlx5_core 0000:01:00.0: E-Switch: Disable: mode(LEGACY), nvfs(0), necvfs(0), active vports(0) [ 28.302930] vmbr0: port 1(eno1np0) entered disabled state [ 28.303137] mlx5_core 0000:01:00.0 eno1np0 (unregistering): left allmulticast mode [ 28.303140] vmbr0: port 1(eno1np0) entered disabled state [ 28.693732] mlx5_core 0000:01:00.0: E-Switch: Disable: mode(LEGACY), nvfs(0), necvfs(0), active vports(0) [ 29.498116] mlx5_core 0000:01:00.0: E-Switch: cleanup [ 30.073545] mlx5_core 0000:01:00.1: E-Switch: Unload vfs: mode(LEGACY), nvfs(0), necvfs(0), active vports(0) [ 30.086121] mlx5_core 0000:01:00.1: E-Switch: Disable: mode(LEGACY), nvfs(0), necvfs(0), active vports(0) [ 31.705745] mlx5_core 0000:01:00.1: E-Switch: Disable: mode(LEGACY), nvfs(0), necvfs(0), active vports(0) [ 32.602107] mlx5_core 0000:01:00.1: E-Switch: cleanup