I have been running HA in a Proxmox VM for a couple years and it has been great. However, I often saw errors related to what seemed networking issues (disconnections from the internet) and chalked it up to Unifi beta firmware issues or actions I took. Recently, I tried running the command dmesg -w
in the shell of the Proxmox machine running HA and saw lots of these errors:
[90602.829933] e1000e 0000:00:1f.6 eno2: Detected Hardware Unit Hang:
TDH <e2>
TDT <d>
next_to_use <d>
next_to_clean <e1>
buffer_info[next_to_clean]:
time_stamp <101587504>
next_to_watch <e2>
jiffies <101587c20>
next_to_watch.status <0>
MAC Status <80083>
PHY Status <796d>
PHY 1000BASE-T Status <3800>
PHY Extended Status <3000>
PCI Status <10>
[90602.957665] e1000e 0000:00:1f.6 eno2: Reset adapter unexpectedly
[90603.054299] e1000e 0000:00:1f.6: Some CPU C-states have been disabled in order to enable jumbo frames
[90603.054757] vmbr0: port 1(eno2) entered disabled state
[90606.820053] e1000e 0000:00:1f.6 eno2: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
[90606.820125] vmbr0: port 1(eno2) entered blocking state
[90606.820129] vmbr0: port 1(eno2) entered forwarding state
From what I’ve read they appear to be related to a bug in a driver for a common Intel ethernet chipset. I believe my Lenovo Tiny P360 has an “Intel Jacksonville I219LM (w/ AMT 16.x)” (from Lenovo’s site).
Anyhow, I am going to add a Mellanox CX322A (2 x SFP+ 10Gbit) to the computer later today and move HA to use that card as I think that will solve the issue for HA… but I’d still like to fix the issue if at all possible.
I am mentioning it here as there are lots of HA users running HA in Proxmox and given the popularity of the Intel NIC chipset, this may be a really common issue amongst Proxmox + HA users.