Unstable networking on Dell optiplex 7090 sff (e1000e)

Summary

Network instability on Dell 7090 with ubuntu 20.04.3 (e1000e interface)

Details

We recently purchased some nice shiny 7090 SFF (vPro enabled) desktops which came with windows 10 installed, so we re-installed them using our pxe infrastructure which utilises the official ubuntu autoinstall methodology (user-data.yml).

Initially we had issues where the installer would not start, so we switched to the HWE kernel in 20.04.3 and then the install and boot is successful.

Unfortunately we are finding the networking on these machines to be extremely unstable, dropping many packets and appearing generally ‘lumpy’.

MTU is 1500 everywhere relevant.

We are unable to mount our NFS servers at all from these boxes.

Things we tried

• HWE Kernel
• OEM Kernel
• Disabling AMT/MEBx
• 2 different identical boxes
• known-good switch port (swapped with a functioning 7060 running 20.04.3)

Query

Has anyone else encountered this issue and perhaps come-up with a fix? Are there updated drivers or kernel for these boxes? I see that a version of this box is ubuntu certified but it notes that a special kernel was used: https://ubuntu.com/certified/202101-28613

I was considering trying 21.04 next, but this post seems to describe a very similar issue: New Dell 7090 Tower - Installed Ubuntu 20.10, 21.04 and 21.10 - Getting network packet loss until usb thumb drive plugged in

Technical details

The network card as given by lspci -v:

00:1f.6 Ethernet controller: Intel Corporation Ethernet Connection (14) I219-LM (rev 11)
    Subsystem: Dell Ethernet Connection (14) I219-LM
    Flags: bus master, fast devsel, latency 0, IRQ 125
    Memory at 6e900000 (32-bit, non-prefetchable) [size=128K]
    Capabilities: [c8] Power Management version 3
    Capabilities: [d0] MSI: Enable+ Count=1/1 Maskable- 64bit+
    Kernel driver in use: e1000e
    Kernel modules: e1000e

and lshw:

*-network:1
     description: Ethernet interface
     product: Ethernet Connection (14) I219-LM
     vendor: Intel Corporation
     physical id: 1f.6
     bus info: pci@0000:00:1f.6
     logical name: enp0s31f6
     version: 11
     serial: d0:8e:79:04:78:52
     size: 1Gbit/s
     capacity: 1Gbit/s
     width: 32 bits
     clock: 33MHz
     capabilities: pm msi bus_master cap_list ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt-fd autonegotiation
     configuration: autonegotiation=on broadcast=yes driver=e1000e driverversion=5.11.0-34-generic duplex=full firmware=0.4-4 
     resources: irq:125 memory:6e900000-6e91ffff

The Intel I219-LM generally works better with the drivers supplied by Intel

Here’s how to install them:

1. Download the latest driver from the Intel website (3.8.4 as of this answer)
2. Expand the file:

   gzip -d e1000e-3.8.4.tar.gz
   tar -xfv e1000e-3.8.4.tar
   

3. change to the src directory
4. Compile and install the driver

   sudo make install
   

5. Reboot

This has generally been the solution I’ve had to use for both Dell and Lenovo machines with this network device

This appears to now be fixed in the latest 20.04 oem kernel ...

As discussed here: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1927925