Failed to set group_fwd_mask on bridge (80028) after upgrade to 4.0.1-69-PRO

[TerpEE93]

I recently upgraded my EVE-NG instance from 4.0.1-65 to 4.0.1-69. After the upgrade I started getting a bunch of "Failed to create network" errors when starting images. In many cases the images wouldn't start at all. For the ones that would, I noticed that LLDP messages were not being forwarded between images.

Digging into the log files, I see this when I try to start a vEOS or vQFX image:

Code: Select all

Apr 05 12:24:37 Apr 05 12:24:37 Online Check state: Valid
Apr 05 12:24:38 Apr 05 12:24:38 ERROR: sat=0
RTNETLINK answers: Invalid argument
Apr 05 12:24:38 ERROR: echo 65535 > /sys/class/net/vnet0040001000e/bridge/group_fwd_mask 2>&1 --- Failed to set group_fwd_mask on bridge (80028).
Apr 05 12:24:38 net.ipv6.conf.vnet0040001000e.disable_ipv6 = 1
Apr 05 12:24:38 Apr 05 12:24:38 ERROR: Failed to set group_fwd_mask on bridge (80028).
Apr 05 12:24:38 Apr 05 12:24:38 ERROR: Failed to create network (11).

When I try to set the group_fwd_mask on an existing bridge, I see that I can not:

root@col-fed-lab-eve-67:~# echo 65535 > /sys/class/net/vnet00200010006/bridge/group_fwd_mask
-bash: echo: write error: Invalid argument

This behavior manifested immediately after the upgrade to 4.0.1-69-PRO. I upgraded via apt, and I believe there was a kernel update as well. Current OS state is Ubuntu 18.04.6 LTS running kernel 4.15.0-173-generic. Output from eve-info.sh is attached.

Any idea how to correct the issue?

Thanks!

[TerpEE93]

OK, I figured it out. And the issue was right there in my problem statement.

For whatever reason, the server booted with the stock Linux kernel after the update. It should have been running the 4.20.17.2-eve-ng-uksm-wg+ kernel. Setting GRUB to boot the 4.20.17.2-eve-ng-uksm-wg+ kernel and a reboot fixed the issue.