Problems Nexus Port-Channel & Bond on Linux-Server

[chris-doro]

Hi,
I try to setup port-channel with 2 Nexus (vpc) and a Linux-Server (Rocky 8).
After I configured a while, it was coming up and running. Could connect to server which has IP-address on VLAN of the bond.
This worked.
But when I reboot the server, I always have to play around to bring the bond up again. 
On the Nexus it shows suspended, on the Linux-server down. 
With shut - no shut and changing the status on the server after approx. 30 minutes I get it up again.
But after every reboot same problems. 
It's connected over 2x 10Gbit/s fiber.
Any suggestions?
Thanks.

[MHM Cisco World]

no lacp suspend individual, this my recommend to you
NOTE:- this feature disable in some case not all time

https://davidsudjiman.wordpress.com/2018/01/31/lacp-suspend-individual/

[Reza Sharifi]

Hi,

For the Portchannel connecting the switches to a server, try using mode "on" which basically means static, and test again.

HTH

[chris-doro]

Hello,
thanks for the suggestions.
It looks like the problem discussed here: https://github.com/systemd/systemd/issues/15208
I am using Rocky 8.8 (4.18.0-477.15.1.el8_8.x86_64).
But it's much better when doing the "trick": /etc/systemd/network/98-bond-mac.link
[Match]
Type=bond
[Link]
MACAddressPolicy=none

I still need to manually type the command "nmcli connection up bond0" after the reboot (even if ONBOOT=yes), but at least after this command, now the channel-group is coming up and stable.