[BUG] DELL OS10 does not work correctly with two LAG links

[ipspace]

01-l3-lag:

• Test topology
• Test results

It's weird: OS10 claims one of the LAGs is up, the other is down, while Arista claims only one link in the LAG group is active.

[jbemmel]

Both port-channels show "up" when I replace EOS with Dell OS10 in the test

Vlan 1 on Dell OS10 is inactive

When I remove the LAG towards X2, both ports to X1 send/receive LACP. With the full setup, only port 1/1/2 sends LACP; Dell side claims it's not receiving any LACP PDUs on other ports

Replacing EOS with Linux results in 2 failed LAG links on Dell side

Tried setting LACP rate to slow - same result

[ipspace]

Both port-channels show "up" when I replace EOS with Dell OS10 in the test

Considering Arista interoperates with everyone else (https://tests.netlab.tools/_html/coverage.lag), who do you think the likely culprit is? ;)

Vlan 1 on Dell OS10 is inactive

I have no idea what this signifies.

When I remove the LAG towards X2, both ports to X1 send/receive LACP.

Yes, one LAG link works (or OS10 wouldn't pass the MLAG test)

With the full setup, only port 1/1/2 sends LACP; Dell side claims it's not receiving any LACP PDUs on other ports

Supposedly OS10 does not have "passive LACP" mode, so how could it not send LACP on all LAG ports?

Replacing EOS with Linux results in 2 failed LAG links on Dell side

I think we have a problem with Linux VM (https://tests.netlab.tools/linux/libvirt/lag/01-l3-lag.yml-validate.log), potentially after I changed the VM image to 24.04.