Interfaces mismatched on switch VM

[Kolya]

Hey everyone, need some help, had some success with this over a year but suddenly things changed and wont line up anymore.

My install is: Bare metal, two Intel CPUs, 380GB RAM, 1TB HD, EVE version 5.0.1-19

I have some VM Switches, some Cisco, which work fine, some Aruba CX OVAs which are the problem.
Interface 1/1/1 and 1/1/2 work flawlessly.

I connected 1/1/3 to 1/1/3 and 1/1/4 to 1/1/4 from CX switch 1 and CX switch 2 respectively to do a LACP LAG.
The Lag doesn't come online, so I did what a normal person would do, I right clicked the node, selected capture for 1/1/3, and....it's blank.
Selected 1/1/4 and....also blank. That's weird, they're up up on the CLI of the switches so what gives?

Then I enabled 1/1/5-1/1/12 on both sides for kicks and giggles, I selected 1/1/3 for the pcap, and strangely I see LLDP just fine both directions. but the LLDP reports it's for 1/1/5. So I then open a pcap for 1/1/5 and it shows LLDP too, but for 1/1/7, tried 1/1/7 and it's 1/1/9 and here we can extrapolate the pattern.

EVE shows the interfaces as correct for 1/1/1 and 1/1/2, captures on those and their LLDP is all in order. But anything 1/1/3 + is shifted up by two interface IDs.
I'm at a loss as to why this is the case.
Here's what I've tried:
1. Rebooted the VMs
2. Reconfigured the VMs
3. Deleted the lab
4. Deleted the vmdk and redone it
5. Fixed permissions over and over
6. Checked for updates and upgrades of EVE
7. Tested with VMXNET3, E1000, Virtio-net-pci
8. tested other qemu versions, it defaults to 2.4 or something but I couldn't get it to boot at all with anything above 4
9. I tested the AOS-CX OVA for 10.10 and 10.13(the newest)
10. Issue began on EVE-NG version 2.x which I had to do a new install to 5.x the current version to upgrade.

Nothing fixes it that I can figure. Not sure what the qemu version controls, thought it was worth the shot. FIgured maybe the NIC type would matter, but when this worked fine on all interfaces it never worked with VMXNET3

Any help is greatly appreciated to get the interfaces lined up properly.

[Uldis (UD)]

NIC Type can be a matter, classic is e1000
which is passing all through
vendor images are tricky, and can be different NIC drivers in use...
even with different drivers NIC on diff nodes it works.
But if the LACP is not working on your Aruba, then, sorry
test something same with other vendor, like cisco or arista..
it works
sad to answer, but it is not eve issue here

[Kolya]

Yeah I felt the same, just hoping maybe someone knew a way to kick around in the device shell to fix it or band-aid it.

Thanks though.