LAG 803.ad Connection Issues

[mcpl-rburns]

We have a single Proxmox host running 8.4.1 connected to a Juniper 4400xd-48F switch. The object is to use two SFP+ connections on the Proxmox host to connect to the Juniper switch which will also be networked to our primary storage unit for NFS VM storage.

If we connect to the switch via a simple RR or Active-Backup, we are able to ping the storage and the switch management. But if we enable 802.3ad we can no longer reach either.

We have been working on this for multiple weeks, and not making any progress. Everyone from the Proxmox side of things leans toward the switch, but the switch person at my end leans toward Proxmox. I would like to think it's not Proxmox, but at this point I really don't care, I just want to move forward.

Ideas?

 

[Heracleos]

There could be several causes.
Indicatively, it appears that your switch (or bond on proxmox) is unable to negotiate the correct LACP mode.
Try checking on the switch whether on the LAG, the lacp mode is active and what type passive/active, and if so, the type of balancing (src-ip, dst-ip, l4-src-port, l4-dst-port).
Try to check if spanning-tree on the LAG creates problems and closes ports to avoid loops.