What is the correct topology for 2 ESXI hosts, no shared storage, to set up LAG groups on all nics for vMotion and general VM data?

Here’s my setup:

Homelab for learning purposes

2 ESXi 6.5u2 hosts, VCSA 6.5u3

5 Gb NICs each: 1 for management, 4 for data

no shared storage, individual datastores on each host

My goal:

– use only vDS(s), no vSs

– put each hosts’ 4 NICs into LAG(s)

– have vMotion traffic go through LAG uplink(s) between hosts

– have VM traffic go through LAG uplink(s) between hosts


– I assume each hosts’ NICs would have to be in their own LAG group?

– Can I accomplish this using one vDS? Or would I need two – one for each LAG group?

– Can vMotion traffic and VM traffic both go through the same LAG group of each host?

– Say one host’s VM is in a dPG with a 4 port LAG uplink (theoretical 4Gb bandwidth), and I simultaneously transfer data from that VM to two client machines that are only 1Gb ethernet each. Would each client be able to transfer at 1Gb/s each (since the LAG group consist of 4 ports), or would each client only transfer at 500Mb/s
each, since each of the 2 clients are not in any LAG groups?

I’m asking because I put both hosts NICs into a single LAG group in a single vDS, and notice no change in vMotion transfers. Upon some research, I realized this is the incorrect topology. Any tips or pointers on what the correct topology is would be greatly appreciated.

Thanks in advance, all! Excited to play around more with my homelab!

View Reddit by ltc_proView Source


To see the full content, share this page by clicking one of the buttons below

Related Articles

Leave a Reply