Link Aggregation on vSwitch0

I got asked about getting Static Link Aggregation working on the default Management Network on vSwitch0, which reminded me of this frustrating incident.  It’s one of those quirks that you have to swallow as a quirk and make sure to remember. Link aggregation on vSwitch0 need special attention.

One of the early tasks that occurs during a deployment is to make sure that the environment has resilience in it’s configuration, especially the fundamental network connections (VM Network, Management Network, vMotion, Storage).  ESXi hosts come initially deployed with VM Network and Management Network set up on vSwitch0.  If you’re working quickly and set up vSwitch0’s Link Aggregation (NIC Teaming tab), expecting it to flow through to the pre-configured networks, you’ll be let down.

Make sure you go into the VM Network and Management Networks and make sure on the NIC Teaming tab that the Policy Exceptions and Failover Order are all unchecked (unless you’re know exactly what you’re doing).

vSwitch0 Management Network NIC Teaming

image sources

  • vSwitch0 Management Network NIC Teaming: Owned by the author

Published by


John White is walking the path to virtualization mastery.

8 thoughts on “Link Aggregation on vSwitch0”

  1. John,
    Thanks for answering my question. I had those check boxes checked since that was how they defaulted so I can see where my config was the problem.
    So, just to be crystal clear, you are saying that both the vswitch properties and the port properties must have all these unchecked? Is this issue unique to the management network? I set up vMotion ports on a different vSwitch with multiple vmnics without unchecking all those boxes. The LAG was functional in that I could at least ping the vMotion interface between hosts. How might that be different?

    Thank you for helping out a fellow spicehead.


    1. The vSwitch Properties _set_ the default. I assume you’re asking about the Management and VM Networks?

      I think this issue is unique to vSwitch0, as the vSwitch set up by the installation of the hypervisor.

      1. The settings you show in your screen capture are the actual Port settings for say, the Management network. What about the vSwitch properties. It has most all the same options. Load Balancing, Network Failover Detection, Notify Switches and Failback. Only the last two can you disable. Load Balancing is supposed to be IP hash from everything I have read.


        1. The vSwitch is where you set the configuration. The vmnics in question should be “Active Adapters” and using “Route based on IP Hash.”

          I also generally use Fallback “No.”

  2. And that worked. So just disabling all those settings on the management Port now allows it to work. Did some arbitrary link breakage(wire unplugging) and it survived just fine.

    I called my buddy and he said “do you know how many people google for that and can’t find an answer?” How could something so simple and vital be so obscure and impossible to find and answer for?

    Thanks a ton. I have been beating my head against this problem for a month.

Leave a Reply