Configuring LACP on Distributed Switches – VMware vSphere 8 & Dell VxRail

With Link Aggregation Control Protocol (LACP) support on a vSphere distributed switch, you can connect ESX hosts to physical switches using dynamic link aggregation. Multiple link aggregation groups can be created on a distributed switch to aggregate bandwidth from physical NICs to ESXi hosts connected to LACP port channels.

Most of you have probably created a LACP configuration on a distributed switch in various customer situations or in your home lab. In today’s post, I want to walk you through the process, but also highlight a limitation for Dell VxRail Appliances. I’ll skip the step of creating a Distributed Switch here, as this post is more of an extension for a Distributed Switch and which one is assumed.

Once your distributed switch is configured, select it and then > Configure > LACP > NEW. Under Migrating Network Traffic to LAGs, you get a description of how and what you need to adjust to make the LACP configuration work properly.

You give it a name, define the ports that should be in a LAG. Set the active/passive, load balancing, and timeout modes. If necessary, you can also adjust the port policies.

And now we come to the limitation of VxRail (it doesn’t matter if you use 7.0.xxx or 8.0.xxx versions), for the “VxRail Network” i.e. management etc. a maximum of 2 ports can be combined in a LAG. The VMware HCIA Distributed Switch, which will be created during the inital VxRail deployment, is meant here.

For example, if you want to configure 4 or more ports, the VxRail network will only use 2 ports. And typically, these 2 ports are combined into a LAG for vSAN traffic only. For all other guest networks / distributed switches, you can configure and use more than 2 ports.

After you have created your desired LAGs, right click on the Distributed Switch > Add and Manage Hosts… > Manage host networking
(Do not go to Distributed Port Group as shown in the picture – sorry for that)

Then select the hosts on which you want to run a LACP configuration.

Now set up the uplinks you want to combine into a LAG. As soon as you open one of the uplinks, you will see the defined LAG in addition to the usual uplink, in our case: PO31-38. Distribute these to the desired vmnics.

While you are making those adjustments, your network colleague must also make the LACP configuration on his hardware switches. Otherwise, your applied configuration will not work. Iwant to thank my two big models Marc Huppert and Daniel Simelka for helping me with this Task. That’s it from this blog post, if you have any questions leave a comment below. 🙂