VMware NSX-T was recently announced just a bit ago, so I thought it would be helpful to do a live demo of an upgrade to 2.4.1. In short, there’s good news that all but a few of the upgrades I’ve seen have been successful on the initial attempt thanks to the Pre-Upgrade / Post-Upgrade Checks that are built into each section, but any that weren’t successful on the initial upgrade were due to hitting timeouts waiting for a manager node to respond post upgrade (due to really old, under-performing hardware), which were easily remedied with a check of the error and then retry the Manager Node upgrade.
The 2.4.1 update is a core maintenance release, with a new enhancement for VMware HCX, adding functionality for virtual machine migration to on-premises NSX-T based deployments. I have a client doing this at the moment and can attest that it’s greatly welcome functionality for companies acquiring other companies with overlapping networks. VMware HCX provides capabilities to migrate from disparate versions as well, so migrating workloads from an acquired company with many different ESXi versions, is a real plus.
To begin the upgrade of NSX-T 2.4.1, we download the .mub upgrade file from my.vmware.com and log into NSX manager. Once we’re logged into NSX manager, we navigate to System on the top toolbar and then Upgrades on the left navigation pane. In the Upgrade window, choose the location of the NSX .mub upgrade bundle from local disk or URL. After we’ve chosen our upgrade bundle, we click upgrade and the upgrade bundle is validated by NSX Manager and then staged for upgrades. Once the upload status changes to Upgrade Bundle retrieved successfully, we click Begin Upgrade and the VMware NSX Upgrade Coordinator starts.
There are five steps in the upgrade, separated by a clickable top toolbar. Bundle and Status – Hosts – Edges – Controller Nodes – Management Nodes. After accepting the End User Agreement, we run the Host Upgrade on our compute workload clusters that have NSX installed and then we click on Run Post Checks to ensure they’re operable. Edge upgrades are next with the same process as well as Controller Nodes.
The final step in the NSX-T 2.4.1 upgrade process is Management Nodes, which have an option to return the NSX management cluster into service after a single or a 3-node cluster is formed. As a bit of guidance, it’s always a good idea to wait for the 3-node cluster to be operational before a return to service. However, if you’ve got a short outage window or allowance, you can return to service with a single NSX manager node, but be advised, performance usage will increase greatly as other NSX manager nodes rejoin the cluster and sync.
Check out my YouTube video of the VMware NSX-T 2.4.1 Upgrade for a preview of what to expect: