Edge upgrade or new deployment is in a Failed state

Products

• VMware NSX

Issue/Introduction

• A newly deployed Edge has a Configuration State of Failed with an error

Host configuration: Failed to send the HostConfig message. [TN=TransportNode/<UUID>]. Reason: Mac address for a vnic fp-eth0 is not found on edge node /infra/sites/default/enforcement-points/default/edge-transport-node/<UUID>.

• An Edge upgrade to 3.2.0 or 3.2.0.1 fails with the following error

Edge 3.2.0.0.0.19067070/Edge/nub/VMware-NSX-edge-3.2.0.0.0.19067089.nub switch OS task failed on edge TransportNode <UUID>: clientType EDGE , target edge fabric node id <UUID>, return status switch_os execution failed with msg: An unexpected exception occurred:  <......snip.....>  ERROR: NSX Edge configuration has failed. 1G hugepage support required\n"

• The Edge VM runs on an ESXi host with an Intel CPU generation type of SandyBridge, IvyBridge or Westmere.

Environment

VMware NSX-T Data Center 3.2.x

 

Cause

Previous versions of NSX-T Data Center supported 2MB and 1GB hugepages, however starting from version 3.2.0 Edge VMs support only 1GB hugepages.
On Westmere, Sandy Bridge, and Ivy Bridge generation Intel CPU, ESXi does not expose 1GB pages by default.

Resolution

This is a known issue impacting NSX-T Data Center 3.2.0 and 3.2.0.1.

This issue is resolved in VMware NSX-T 3.2.1 (build number 19801959).

Workaround:
A configuration change must be made to the Edge VM to allow it use 1GB hugepages.

1. Power off the Edge VM
2. On the vSphere Client select Edit Settings
3. Select the VM options tab
4. Expand Advanced and select Edit Configuration
5. Click Add Configuration Params
6. Add featMask.vm.cpuid.PDPE1GB = Val:1 
   
7.  Save the settings
8. Power on the Edge VM

Note:

In the case of an upgrade failure, retry the upgrade once the configuration change is complete.
Alternatively, the Edge VM can instead be moved to an ESXi host with a newer generation CPU which by default allows 1GB hugepages.