Error while deploying OVF package 'Failed to attach vif uuid to network'

Products

• VMware Cloud on AWS
• VMware vSphere with Tanzu

Issue/Introduction

How to remove stale network entry from OVF templates.

Symptoms:

Failed to attach vif uuid to network nsx.LogicalSwitch:#######The requested object : LogicalSwitch/####### could not be found. Object identifiers are case sensitive

VmReconfigure: adding opaque network vnic

In vCenter VPXD:

2023-03-30T10:15:04.967Z info hostd[2113535] [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/vsan:########x/Win2012_New.vmx opID=###### user=vpxuser:VMC.LOCAL\vpxd-extension-#########] Reconfigure failed: N3Vim5Fault19PlatformConfigFault9ExceptionE(Fault cause: vim.fault.PlatformConfigFault

 

Environment

VMware Cloud on AWS 

VMware vSphere with Tanzu

Cause

Stale network segment in OVF template.

Resolution

Editing the templates to point to the correct segments should allow the vCenter to clear these stale network references
Convert the template to VM, edit it virtual hardware then convert t back to template.

Additional Information

Impact/Risks:

Unable to deploy OVF Template

 

This behavior was also observed in vSphere with Tanzu environments:

OVF deployment from content library item implementation plumbs random network to the VM as it does initial deploy. In this case, the network chosen happened to be a Stale/missing NSX network that was deleted from the NSX, but did not get cleaned from VC. 

Resolution for environments with vSphere with Tanzu:
This stale NSX network was removed from the VCDB directly as attempts to do so via the vSphere API (MOB) did not work (task succeeded but network not removed). Once this stale network was removed from the VC, this unblocked the CL OVF deployment flow.

The controllers running on the SV were able to make forward progress towards the desired state and the clusters came back healthy as expected.