PAIF 'AI Workstation' Deployment Fails Due to GuestBootstrap Error and NGC Authentication Failure

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

Deployment of the 'AI Workstation' catalog in PAIF fails with the error:
Failed due to status condition. GuestBootstrap: 'False', reason: 'Failure'.
The namespace creation completes successfully, but the deployment fails during the GuestBootstrap phase.

Environment

Aria Automation 8.18.x

Cause

The deployment failure is linked to a Docker login failure when the DL workload attempts to pull containers from nvcr.io.
The error from /var/log/dl.log indicates unauthorized access due to missing or incorrect credentials:
Docker Login Failed: Error response from daemon: Get "https://nvcr.io/v2/": unauthorized.

Resolution

Ensure the correct credentials are used for Docker login (correct credentials are to be used as input in the catalog request forum)

Username: $oauthtoken
Password: The valid NGC API key

Additional Information

The journal logs also referenced an NVIDIA Virtual Applications error:
Error: No pool features found, which may be related to GPU resource configuration but was secondary to the main authentication failure.
Key log files reviewed:

/var/log/dl.log
/var/log/vgpu-install.log