WilliamLam.com

VCF 9.1.1 - OCuLink External Graphics (eGPU) Dock with integrated PSU for GPU Passthrough with VKS

by // 1 Comment

For those who have been following my VCF Private AI Services (PAIS) exploration with deploying my own AI models and using them with Coding Agents running on VCF, I have found that these AI workloads can really push the thermals on my NVIDIA RTX 4000 running inside my Minisforum MS-A2, eventually causing the GPU to start throttling.

Earlier this year, I had shared a solution from a colleague who wanted to leverage a more capable GPU with vSphere Kubernetes Service (VKS) and was able to externalize the GPU using an OCuLink-based eGPU dock that is also compatible with the MS-A2.

I was initially going to use the same eGPU solution, which would have also required purchasing a separate PSU to power the GPU. While looking at my options, I came across another OCuLink eGPU dock (AG01) from Aoostar that had an integrated PSU (up to 800W), was plug-and-play ready, and certainly had a much cleaner aesthetic design.

[Read more...]

VCF 9.1.1 - Single ESX Host for VCF Fleet, VCF Workload Domain or VCF Cluster

by // Leave a Comment

For a production deployment of VMware Cloud Foundation (VCF), there is a good reason why you might not want to run everything on a single ESX host. 😅

Assuming you understand how the minimum ESX host requirements are derived for VCF deployments, there are certainly benefits to deploying a resource-optimized VCF environment for lab or proof-of-concept (POC) purposes. Being able to exercise workflows such as creating a new VCF Workload Domain or adding a new VCF Cluster with just a single ESX host can be extremely useful.

In fact, this has been possible since VCF 5.x and is something I have personally used to quickly stand up a full VCF environment or expand an existing one, which I have documented over the past few years:

While the scenarios above are the most common use cases for a single ESX host, deploying a new VCF Cluster with just one host has been a less common request, but it has also been possible.

To be able to deploy a new VCF Fleet, Workload Domain or Cluster using just a single ESX host requires applying a workaround, which I have captured in my comprehensive VCF 9.1.x Installer and SDDC Manager Configuration Workaround for Lab Deployments blog post.

[Read more...]

Bypassing Minimum Unique Flows & Metric Collection Period for vDefend Security Services Platform (SSP) 5.2 Lab Deployments

by // Leave a Comment

After successfully deploying vDefend Security Services Platform (SSP) 5.2 in my VCF 9.1.1 environment, I came to learn that SSP requires at least 1,000 unique flows before it can generate a security assessment of your environment.


Similarly, SSP also requires a minimum of three days of collected flow data before it can generate a flow capacity assessment.


If you have a small lab environment like myself, generating 1,000 unique flows can be difficult, and having to wait several days to demonstrate certain SSP capabilities may not be ideal, especially for proof-of-concept deployments. Here is how you can work around these default requirements when deploying SSP in a lab environment.

Disclaimer: This is not officially supported by Broadcom, please use only in a non-production environment for lab and learning purposes.

[Read more...]