New Mobile Home Lab Hardware..
Since posting a small teaser of the new home lab hardware on Twitter last week I've had more than a few enquires asking what the hardware configuration is of the fantastic SMC Super Servers, how I've got them configured from an infrastructure perspective and perhaps the most interesting, folks wanting to know what I'm going to be using them for.
Well firstly I've asked a guest blogger (Simon Richardson) to describe the hardware and infrastructure configuration in more detail, given his background and the questions from twitter.. my description of "Some speedy processors, lots of memory and lots of network ports" will not satisfy anyone curiosity. Look out for Simon's first blog on the hardware later this week.
As to what I will be using the lab for, well it probably comes as no surprise that the arrival and configuration of the new hardware coincides with the General Availability of vSphere 7 (with Kubernetes) and Tanzu Kubernetes Grid.
Check Here & Here for more information on each of those GA announcements.
It has been an incredible journey from the early days of vSphere Integrated Containers ( 2015 ! ) to where we are today with a fully integrated Kubernetes platform being part of vSphere and VMware Tanzu Kubernetes Grid being a strong upstream Kubernetes solution for vSphere and beyond, but I digress.
Back to the question at hand, what will I be using the new home lab and VMware Kubernetes solutions for, well quite simply a great deal from the software engineering perspective. The lab is sufficient to build and more importantly demonstrate some very complex Cloud Native application and developer scenarios, everything from migrating traditional Java (.. other languages exist too I'm told) applications to the Kubernetes platform up to some exciting machine learning and analytics solutions. From simulating large(ish) data models and apps to showing what is possible running Kubernetes at 'the edge' for things such as retail store and isolated environments.
My primary driver here is the software and Kubernetes layers; Simon is going to make sure the hardware and infrastructure can deliver.. and probably, more importantly, fix it when I inevitably I break things.
Running the virtual appliance across all the servers (32 cores) 24x7 for 6 days made for a great test of the hardware stability alongside vSphere & vSAN 7. Everything performed perfectly.
As part of the testing, I deployed the folding @ home application in two form factors.
The vApp virtual appliance configured as a large vApp consuming 8 cores (1 whole server ) across 3 of the physical hosts.
The final host is a Tanzu Kubernetes Grid K8S Cluster running the Folding @ Home Applications are a smaller k8s pod that I could quickly scale out and in allowing me some compute cycles to tinker.
If you would like to get involved and join the team, check out these links.
As we are on the subject, this repo is an example of how to deploy and use GPU's with K8S clusters - not something in scope for this project (yet) but exciting and would work nicely running something like DockerDesktop on a GPU enabled workstation. ( https://github.com/RobbieJVMW/k8s-fah )
With my developer head on I intend to re-configure the K8S app to be a little more Kubernetes friendly by storing the team ID and other config data inside the k8s cluster rather than coded into a config file .. it's on my //TODO.