Automated Nested Lab Deployment on SDDC Part 1: VMware Cloud on AWS

While preparing for this years VMworld, I had the unique opportunity to work across a number of VMware Cloud SDDC solutions such as VMware Cloud on AWS (VMConAWS), Azure VMware Solution (AVS), Google Cloud VMware Engine (GCVE) and Oracle Cloud VMware Solution (OCVS). During the the development of several demos, I found it invaluable to be able to deploy a Nested vSphere environment to validate my configurations prior to connecting our real vSphere on-premises infrastructure.

Putting aside VMworld demos, this can certainly be extended to other use cases such as accelerated pilots, proof of concepts and lab/development purposes. Customers have been leveraging Nested Virtualization technology for more than a decade plus now and it definitely makes sense that they would also want to do the same for certain workloads running within a VMware Cloud SDDC. With that said, Nested ESXi is not officially supported by VMware or on any other VMware-based platform.

As part of building my VMworld demos, I also had spent some time on creating some automation that would make it easier for me to re-deploy these Nested Lab environments and also being aware of the specific VMware Cloud SDDC solutions, so that I only have a single script to maintain. In case folks are also interested in being able to do this, you can follow this 4-part blog series which I will be kicking it off with VMware Cloud on AWS (VMConAWS).

Disclaimer: Nested ESXi is not officially supported on VMware Cloud on AWS or by VMware.

Prior considerations and network constraints still apply when running Nested ESXi on VMConAWS, for more details I recommend taking a look at this blog post I published in 2018. While network connectivity to the inner-VM workloads is not possible due to the lack of MAC Learning support, basic vSphere infrastructure operations are fully functional. To simplify the setup, I have also created a simple NFS Photon OS OVA (x86 at bottom of blog post) which is used for VMware Cloud SDDC which do not support the ability to run Nested vSAN on top of a physical vSAN infrastructure such as VMConAWS as an example.

Pre-Requirements:

1, 2 or 3-Node SDDC already deployed
Bastion / Jumphost which has network connectivity to the SDDC Management network. In my setup, I have deployed a Windows Server VM which acts as both a local DNS Server but also the Jumphost. You can follow this blog post to configure the NSX-T firewall to allow access through the Management Gateway construct
PowerCLI 12.x installed on the Bastion/Jumphost
Download the desired version of OVAs (vCenter Server Appliance (VCSA), Nested ESXi Appliance and NFS PhotonOS OVA)

Step 1 - Download the nested-sddc-lab-deployment.ps1 script and transfer that and the OVAs to the Bastion/Jumphost.

Step 2 - Update the script (details can be found on the Github repo) that reflects your environment. For those who have used my previous Automated Nested Lab Deployment scripts, this should feel very simliar. The only key difference is specifying the SDDC Provider ID which the script will properly handle the uniqueness for each respective VMware Cloud SDDC environment.

Step 3 - Once you have saved your changes, you can execute the script and a summary output as shown in the screenshot below will be provided prior to actually starting the deployment.

If everything was setup correctly, the script will take ~22minutes to deploy a fully configured VCSA with 3 x ESXi VM (default) and attached to our NFS VM to provide shared storage across the ESXi hosts.

If you have DNS configured and enabled in the script, you can then connect to your VCSA instance using the various CLI/API or the vSphere UI of the FQDN that you had specified for the VCSA. If not, then you would connect using the IP Address. You will notice that all VMs deployed as part of the script will be placed inside of a vApp construct.

Stay tuned for Part 2, 3 and 4 of this blog series which I will be covering the other VMware Cloud SDDC solutions.

More from my site

Comments

Thanks for the comment!Cancel reply