After installing the nested ESXi, I had to install VCSA. As mentioned initially, I ended up deploying VCSA over the CLI, hosting it within my physical ESXi. Essentially, I end up with two separate VMs – the nested ESXi and the VCSA.

Before continuing, keep in mind the following:

  1. The deployment scripts and templates (see below) are located on the CD (iso image)
  2. Since I will deploy from my Mac, to speed things up a bit, I’ve copied the CD content to my local SDD drive

I – Get the portgroup and datastore names

When deploying VCSA using the CLI I will need to pass specific parameters to the scripts enabling the actual deployment. These parameters are all compiled into a .JSON file (Java Script Object Notation) which is passed further to the deployment script.

In my case, there are two parameters I needed to know the name of exactly, as I would be referring to them in the deployment template file:

  1. the portgroup name of the management network (deployment network)
  2. the datastore which will host the VCSA .vmdk files

For these I’ve issued the following commands on my physical ESXi host:


II – The Deployment Template

VMware was kind enough to provide built-in base templates which we could adapt to our own needs. These can be found here:


Since I will be doing an embedded installation, I’ve picked and adapted the template named embedded_vCSA_on_ESXi.json

This is how my template looks like:


At last, I’ve renamed the file to vcsa-deploy-ph.json

III – Start the Deployment

  1. On my mac, I opened a Terminal window
  2. Change directory to the location where the main deploy:
  3. Issued the following command:
    ./vcsa-deploy -v –accept-eula /Users/rafael/Downloads/vcsa-deploy-ph.json

Next you should start seeing some activity on the screen, including a progress update as below …


… and after a while, you will see (hopefully) something similar to the following:


Note: This process takes a while to complete – just be patient. If it does fail, you will have to be mindful at all times of the stage at which the failure occurred since there won’t be any automatic rollback.

For example, if the process fails after creating the VCSA VM, if you try to restart the script, you will get an error stating the VM already exists. You will have to first destroy the VM before trying again.

At this stage, refresh your vCenter 5.x view and you should see the VM listed and started:


You should also be able to ping it:


IV – First time connection to VCSA v6

At last, you will have to finish off your deployment by adding your host to VCSA, license the host and the vCenter, etc.


To login, use the url http://<VCSA-IP-Addr>/vsphere-client/ and the username administrator@<your domain name> – as specified in the deployment template file.

You could also login as root – though since the root account does not have the right privileges by default, you will see an empty inventory!

To wrap up, here is a screenshot of the new VCSA v6, with a Datacenter and the nested ESXi:


Thank you,
View Rafael A Couto Cabral's profile on LinkedIn

Comments are closed.