Build a VMware Cloud on AWS Content Library using AWS S3

Gilles Chekroun
Lead NSX Systems Engineer - VMware Europe

What is a Content Library?

DISCLAIMER: External content libraries are not supported by VMware GSS !

vSphere Content Library is a repository where you can store VM templates, vApp templates and other kind of files like ISO. Administrators can use the templates to deploy VMs and vApps. Each item library can contain multiple files. The basis is .ovf together with the description files like .mf and .vmdk
If you have .ova files, you can simply un-compress them using the linux command: 
tar -xvf SomeFile.ova

There are two kind of Libraries: Local and Subscribed.

Local Library

Local Library stores items in a single vCenter. VM and vApp templates are stored as OVF file format.

Subscribed Library

You can create a Subscribed Library in the same vCenter and to ensure the content is up-to-date, the library can automatically sync with the source on a regular basis.
You can also use the option to download content from the source immediately or only when you need. this will help to manage local storage.
You can also delete local content (look at it like un-sync) to save space as well.

In this article, we will build the content in AWS S3 and use that as the source for publication.

Build the source

NOTE: For this first release, I will build the source from local computer and duplicate to S3. Clearly the goal is to create the source in S3 directly but that will need a few tweaks in the python script that creates the metadata indexing. Stay tuned for some updates . . .

File structure

The Content Library uses Open Virtualization Format (OVF). An OVF package consists of several files placed in one directory. There is ONE .ovf file descriptor and typically one or more disk images, certificates and other auxilliary files.

Note the directory structure with sub-directories for each template.

Create the metadata indexes

The Content Library subscription needs a lib.json file that will be created using the following python script: (Thanks to William Lam and Eric Cao).
Refer to William Blog here with full details.

We use pthyon 2.7 and the usage is:
python ContentLib <path-to-your-ContentLib>
This creates the needed indexes.

Duplicate to AWS S3

Get the "lib.json" URL
Make the Content Library bucket public

Create the Content Library

Open vCenter and select Content Libraries
Add a Library
Give it a name and click NEXT
Use Subscribed Library and paste the S3 URL of lib.json.
Select Download content when needed to save disk space.
Select storage
Review and click FINISH
Note that the new Library has 3 templates but 0 Bytes size. The content will sync on demand.
Open the Library. Note the 0 Bytes size.

Deploy new VM from Template

Give it a name and chose the Workloads Folder
Continue on a classic VM deployment steps and click FINISH
The VM deployment has triggered the synchronisation of the specific content.
VM is deployed
From here, we can power it on and go ahead.

Save disk space and Delete Library item

To save some space on the local vSAN we can "un-sync" items from our library.
The items remain in S3 obviously for future deployments.
At this stage, we are back to a zero size Content Library.



Egress VPC and AWS Transit Gateway (Part1)

AWS Transitive routing with Transit Gateways in the same region