One of the biggest questions I had when I first started diving into Big Data Clusters was, “What about licensing….how will that work?” With so many different instances running on the storage pool, data pool and compute pool nodes will licensing cost too much? The answer I got from Microsoft was that it will “be competitive”.
Yesterday was the release of SQL Server 2019 CTP 3.2. The biggest change in CTP 3.2 is that Big Data Clusters is now in public preview. That means anyone can go download and deploy it. Prior to CTP 3.2, you had to sign up for the “Early Adoption Program”, wait until you received an email with your Docker credentials, etc. With CTP 3.2, Microsoft has actually done away with Docker credentials. You no longer need that to create your Big Data Cluster as the images needed are on Microsoft’s public repo.
Yay!! SQL Server on Windows Containers! Hey wait a minute, I thought you could already install Docker Desktop on Windows? Yes, but behind the scenes, Docker uses HyperV to create a Linux VM (called MobyLinuxVM). So even though it was on a Windows machine, it still used a Linux VM on the backend.