Why we don't use your grandfather's storage for K8's?
I decided to write this blog because I get asked this question a lot. Why should I use Portworx, when I can use CSI from my existing storage array which is free?
This is a very common question from storage teams and infrastructure teams. Why do you need to pay for storage when I already have it? The answer is a bit complex with multiple rabbit holes. But let's look at this from the VMware, Storage, Core infrastructure teams' perspective.
From that lens, we are accustomed to things like VMware SRA (the site recovery adapter) for plugging storage into VMware SRM. A very mature tool for recovering a VM in its entirety from one data center to another, or from tools like synchronous replication. capable of having Active-Active data between 2 data centers.
I want this for my Kubernetes environment! I'm going to build it using the tools I already poses. This is a common discussion. Often it's revisited with. It's a lot of work we don't have resources to do this, or there are gaps that I can't seem to get over, or How do I back this up now?
This always brings me back to the point of data management. You have the data. No question your array can store it and snap it and migrate it. But being able to manage it in such a way that can make your K8s application portable is the challenge. For this, we need data management as close to the k8s control plane as possible. We need to look at objects in k8s and treat data as another object.