VMware Lightboards - Cloud Native Fundamentals, 1 Dec 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Persistent Container Storage with Project Hatchway

Description

This video will guide you through the basics on persistent container storage with project Hatchway (vmware.github.io/hatchway)

For more information on VMware's Kubernetes based solutions, please visit VMware Pivotal Container Service (PKS) website:
https://cloud.vmware.com/pivotal-container-service

A

Hi, my name is steve juggler, I'm a director of systems, engineering in VMware's cloud native apps business unit, and this quick light board session is all about persistent container storage and we do that at VMware with an open source project called project hatchway, you can see up here. I've got the VMware github io / hatchway, which goes through the details of how to actually configure and do this in your VMware environments. Now most of us know that a lot of times containers are seen as being used for stateless workloads. The reality is with this project.

A

Hatchway we've got a couple of different options to actually maintain some state and use persistent volumes with container workloads. So I've got two examples. Here are pre-written to save some time here. I've got our standard docker host environment. So if you have a docker host virtual machine running in vSphere, this is something you can do it. Oh sorry, vSphere specifically just ESX for that matter, free version of ESX this works, but in effect it's a regular docker host VM.

A

So what what makes it a docker host is the fact that we've got the docker engine running okay. So, as an administrator there's a couple of administrative things, I need to do up front to get this ready. The first is that I need to install the vSphere.

A

Docker volume service vibe into ESX, so do this the way you normally install vibs just get that installed. The next administrative tasks before we're ready is that we need to basically install something in the container host called the vSphere volume plug-in so vSphere volume plug-in and now we're set up and ready.

A

Once we get those two configured, then, if I have a user that wants to create a container and in turn wants to create a volume or persistent volume with that container through the docker CLI or API, they can issue a request to create a persistent volume, and so what we do, then? What happens is that through the docker engine through the API, it calls the vSphere volume plug-in and then in turn calls the vSphere docker volume service and inside our data store inside ESX.

A

We end up creating a VMDK down here, okay and so that VMDK, then in turn, is presented up to the actual docker host. We've got our persistent volume here and we can connect that up. So you can iterate on your containers as much as you want. You can always go back to that. Persistent storage. That's going to exist, so that's our standard, docker, host environment, second environment here is when we have a kubernetes environment. Okay, it's a little bit different.

A

The big thing you'll probably see here is that it isn't just ESX it's it's actually vSphere, okay and the the difference is that we're going to be interacting with vCenter to provision some of these persistent volumes? Okay. So what are the? What are the components that are required here? Well, in kubernetes, we have something running called a cubelet, and that is the thing that interacts with the kubernetes cluster services, so to speak, we've got on top of those cluster services.

A

We've got an API, ok, so we've got the the cubelet in in the node. We've got the API. The second thing we actually need in the node here is the kubernetes vSphere cloud provider, and so what that piece does is that that will in turn go talk to V sender. So, let's walk through this there's actually a couple of different ways. You can do persistent volumes and kubernetes.

A

For the purposes of this, I'm going to talk about a situation where a user wants to create their own volume or position volume claim, and so this user intern will go through the kubernetes api will issue. This persistent volume claim and much the way it happens over here through the api and the couplet we're going to go to the the kubernetes vSphere cloud provider. That's going to call vCenter down here.

A

The center will in turn direct the creation of again another VMDK down here in the datastore and then that VMDK is in turn mapped up to a kubernetes volume.

A

Okay, so now I run my pod and I have persistent storage associated with that pod. So that's the basics. With with kubernetes, persistent storage, there's, actually quite a few topics related to kubernetes and storage. So I mentioned a persistent volume claim, there's also just persistent volumes. If they've been pre-created for you, the persistent volume claim which is creating them. There are storage classes and there's also these things called stateful sets which is encompassing a lot of the storage configuration so that it's all-encompassing.

A

So there's these four other topics which we could probably do some additional light boards on. So look look to those in the future. Thanks for watching.

A