Cloud Native Computing Foundation KubeCon + CloudNativeCon Europe 2020 - Virtual, 15 Sep 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Sponsor Demo: Dell Technologies - Enterprise Data Management and Protection for Kubernetes Workloads

Description

Dell EMC Enterprise Storage and Data Protection solutions come with all the required Kubenernetes integrations for workload deployment, data management and protection as well as migration across the hybrid cloud. In this three-part demo we will cover some key enterprise capabilities of the Dell EMC storage and data protection solutions for Kubernetes workloads. The first part covers how CSI snapshots can be used to repurpose production data for test and dev environments. The second part shows the deployment of a YugaByte database based application deployment using VVOLs on PowerStore with VMware Tanzu Kubernetes Grid. Finally, in the third part, we show complete application protection of a MYSQL workload with application consistency and use the backup copy to spin the entire app on a new Kubernetes cluster all on AWS.

A

In this three-part demo, we'll cover some key enterprise capabilities of the dell emc storage and data protection, solutions for kubernetes workloads.

A

In software development, it is always a good idea to have test cases that are realistic in this demo, video. We are going to take a look at how you can take the copy of production data of a containerized application and use it to spin up a test or a dev instance of the application so that the application can be tested with the latest production data.

A

The storage platform here is powermax with the container storage interface. Kubernetes is the container orchestration platform and we are using gitlab for the devops automation.

A

Here is a sample application that we forked from github. It is a bare bones web app to manage a list of to-do items with sqlite as the backend database to store the to-do item. Information- and here is where the magic is happening. This is the persistent storage claim for the sql lite part. That is part of the to-do application.

A

Here we are actually using a tag variable of the application instance to dynamically allocate the right storage volume. If the tag indicates that its production instance then a claim against the production volume of the storage class sc and in a little while uh we'll see how to pass power max as the value for sc. If it's not a production instance, then a snapshot of the production volume is used instead, if the snapshot pv doesn't exist, kubernetes creates a snapshot of the production, persistent volume and then uses it.

A

This way, every non-production instance will have its own persistent volume. That is actually the latest copy of the production data, and here is the yaml for the snapshot. Definition. The storage class that will be used here will be max dash snap class. Of course, sc is the variable here and we'll see how to pass power max for the variable sc.

A

Let's see this in action here is the production version of the app. Let me add a few items that are, in fact a must in today's world.

A

Now, let's pin up a dev version of this application to do some changes to the app before we do that, a quick check on pvcs and pvs in the namespace.

A

You can see the persistent volume claim and the persistent volume for the production instance, and there are no snapshot volumes here cool now I create a new branch called colors and the change I'm going to make is to change the color of the title on the web page.

A

I can do a quick div to make sure and then commit and push the change. This creates a new pipeline in gitlab. Let me navigate to the.

A

Build- and here is the helm command for the application deployment in kubernetes, where we set things like a unique namespace for the instance and passing relevant variables. Of course, the variable sc, which stands for storage class is being set to powermax.

A

The build and deployment are completed, and here we can see that the color of the title has been changed and if you notice the data is the latest production data. Since it's a snapshot of the production volume.

A

Now we can go back and check to see that a new snapshot volume is created and we can also see the corresponding persistent volume and volume claim in addition to the production objects.

A

I can also check the unisphere environment to see that the snapshot has been created as part of the deployment within the same storage.

A

A

This same workflow can be used for any of the dell emc storage platforms listed here. These are basically the respective snapshot storage classes for each of the platforms.

A

In the second demo, we'll see how dell emc power store can be used to deploy containerized applications on vmware's tanzu kubernetes grid.

B

Here we click on create a new namespace. We select our wcp cluster and specify the name and then just click create. We can see that the namespace creation completed successfully and the kubernetes status is active at this point. I'm adding the dell emc powerster storage policy- I just created to our namespace.

B

This will allow me to create the tkg persistent volumes on my dell emc parse store vivo data store.

B

Next, I'm logging in to my v3 kubernetes, using the kubectl command, I'm authenticating with the administrator at vsphere local user.

B

Then I am switching to the new namespace. I just created using the kubectl config use context command by running ctl get sc. We can see the firestore policy. This appears as a storage class within the namespace.

B

As you can see, I'm using this storage class in my manifest file of my tkg cluster, this storage class is used for both the control plane and the worker nodes. You can also see the namespace and the tkg name. Now, I'm just going to create it. So now we can go back to the vsphere ui and see what's actually happening in our namespace.

B

We can see that the new tkg cluster has been created and the control plane vm has been built out but hasn't been powered on yet I'll speed things a little bit as this process takes about 10 minutes to complete.

B

By navigating to nsxt, we can see that a tier 1 gateway is being created for the namespace, as well as a new load. Balancer.

B

Now, let's connect to our power, store cluster and navigate to the vivos tab. Here you can see that we have multiple vehicles and each of them belongs to a particular tkg vm.

B

The type of evolves provisioned depends on the type of data that is being stored, config vivo stores, standard vm configuration data such as vmx log and nvram files.

B

Data vehicles store data such as vmdks snapshots and clones, swap vivo stores a copy of the vm memory pages when the vm is powered on now, let's log in to the guest tkg cluster, I just created by specifying the tenzu kubernetes cluster name and namespace.

B

We can also take a look at the nodes to make sure we are in the right context. Here we can see the single control, plane, node and the three worker nodes.

B

Now, let's create a stateful application inside the tkg cluster. For the purpose of this demo, I am deploying an application called yoga by db, which consists of 3 master nodes and 3 worker nodes. Each has a persistent volume claim, which will represent a single virtual volume.

B

We can see that within a few seconds all pods are up and running. We can access the application ui by navigating to the service, address and specify the application port.

B

Persistent storage is presented through the vmware csi driver called cns cns uses existing storage options for storage provisioning in a new way. First, it is based on storage policies, as we saw before, and, furthermore, it uses first class disks instead of standard disks.

B

A

Emc power, protect data manager and power, protect dd series appliances are the industry leading data protection, software and purpose-built hardware? You can deploy power, protect data manager, capabilities for your kubernetes environment to protect complete applications with the benefits of application, consistency and the most efficient data reduction.

A

You can do this both for on premises as well as for a public cloud deployment of kubernetes.

A

For this demo, we have chosen a kubernetes deployment in aws to show the data, protection and management workflows. Our sample application is a my sequel namespace for a wordpress blog.

A

Let's take a quick look at the app namespace that we are protecting and you can see the details of the parts, pvcs, config maps, etc for the application.

A

To enable application consistency, we introduced the concept of application templates.

A

We are already shipping templates for mysql and mongodb and more will be added in the future, and you can also modify these or create new ones yourself, based on the requirements of your database in the application template yaml file, the required calls for queries in the database are under pre-hooked section and the unquizzing calls are under the post hook section.

A

The nice thing about these templates is that they run agentless, meaning the hooks do not require any agent to be pushed onto the production cluster. To make these questing and unquestion calls to use this template for our application. We'll also make sure that we are using the right application label.

A

Now, let's create an app consistent protection plan for this mysql application, we'll go to the power, protect data manager ui and create a protection policy for the app using the policy.

A

Wizard select kubernetes for application type.

A

Leave the default crash consistent option since the application consistency is coming from the template, yaml file that we already assigned to the application using the appropriate label.

A

Here we select the mysql app that we want the policy to.

A

A

Under schedule, we specify the backup frequency and retention.

A

A

Now let us go ahead and initiate a backup in the backup now wizard using the policy we just created.

A

Once the job is done, we can check the new copy of the application created on the data domain target.

A

Now let us see how we can use this copy of the application deployment to spin it up on an entirely new kubernetes cluster. Let me go back to the data manager ui and initiate a restore using the asset, restore wizard.

A

You select the cluster to restore the app to, and here I'm selecting the cluster called target for the app to be restored.

A

And here I'm selecting the entire application to be restored and not just the database pvcs you can restore to one of the namespaces that are already available on the cluster or can specify a new namespace for the restored app.

A

Pvcs is already included for restore, I continue and on the summary tab, you can check the details and hit the restore.

A

A

Once the restore job is done, we can go back to the command line of this new cluster and check the newly created namespace and all the objects of the application's stateful set the parts, pvcs, config maps etc.

A

A

A