Cloud Native Computing Foundation KubeCon + CloudNativeCon Europe 2021, 15 May 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Sponsored Session: Commvault - Many Apps. Many Locations. One Storage Solution.

Description

Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon North America 2021 in Los Angeles, CA from October 12-15. Learn more at https://kubecon.io The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.

Sponsored Session: Commvault - Many Apps. Many Locations. One Storage Solution.

Speakers: Abhijith Shenoy, Xiangyu Wang

Enterprises are moving their applications to the cloud at an unprecedented pace. The pandemic has accelerated this journey and hybrid-multi cloud is now a foundation of modern enterprise architectures. Along the way, you’ve become comfortable optimizing for where, when, and how your applications run. So, why wouldn’t you be equally comfortable using distributed storage that:

• Provides location transparency for all your stateful applications
• Runs anywhere SDS for predictable enterprise-grade resilience, scale, and performance delivered in an application-aware model
• Integrates natively with K8s and your existing workflows

Containerization promises application portability, yet legacy infrastructure relies on traditional snapshot and replication technology to migrate data across locations, becoming a bottleneck for modern DevOps environments.
With distributed storage, you can provision cloud-native, programmable storage when you need it with self-service access and know that your data is instantly available across multiple locations (edge, core, clouds) using our distributed architecture. In this session, you will see distributed storage in action in a K8s environment and learn how to:

• Easily setup and manage persistent volumes using our new Commvault Distributed Storage Operator
• Establish enterprise-grade high availability for stateful applications across regions and zones
• Benefit from the simplicity and efficiency of sharing the same storage across multiple OpenShift and K8s clusters

Modern solutions don’t come from bolt-ons to legacy technology – they come from a foundational architecture rooted in multi-hybrid cloud distribution.

A

Hello all and welcome to the kubecon presentation of commvault distributed storage for stateful containerized applications. My name is abjet and I am a distinguished engineer at commvault joining me today is shaggy wang, a principal engineer at commvault who will demonstrate the key capabilities we have developed for container ecosystems.

A

Storage can sometimes be seen as an obstacle adding overhead and complexity to application workflows, but data is fueling applications and persistent volumes. Allow for stateful applications to become mainstream. Disputed storage is changing the way we think about storage, with a focus on data wherever it lives. We do this by centering on three key architectural principles: one an agile infrastructure that unifies block file and object, storage in a single platform, two location, transparency of data and three fully programmable infrastructure to seamlessly integrate with your existing workflows.

A

All of this, without compromising on enterprise grade resiliency security, governance, scale or performance comma distributed storage is a software defined storage technology that supports a breadth of workloads. While I would love to talk more about this, the focus of this presentation is limited to how we deliver on this promise for container ecosystems.

A

Let's take a look at how combat distributed storage adds value to your kubernetes environments. First, let's talk about simplified, install and healing. We recently launched our official operator to manage all our storage components and kubernetes and openshift clusters. This includes day one installs and seamless upgrades in a predictable way that fully aligns with how you've been managing your kubernetes clusters. Already. Our solution is programmable, with kubernetes apis and industry, standard orchestration and visualization tools.

A

Our solution is also csi compliant supporting both block and file storage, along with other capabilities such as online volume, expansion, snapshots and clones location, transparency of data for application, portability, commvault distributed storage cluster can span across racks data, centers, public clouds or even a mix of both private and public clouds.

A

A single distributed storage fabric is presented to the kubernetes clusters, with policy driven data placement to ensure that data is protected and available to your applications within the determined sla enterprise grade storage for stateful applications for security, commvault, distributed storage, fully integrates with any k-map compliant key management systems enabling users to bring their own keys. We provide storage efficiency through features such as in-line, global deduplication and compression, which can be dynamically configured through storage classes. We also support point in time, snapshots and clones for restoring old data or for test dev environments.

A

Before I hand it off to bang for a quick demonstration of the platform. Let me summarize what we will be showcasing in the demo. You will effortless you'll see how we can effortlessly manage the installation and upgrade of uh storage components in kubernetes using our enterprise operator. You'll also see how multiple kubernetes, as well as openshift clusters, can provision storage out of the same multi-site hedwig cluster through a single distributed storage fabric.

A

We will also showcase how you can provision application aware volumes with the desired level of fault, tolerance across racks and data centers with ease, and we will also finally showcase other capabilities, such as seamless expansion, snapshots and clone cloning of volumes with absolutely no application downtime throughout the demo, you may observe references to hedwig. Please note that we are actively undergoing a rename of hedwig to combat distributed storage. Having said that, I'm going to hand this off to bank so that he can quickly take you guys through a demo of all the capabilities.

B

Thanks rpgs hello, everyone in this part of presentation. I will show some demo about how commode enables stateful applications in kubernetes. For this demo, we have a combo distributed storage cluster with two disks on prime and one. This is on aws and we have one aws eks cluster, one on prime openshift cluster.

B

Let me first show how easily we can install hardware csi on openshift, using operator hadwig operator is listed on operator. Hub can be installed directly from openshift ui. Here, with one click on install, you will get it after install done. Go into the hardware operator.

B

You can easily create a hardware deploy which contains the detailed information about your command storage.

B

Cluster, I already create a yam fill for the multi-side cluster. We are using. The c's are part of the nodes in that cluster. As part of the hardware deploy object, the operator automatically deploys the csi driver and the proxies for managing both proxy, both block and nfs volume. In the rest of the demo, I will use oco cube cutter command to showcase how convoluted distributed storage is fully programmable through kubernetes apis. Now, let's create a elastic search application using command storage. Here is the client machine point to the openshift cluster to use command storage.

B

We need to first have a storage class in the storage class. You can choose the replication policy to decide which aware you want to place your data. How many copies you want to keep and also turn on some enterprise level feature like a duplication, comparison or encryption.

B

Here I already have a storage class for nfs replication policies, data center, oer and the data will be placed on all three data centers on the commode cluster. I already have a young field for the elastic search in this young field. You will use the storage class. We have.

B

And the mount the volume to elastic search data.

B

B

Okay, let's check out if the application is up and running by issue or call command to elastic search servers.

B

See we can get a response from.

B

B

Now, let's log into the container.

B

You see the converter, nfs storage is mounted to the elasticsearch data path.

B

We have seen how persistent volume data can be replicated across multiple set. Let's explore a scenario where you want to your data to be located within a single cycle for a pre-production or test use case and how we provide the flexibility to do so here we have an aws eks cluster on us west, 2 region, all the css driver and hardware proxy already set up here and the storage class for block already created.

B

It will only use the uswest tool data center to place the data. Now, let's create an application using this storage class in this young field, it will create a container and using the storage class we created before and a month the volume tool, postgres data.

B

B

Now, incoming ui, if you see all your data will be located only on the aws data center for this block.

B

B

A

The container is up and.

B

B

Let's go into the container.

B

You see the convoluted block. Storage is mounted to the post request data pass.

B

And the size for this processor volume claim is 10 gig. No, sometimes people might need to extend the volume as their data grows for commodore. We can easily do the results and you don't need to take down your application here. I will show.

B

You now the pvc is 10 gig. I will change it to 20.

B

B

You can see the pv is already resized to 20 gig now and if you come to the con motor ui. This is a block virtual desk that we created before for this poster glass container.

B

Now the container size is 20 gig.

B

B

Okay, now the pvc is also reset to 20 gig. Let's check the size inside the container, let's run the df7 kh. Now you can see within the container this block device also resize to 20 gig, and you don't need to take down your application totally and we can also create a snapshot for the volume. In case you need to go back your data very easily.

B

Here I have a young field to create the snapshot for the poster glass volume we created.

B

B

Now the snapshot is ready to use. Now, if you come to cambod ui, you can easily find there is a snapshot we just created for this block virtual desk.

B

And later on, you can't revert to this snapshot I'll, create a clone volume from it very easily. Let's now create a clone from it.

B

I have a persistent volume claim yum fuel which we'll use in the snapshot that we just.

B

Created now you can see we have a new processor volume claim which is 20 gig now and if you come to commod ui, you will find a new uh block device creator here.

B

The digital information will show it is the clone desk from the snapshot that we created before. That's all what I want to show here thanks everyone.

A

Thanks a lot wang that was a very smooth and wonderful demo for our audience. uh If you have any more questions or if you would like to discuss any particular feature or anything else further with us, please visit us at our virtual commvault booth feel free to also connect with us on our slack channel and to get more information about our complete offering for containers.

A

You can always visit www.commvault.com containers. Thank you all and have a nice.

A