We are not stopping with the integration around and above Kubernetes To improve competitiveness in a fast-paced world, organizations are embracing software as a prime differentiator. A software-driven business is […]
We are not stopping with the integration around and above Kubernetes
To improve competitiveness in a fast-paced world, organizations are embracing software as a prime differentiator. A software-driven business is more agile and efficient, innovates faster and responds dynamically to changing market and customer demands. Increasingly, gaining these compelling advantages requires adopting cloud-native applications.
Kubernetes is the de facto standard for container orchestration for microservices and applications. However, enterprise adoption of big data and databases using containers and Kubernetes is hindered by multiple challenges, such as complexity of persistent storage, availability, and application life-cycle management. Kubernetes provides the agility and scale that modern enterprises require. However, Kubernetes provides the building blocks for infrastructure, but not a turnkey solution.
Originally containers and Kubernetes were optimized for stateless applications, restricting enterprises to use container technology for workloads, such as websites and user interfaces. More recently, stateful workloads have been increasingly run on Kubernetes. Whereas it is relatively easy to run stateless microservices using container technology, stateful applications require slightly different treatment.
There are multiple factors which need to be taken into account when considering handling persistent data using containers, including:
- Containers are ephemeral by nature, so the data that needs to be persistent must survive through the restart/re-scheduling of a container.
- When containers are rescheduled, they can die on one host, and might get scheduled on a different host. In such cases, the storage should also be shifted and made available on a new host for the container to start gracefully.
- The application should not have to deal with the volume/data and underlying infrastructure, and should handle the complexity of unmounting and mounting, etc.
- Certain applications have a strong sense of identity (such as Kafka, Elastic, etc.), and the disk used by a container with a specific identity is tied to it. If, for any reason, a container with a specific identity becomes rescheduled, it is important that the disk specifically associated with that identity, gets reattached on a new host.
In this white paper, we present a solution that provides the robust capabilities and built-in high availability and scalability of XtremIO and PKS, bringing simplicity to Kubernetes, and creating a turnkey solution for data-heavy workloads. XtremIO CSI Plugin provides Pivotal PKS technology with built-in enterprise-grade container storage and uncompromising performance that extends Kubernetes’ multi-cloud portability to the private cloud.
With this proven integration, organizations can accelerate their digital business objectives via cloud-native applications, from development through testing and into production, on a single platform with ease and agility. Most importantly, eliminating the risk of application downtime preserves critical revenue streams, helps ensure customer satisfaction and strengthens your competitive advantage.
Planning, designing and building a private or hybrid cloud to support cloud-native applications can be a complex and lengthy project that does not address immediate business needs. IT also must maintain and manage the infrastructure that supports these new applications, to ensure that the environment is reliable, secure and upgradeable.
Kubernetes offers great capabilities. However operationalization of these capabilities is not a trivial task. There are many layers involved, such as hardware, operating system, container runtime, software-defined networking, and Kubernetes itself. And all the layers require Day 0, Day 1 and Day 2 activities. Most of these layers also require constant patching (for example, Kubernetes ships a new version every three months), so the effort to maintain this infrastructure grows exponentially.
Kubernetes is the de facto standard for container orchestration platforms. However, Kubernetes can be difficult to operationalize and maintain for the following reasons:
- High availability — There is no out-of-the-box fault-tolerance for the cluster components themselves (masters and etcd nodes).
- Scaling — Kubernetes clusters handle scaling the pod/service within the nodes, but do not provide a mechanism to scale masters and etcd Virtual Machines (VMs).
- Health checks and healing — The Kubernetes cluster performs routine health checks for node health only.
- Upgrades — Rolling upgrades on a large fleet of clusters is difficult. Who manages the system it runs on?
- Operating systems — A Kubernetes installation requires multiple machines where an OS is installed. Installing and patching these systems is an overhead.
Pivotal Container Service (PKS) is a purpose-built product that enables enterprises to deploy and consume container services with production-grade Kubernetes, built with high availability, security, multi-tenancy and operational efficiency across private and public clouds.
It effectively solves the complexities around Kubernetes infrastructure and enables you to focus on your business applications. Furthermore, it benefits your developers, providing them with the latest version of Kubernetes and simple integrations with monitoring and logging tools.
With Pivotal Container Service (PKS), not only do you receive a vanilla Kubernetes distribution supported by Pivotal and VMware, but you also receive the automation to maintain the Kubernetes stack with reduced effort, including the operating system, container runtime, software-defined networking and Kubernetes itself. The automation provided by BOSH enables you to install (Day 0), configure (Day 1), patch and maintain your Kubernetes infrastructure in a healthy state over time (Day 2).
This automation also extends your capabilities to have an effective Kubernetes as a service offering, and lets you create Kubernetes clusters with a single command, to either expand or shrink the clusters, or to repair them as needed. This is a unique capability that enables you to provide your Kubernetes users with as many clusters as they need, and to resize them, leveraging automation as your workload requirements change.
PKS also simplifies the networking configuration, by providing you with VMware NSX-T, which is a software-defined network infrastructure to build cloud-native application environments.
PKS also provides you with a container registry to host your containers, namely, Harbor, a graduated CNCF project.
- Pivotal PKS Components
Dell EMC XtremIO X2: The Next Generation All-Flash Array
Dell EMC XtremIO is a purpose-built all-flash array that provides consistent high performance with low latency; unmatched storage efficiency with inline, all-the-time data services; rich application integrated copy services; and unprecedented management simplicity.
The next-generation platform, XtremIO X2, builds upon unique features of XtremIO to provide even more agility and simplicity for your data center and business. Content-aware in-memory metadata, and inline all-the time data services have made XtremIO the ultimate platform for virtual server and desktop environments and workloads that benefit from efficient copy data management. The architecture offers VMware XCOPY operations that are twice as fast, which can reach as high as 40 GB/s bandwidth, due to them being in-memory operations.
X2 includes inline, all-the-time data services, thin provisioning, deduplication, compression, replication, D@RE, XtremIO Virtual Copies (XVC), and double SSD failure protection with zero performance impact. Only XtremIO’s unique in-memory metadata architecture can accomplish this.
This white paper is designed for developers, system architects and storage administrators tasked with evaluating or deploying Pivotal PKS, and interested in persistent storage within their deployment. While this document focuses on Dell EMC XtremIO X2 products, the concepts discussed and implemented can be applied to other storage vendor products.
Understanding the material in this document requires prior knowledge of containers, Kubernetes, PKS and of the Dell EMC XtremIO X2 cluster type. These concepts are not covered in detail in this document. However, links to helpful resources on these topics are provided throughout the document and in the References
section. In-depth knowledge of Kubernetes is not required.
Click the screenshot below to download the white paper.
You can also watch the below demos to see how it all works