Latest Version in Kubernetes: 1.17 Release

Latest Version in Kubernetes: 1.17 Release

February 25, 2020 0 By Hoofer

Introduction:

Kubernetes is an open-source platform used to automate the operations of Linux containers. It eliminates the manual process of scaling and deploying containerized applications and also manages the clusters efficiently. Those clusters host public and private clouds; hence Kubernetes is referred to as an ideal platform to host cloud-native applications for rapid scaling. Google engineers designed Kubernetes. Google was an earlier contributor to Linux container technology, and it can generate more than 3 billion deployments of Linux containers in a week.

Features of Kubernetes:

Here are some of the Kubernetes features:

  • Capabilities of self-healing
  • Load Balancing and Horizontal Scaling
  • Loose coupling of Infrastructure to each component
  • Resource utilization has a high density
  • Centric-Application Management
  • Predictable Infrastructure creation
  • Automated Scheduling
  • Automated rollback and rollouts
  • Consistent development in production and testing
  • It has Enterprise-ready features
  • Auto-scalable Infrastructure

Advantages of Kubernetes:

Here, are some of the advantages of Kubernetes:

  • It can arrange containers across many hosts.
  • Automate and control application deployments.
  • Scale containerized applications.
  • Self-heal the applications with auto-replication, auto-restart, auto-scaling, and auto-placement.
  • Better usage of hardware to increase resources.
  • Add storage to run applications statefully.
  • Manage services to deploy applications.
  • The containers in Kubernetes perform the updates without service interruptions.
  • Kubernetes run with on-premises OpenStack, Microsoft Azure, Public Clouds Google, Amazon Web Services (AWS) and so on.
  • Master, Cluster, and Node are the important basics of Kubernetes.
  • Replication centres, Deployments, Replication Controllers, Daemon sets, Stateful sets are some of the important terms in Kubernetes.
  • Kubernetes is a good example of the Container Management System.
  • It also offers self-healing and automated scheduling.
  • Kubernetes allow auto-scaling.

Kubernetes 1.17 Release:

The latest version of Kubernetes: 1.17 was released last month (December-2019). It consists of 22 enhancements, out of those, 14 are stable, 4 in alpha and 4 in beta. Kubernetes 1.17 version improves scalability in the cluster network. Also, it introduces the pod schedule, cluster storage, and runtime options for custom resource support.

New Features in Kubernetes 1.17:

  • In Kubernetes Certification 1.12, volume snapshots were introduced in alpha, but now Kubernetes version 1.17 promotes those volume snapshots to beta. This advanced feature can snapshot the cluster in a moment. Those snapshots can provide a new volume with old snapshots’ data. This helps to elaborate code-version and data-version operations.
  • The default storage plug-ins are moved to CSI (Container Storage Interface), and it leads to independence on drivers for the core version.
  • Earlier in Kubernetes 1.12, cloud provider labels were introduced only in beta. But, now it is available in both beta and alpha because of Kubernetes 1.17 version.

Volume Snapshots in Kubernetes:

Most of the storage systems like Amazon Elastic Storage, Google Cloud Disk create persistent volume snapshots to represent a point time in volume copy. This snapshot is used to restore the existing volume to the previous state. Kubernetes volume plugin provides an abstraction to automate the provision and attachment of storage file. Kubernetes also creates an abstraction layer between underlying systems and distributed application systems to agnostic the cluster specifics and application deployment. SIG, which is a Kubernetes storage, identifies critical functionality of snapshot operations for stateful workloads. Kubernetes users can handle use cases without API by providing trigger snapshot operations. At present, Kubernetes users empowered snapshot operations with cluster regardless of storage underlying. Also, Kubernetes snapshot acts as building blocks to develop storage administration and enterprise-grade features with backup solutions.

CSI Migration Beta:

Kubernetes provides volume plugin systems in which “in-tree” was code for the core Kubernetes. As we know, to add new volume, support was a tough challenge. Vendors wanted to support the storage system to Kubernetes, but it was not possible. Besides, the storage in the third-party code causes security issues and reliability issues in core binaries and the code was complicated to test. Later on, Container Storage Interface solves those issues. As there were many CSI drivers, the production became ready for the CSI model benefits. However, we didn’t force users to make the workload changes by breaking available API storage. The only possible way is to replace “in-tree plugin” with CSI.

CSI(Container Storage Interface) migration enables in-tree storage plugin replacement with corresponding CSI drivers. User cannot be able to identify the differences if CSI migration works properly. Users may depend on the in-tree storage functionality with the existing interface. If Kubernetes updates the cluster to enable migration of CSI, then the stateful workloads and deployments continue to function behind the scenes of storage management operations. Kubernetes 1.17 version ensures storage APIs stability for the smooth upgrade promise. This involves backward capabilities. Existing feature behaviours, and API stability.

New features in Beta after Kubernetes 1.17:

  • Revamp of snapshot API volume.
  • Added deletion secret to the snapshot content as an annotation.
  • CSI snapshooter splits into two controllers external-snapshooter sidecar and common snapshot controller.
  • Added a new finalizer to prevent the content API object of volume snapshot.

Volume Snapshots Requirements:

  • Volume Snapshot controller.
  • Volume CRDs.
  • CSI driver with volume snapshot beta.

Drivers supporting Volume Snapshots:

Volume snapshot feature depends on a common snapshot controller to the CRDs. Both the CRDs and snapshot controller are independent of CSI driver. There must be only one instance snapshot controller to install CRDs per cluster. Hence, Kubernetes distributors deploy the CRDs as the cluster management process. Kubernetes supports Flex, in-tree, and CSI volume plugins. But, whereas snapshot supports CSI drivers to use Kubernetes features on the cluster. The following CSI drivers support volume snapshots:

  • Portworx CSI Driver
  • GCE Persistent Disk
  • NetApp Trident

Conclusion:

We have discussed the features of Kubernetes 1.17 release. These features enable user’s access with a single-line command to the latest upstream in 60 seconds. Snap store proxy controls the `Kubernetes-worker` charms, and CIS (Center for Internet Security) maintains the benchmark of Kubernetes 1.17. This is how Kubernetes 1.17 works.

##

About the Author

Anji Velagana, a graduate in Electronics and Communication Engineering from Jawaharlal Technological University, Kakinada. He is currently working as a Digital Marketing Analyst and Content Contributor for Mindmajix. He writes about various platforms like Data Science, Servicenow, Kubernetes, Business analysis, Performance testing, Mulesoft, Azure, SaltStack, and a few other courses. Contact him via anjivelagana@gmail.com and LinkedIn.