In the past week the launch of the new version of the container orchestration platform Kubernetes 1.18, version that includes 38 changes and improvements, of which 15 are in stable state and 11 are in beta state, in addition to 12 new alpha state changes are proposed. In preparing the new version, equitable efforts were directed both to the refinement of various functions and to the stabilization of experimental capabilities, as well as the incorporation of new developments.
For those unfamiliar with Kubernetes, you should know that this is a container orchestration rig which allows you to manage a cluster of isolated containers as a whole and provide mechanisms for deploying, maintaining and scaling applications that run in containers.
The project was originally created by Google but was later transferred to a separate platform, curated by the Linux Foundation. The platform is positioned as a universal solution developed by the community, not linked to individual systems and capable of working with any application in any cloud environment. The Kubernetes code is written in Go and is distributed under the Apache 2.0 license.
What's new in Kubernetes 1.18?
This new version of Kubernetes comes with various enhancements for Kubectl, of which it is mentioned in the ad that added an alpha version of the "kubectl debug" command, which makes it easy to debug in pods when running containers with debugging tools.
While the command "Kubectl diff" has been declared stable, which allows you to see what will change in the cluster if you apply the manifest.
As well all "kubectl run" command generators have been removed, except for the single pod generator startup, plus the indicator –Dry-run was changed, depending on its value (client, server and none), the test execution of the command is done on the client or server side.
The code kubectl is assigned to a separate repository. This allowed us to separate kubectl from internal kubernetes dependencies and made it easy to import code into third-party projects.
As for the network changes, it is noted that IPv6 support is now in beta, added PVC cloning, the possibility of network blocking raw devices such as permanent disks, support for blocking raw devices in CSI, transfer of information about the drive requesting to connect a disk to the CSI controller, plus that a new "immutable" field has been added to the ConfigMap and Secret objects.
Of the other changes that stand out:
- The ability to use the deprecated API group / v1beta1 apps and / v1beta1 extensions was finally removed.
- ServerSide Apply updated to beta2 state. This enhancement brings kubectl object manipulation to the API server.
- CertificateSigningRequest API declared stable.
- Support for the Windows platform.
- Windows node support continues to expand
- CRI-ContainerD support
- RuntimeClass implementation
- CSI proxy
- Transferred support has been stable
- Group managed service account
- RunAsUserName
- Topology Manager has received beta status. The feature includes NUMA distribution, which prevents performance degradation on multi-socket systems.
- Beta status was obtained using the PodOverhead function, which allows you to specify in RuntimeClass the additional amount of resources required to start the home.
- Extended hugepages support, alpha isolation status added to container and support for multi-level hugepages sizes.
- Added AppProtocol field where you can specify which protocol the app uses
- Translated to beta state and enabled by default EndpointSlicesAPI, which is a more functional replacement for regular Endpoints.
- An IngressClass object has been added, indicating the name of the input controller, its additional parameters, and the sign to use it by default.
- Added the ability to specify in the HPA manifest the degree of aggressiveness when changing the number of homes in operation, that is, when the load increases, it immediately starts N times more copies.