Shulou(Shulou.com)05/31 Report--

This article introduces the relevant knowledge of "what are the main functions of 2020Kubernetes". In the operation process of actual cases, many people will encounter such difficulties. Next, let Xiaobian lead you to learn how to deal with these situations! I hope you can read carefully and learn something!

Kubernetes Topology Manager

In many high-performance workloads, a combination of CPU and hardware accelerators is used to provide high-throughput parallel computing. To get the best performance out of a workload, various optimizations are required, such as CPU isolation and memory and device allocation. However, in Kubernetes, these tasks are performed against disjoint sets of components, because the device manager and CPU both work independently when allocating resources. This can result in performance degradation and increased latency.

Kubernetes Topology Manager was released as a beta version in Kubernetes v1.18. The topology manager is a kubelet component that helps reduce latency and improve performance for mission-critical applications. Topology Manager provides a single source for each component through an interface for a component called Hint Providers. Components can use this interface to make topology-consistent resource allocation decisions, providing low latency and optimized performance for critical workloads.

To enable this feature, TopologyManager feature gate needs to be enabled. In v1.18 and later, this feature is enabled by default.

Kubectl node debug

Kubectl node debug was launched as alpha in Kubernetes v1.18 and has been upgraded to beta in v1.20. This feature allows end users to debug their nodes via kubectl. Users can debug running Pods without restarting them or relying on entry containers to perform debugging tasks, such as checking file systems, executing other debug utilities, and performing initial network requests through host namespaces. Users can now do the following:

Troubleshooting workloads that crash at startup by creating copies of pods using different container images or commands. Troubleshooting Distroless containers by adding a new container with debugging tools to a Pod copy or temporary container. Troubleshooting nodes by creating a new container that runs on the host workspace and has access to the host file system. With this feature, Kubernetes hopes to no longer use SSH for node debugging and maintenance. Kubernetes v1.20 will enable kubectl debugging by default.

V1 Ingress API

Since Kubernetes v1.1, the API has been available as a beta version with various enhancements. Ingress has become a popular API among users and Load Balancer, and it has become a de facto GA in the K8s community. The Ingress API handles external access to services by exposing relevant HTTP and HTTPS routes. Ingress performs various tasks such as load balancing, providing name-based virtual hosting, and terminating SSL/TLS. Ingress resources depend on the Ingress controller. K8s currently supports various Ingress controllers such as nginx. In v1.18, Kubernetes made some key changes to Ingress. The new Path Type field becomes available and is set to ImplementationSpecific by default. Users can now specify path types using Exact and Prefix. The user can use the IngressClass field to specify the Ingress type in the K8s cluster. Starting with Kubernetes v1.18, a new field named IngressClassName is provided to specify the Ingress class, replacing the previous kubernetes. io/ingress.class.

Volume Snapshot Operations

Snapshots are not the most reliable backup and restore solution when it comes to high-volume workloads. But when used correctly, snapshots can help provide backup and restore functionality without downtime. This enhancement was enabled in Kubernetes v1.16 as a beta release and provides API support for Kubernetes Container Storage Interface (CSI) plug-ins to take PersistentVolume snapshots and restore them when needed. To ensure reliable snapshots, users should ensure data consistency across the application level, host operating system, and storage system. Snapshots taken before application data is stored in memory are useless. In Kubernetes v1.20, volume snapshot operations were GA. This feature allows users to take snapshots of data volumes in a standardized manner to ensure reliability. Snapshot operations are portable and can be integrated into various Kubernetes environments or supported storage providers. These snapshot primitives can be used to develop advanced storage management capabilities for the K8s for cluster and application-level backups.

Node Elegant Close

This new feature was introduced in alpha version in the latest K8s release. This feature solves a problem many users and cluster administrators face when Pod is down. Pods do not always follow the Pod lifecycle. This can happen when a node system in a running Pod is shut down. Since the Pod has no way of knowing that the node is down, the Pod may experience problems and fail to function properly. The GracefulNode Shutdown feature solves this problem by making the kubelet aware that node shutdown causes a Pod to terminate normally.

HPA Rate Control

The Pod Horizontal Auto-Scale (HPA) API is an alpha feature that allows a Pod to automatically scale into a replica set based on certain metric values. This feature is useful when it comes to applications that may be affected by traffic fluctuations. The HPA now comes with optional field behavior, allowing users to set different scaling ratios for different applications based on their capabilities and known behavior.

"What are the main functions of 2020Kubernetes" is introduced here. Thank you for reading. If you want to know more about industry-related knowledge, you can pay attention to the website. Xiaobian will output more high-quality practical articles for everyone!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.