Kubernetes
Spark on Kubernetes - better handling for node shutdown
Spark 3.1 on the Kubernetes project is now officially declared as production-ready and Generally Available. Spot instances in Kubernetes can cut your bill by up to 70-80% if you are willing to trade in reliability. The new feature - SPIP: Add better handling for node shutdown (SPARK-20624) was implemented to deal with the problem of losing an executor when working with spot nodes - the need to recompute the shuffle or cached data.
Tại sao nên triển khai Apache Spark trên Kubernetes
Spark đã quá nổi tiếng trong thế giới Data Engineering và Bigdata. Kubernetes cũng ngày càng phổ biến tương tự, là một hệ thống quản lý deployment và scaling application. Bài viết này bàn đến một số lợi ích khi triển khai ứng dụng Apache Spark trên hệ thống Kubernetes.
Spark History Server on Kubernetes
The problem with running Spark on Kubernetes is the logs go away once the job completes. Spark has tool called the Spark History Server that provides a UI for your past Spark jobs. In this post, I will show you how to use Spark History Server on Kubernetes.
3 ways to run Spark on Kubernetes
Spark can run on clusters managed by Kubernetes. This feature makes use of native Kubernetes scheduler that has been added to Spark.
