Book Image

Running Windows Containers on AWS

By : Marcio Morales
Book Image

Running Windows Containers on AWS

By: Marcio Morales

Overview of this book

Windows applications are everywhere, from basic intranet applications to high-traffic public APIs. Their prevalence underscores the importance of combining the same tools and experience for managing a modern containerized application with existing critical Windows applications to reduce costs, achieve outstanding operational excellence, and modernize quickly. This comprehensive guide to running and managing Windows containers on AWS looks at the best practices from years of customer interactions to help you stay ahead of the curve. Starting with Windows containers basics, you’ll learn about the architecture design that powers Amazon ECS, EKS, and AWS Fargate for Windows containers. With the help of examples and best practices, you’ll explore in depth how to successfully run and manage Amazon ECS, EKS, and AWS Fargate clusters with Windows containers support. Next, the book covers day 2 operations in detail, from logging and monitoring to using ancillary AWS tools that fully containerize existing legacy .NET Framework applications into containers without any code changes. The book also covers the most common Windows container operations, such as image lifecycle and working with ephemeral hosts. By the end of this book, you’ll have mastered how to run Windows containers on AWS and be ready to start your modernization journey confidently.
Table of Contents (22 chapters)
1
Part 1: Why Windows Containers on Amazon Web Services (AWS)?
4
Part 2: Windows Containers on Amazon Elastic Container Service (ECS)
9
Part 3: Windows Containers on Amazon Elastic Kubernetes Service (EKS)
14
Part 4: Operationalizing Windows Containers on AWS

Avoiding pod-scheduling disruption

Running a heterogeneous Kubernetes cluster imposes some particularities that need to be considered to reduce application deployment disruption. By default, Amazon EKS uses kube-scheduler as the default scheduler for Kubernetes; when you schedule a pod, kube-scheduler uses a combination of filtering and scoring to select the optimal node to run the pod, which works nicely when you only run Linux-based applications.

kube-scheduler doesn’t take the OS into account and as a result, it is common to see Linux-based sidecar containers or plugins running as DaemonSets failing to be scheduled just because it doesn’t treat nodeSelector in their manifest. For instance, a DaemonSet ensures that all nodes run a copy of a pod, but if you have a heterogenous cluster, Windows pods cannot be deployed on Linux-based nodes and vice versa, failing to schedule the pod; thereby, your deployment will never have a 100% success rate.

We can work around...