Book Image

Learn OpenShift

By : Denis Zuev, Artemii Kropachev, Aleksey Usov
Book Image

Learn OpenShift

By: Denis Zuev, Artemii Kropachev, Aleksey Usov

Overview of this book

Docker containers transform application delivery technologies to make them faster and more reproducible, and to reduce the amount of time wasted on configuration. Managing Docker containers in the multi-node or multi-datacenter environment is a big challenge, which is why container management platforms are required. OpenShift is a new generation of container management platforms built on top of both Docker and Kubernetes. It brings additional functionality to the table, something that is lacking in Kubernetes. This new functionality significantly helps software development teams to bring software development processes to a whole new level. In this book, we’ll start by explaining the container architecture, Docker, and CRI-O overviews. Then, we'll look at container orchestration and Kubernetes. We’ll cover OpenShift installation, and its basic and advanced components. Moving on, we’ll deep dive into concepts such as deploying application OpenShift. You’ll learn how to set up an end-to-end delivery pipeline while working with applications in OpenShift as a developer or DevOps. Finally, you’ll discover how to properly design OpenShift in production environments. This book gives you hands-on experience of designing, building, and operating OpenShift Origin 3.9, as well as building new applications or migrating existing applications to OpenShift.
Table of Contents (24 chapters)

Network topology in OpenShift

In order to provide a common medium for containers to communicate with each other, OpenShift makes use of an overlay network that's implemented via VXLAN. The Virtual eXtensible Local Area Network (VXLAN) protocol provides a mechanism for transferring Layer 2 (Ethernet) frames across Layer 3 (IP) networks. Depending on the SDN plugin being used, the scope of communication may be limited to pods within the same project or maybe completely unrestricted. No matter which plugin is used, the network topology is still the same.

When a new node is registered in etcd, the master allocates a private /23 subnet from the cluster network. By default, subnets are allocated from 10.128.0.0/14, but can be configured in the networkConfig stanza of the master configuration file. The following is an excerpt from the file containing the relevant parameters...