Chapter 2. Architecture overview
This is a reference architecture for running OpenShift Container Platform 3.11 on Red Hat OpenStack Platform 13. In this reference architecture, Red Hat OpenStack Platform is deployed to physical servers; OpenShift Container Platform is deployed to virtual machines running on Red Hat OpenStack Platform’s hypervisor. The architecture is highly available and suitable for production.
Both OpenShift Container Platform and OpenStack Platform are modular and configurable, and there are many supported ways to install OpenShift on OpenStack. Red Hat field consultants developed the recommendations in this reference architecture based on their experience deploying to production. This document describes the reference architecture as it was deployed and tested in a lab environment. It also shares important design considerations and key integrations between the products.
Step by step instructions for installing Red Hat OpenStack Platform and Red Hat OpenShift Container Platform is beyond the scope of this document. For detailed installation instructions, see the OpenStack Platform 13 Director Installation and Usage Guide and the OpenShift Container Platform 3.11 product documentation.
2.1. Relationship between OpenShift and OpenStack
Figure 1 shows a block diagram of OpenShift Container Platform running on OpenStack Platform. Starting from the bottom, Red Hat OpenStack Platform is deployed on commodity hardware; OpenShift Container Platform is installed on either virtual machines or bare metal servers provisioned by OpenStack. Containerized applications run in OpenShift alongside applications running on OpenStack virtual machines.
The relationship between OpenStack and OpenShift is complementary: OpenStack exposes resources that OpenShift consumes. It provides OpenShift with compute, storage, and networking infrastructure plus additional resources such as self-service load balancers and encryption. OpenShift runs its containerized applications on the infrastructure provisioned by OpenStack. The products are tightly integrated, allowing OpenShift to consume OpenStack resources on demand and without user intervention.
Figure 2 illustrates the relationship between OpenShift and OpenStack by showing a typical interaction between a containerized application and the infrastructure; this example relates to storage. When an application running on OpenShift needs a persistent volume, it submits a persistent volume claim to the OpenShift API. This OpenShift API call is translated into a Cinder API call to create a volume. When the volume is ready it is presented back to OpenShift, and attached to the requesting pod. The persistent volume claim only needs to include the volume size and access mode. The backend implementation details of how and where the volume is created are handled by OpenStack. The OpenShift API abstracts them from the user making the resource claim. This pattern of interaction is repeated for other OpenShift infrastructure requests to OpenStack.