Chapter 2. Identify your Use Case
This chapter provides a brief introduction of the two use cases available in Containerized Red Hat Gluster Storage.
Red Hat Openshift Container Storage does not support a simultaneous deployment of converged and independent mode with ansible workflow. Therefore, you must deploy either converged mode or independent mode: you cannot mix both modes during deployment.
Red Hat only supports Heketi inside OpenShift Container Platform in OCS.
2.1. Converged Mode
Converged mode was earlier called as Container-Native Storage.
This deployment delivers a hyper-converged solution, where the storage containers that host Red Hat Gluster Storage co-reside with the compute containers and serve out storage from the hosts that have local or direct attached storage to the compute containers. This solution integrates Red Hat Gluster Storage deployment and management with OpenShift services. As a result, persistent storage is delivered within an OpenShift pod that provides both compute and file storage.
Converged Mode for OpenShift Container Platform is built around three key technologies:
- OpenShift provides the platform as a service (PaaS) infrastructure based on Kubernetes container management. Basic OpenShift architecture is built around multiple master systems where each system contains a set of nodes.
- Red Hat Gluster Storage provides the containerized distributed storage based on Red Hat Gluster Storage 3.5 container. Each Red Hat Gluster Storage volume is composed of a collection of bricks, where each brick is the combination of a node and an export directory.
- Heketi provides the Red Hat Gluster Storage volume life-cycle management. It creates the Red Hat Gluster Storage volumes dynamically and supports multiple Red Hat Gluster Storage clusters.
The following list provides the administrators a solution workflow. The administrators can:
- Create multiple persistent volumes (PV) and register these volumes with OpenShift.
- Developers then submit a persistent volume claim (PVC).
- A PV is identified and selected from a pool of available PVs and bound to the PVC.
- The OpenShift pod then uses the PV for persistent storage.
Figure 2.1. Architecture - Converged Mode for OpenShift Container Platform
Red Hat Openshift Container Storage does not support a simultaneous deployment of converged and independent mode with ansible workflow. Therefore, you must deploy either converged mode or independent mode: you cannot mix both modes during deployment.
2.2. Independent mode
Independent mode was earlier called Container-Ready Storage.
When Red Hat Gluster Storage is deployed as stand-alone storage, providing storage to containers, that is called independent mode. In this mode, the life cycle of the storage platform is maintained independently from the life cycle of the container platform.
When Red Hat Gluster Storage is deployed on top an OpenShift cluster, it is called converged mode.
Independent mode provides dynamically provisioned storage, statically provisioned storage, RWO support, and RWX support. Further, it provides full support for OpenShift Container Platform infrastructure services like logging, metrics, and registry services.
For users of persistent storage, the deployment modes are completely transparent. Administrators, however, will see a variation in how they set the system up, manage, and scale. In independent mode, storage is managed like Red Hat Gluster Storage.
Following are some of the key drivers of choosing independent mode of deployment:
- OpenShift Container Platform administrators might not want to manage storage. Independent mode separates storage management from container management.
- Leverage legacy storage (SAN, Arrays, Old filers): Storage arrays from traditional storage vendors often have either limited or no support for OpenShift. Independent mode allows you to leverage existing legacy storage for OpenShift Containers.
- Cost effective: In environments where costs related to new infrastructure is a challenge, you can repurpose existing storage arrays to back OpenShift under independent mode. Independent mode is perfect for situations where you can run Red Hat Gluster Storage inside a VM and serve out LUNs or disks from these storage arrays to OpenShift offering all the features that the OpenShift storage subsystem has to offer including dynamic provisioning. This is a very useful solution in those environments with potential infrastructure additions.
Independent mode may have Heketi, and other provisioners (components of independent mode) deployed on top of OpenShift Cluster nodes. Red Hat only supports Heketi inside OpenShift Container Platform in OCS. Heketi is a service endpoint for automated Red Hat Gluster Storage volume provisioning, where requests for allocation of Red Hat Gluster Storage volumes to back OpenShift PVs land from kubernetes. Heketi manages allocation and de-allocation of Red Hat Gluster Storage volumes dynamically.
Red Hat Openshift Container Storage does not support a simultaneous deployment of converged and independent mode with ansible workflow. Therefore, you must deploy either converged mode or independent mode: you cannot mix both modes during deployment.
In independent mode, Heketi must have complete control of the Gluster cluster.
