Skip to content

Commit

Permalink
chore: Document hardware support
Browse files Browse the repository at this point in the history
Signed-off-by: Denilson Nastacio <[email protected]>
  • Loading branch information
nastacio committed Feb 1, 2024
1 parent 514a7de commit 76ff6b7
Showing 1 changed file with 23 additions and 6 deletions.
29 changes: 23 additions & 6 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,8 @@
- [Shared cluster](#shared-cluster)
- [GitOps](#gitops)
- [Governance Policies](#governance-policies)
- [Hardware](#hardware)
- [Hardware support constraints](#hardware-support-constraints)
- [Storage](#storage)
- [Installation](#installation)
- [Individual clusters](#individual-clusters)
Expand All @@ -16,7 +18,7 @@

This repository contains Argo CD `Application` resources representing sample deployments of IBM Cloud Paks, and, as such, they are meant for inclusion in an Argo CD cluster. Different Cloud Paks are represented with different `Application` resources and grouped by a resource label tied to each Cloud Pak.

**Important**: This repository is meant as a demonstration of how Cloud Pak deployments can be deployed and managed with GitOps practices. Adoption in a production environment can start from a repository fork, followed by customization of folders and files to match the desired configuration.
**Important**: This repository demonstrates how Cloud Pak deployments can be deployed and managed with GitOps practices. Adoption in a production environment can start from a repository fork, followed by customization of folders and files to match the desired configuration.

You may decide to include one or more of these `Application` objects to the target cluster and then determine which ones you want to synchronize into the cluster.

Expand All @@ -40,9 +42,9 @@ Supported versions:

### Shared cluster

Starting with the v0.22 release, it is possible to deploy Cloud Paks using dedicated automation foundation instances.
Starting with the v0.22 release, deploying Cloud Paks using dedicated automation foundation instances is possible.

At the root of this configuration, lies a pre-synchronization hook inside the `cp-shared` application, which creates a default "common-service-maps" ConfigMap under the `kube-public` namespace, according to the instructions listed under <https://www.ibm.com/docs/en/cloud-paks/1.0?topic=cfs-installing-cloud-pak-foundational-services-in-multiple-namespaces>
At the root of this configuration lies a pre-synchronization hook inside the `cp-shared` application, which creates a default "common-service-maps" ConfigMap under the `kube-public` namespace, according to the instructions listed under <https://www.ibm.com/docs/en/cloud-paks/1.0?topic=cfs-installing-cloud-pak-foundational-services-in-multiple-namespaces>

Note that if you want to enable this feature and customize the target namespaces for Cloud Paks, you must update the parameters to the `cp-shared-app` application to override that default location.

Expand All @@ -52,18 +54,33 @@ GitOps is a declarative way to implement continuous deployment for cloud-native

### Governance Policies

Practicing GitOps at scale, with dozens or even hundreds of clusters, benefits from a level of abstraction where each cluster follows a few select policies. This repository contains a simple deployment of governance policies for the deployment of OpenShift GitOps and Cloud Paks to a fleet of clusters.
Practicing GitOps at scale, with dozens or even hundreds of clusters, benefits from a level of abstraction where each cluster follows a few select policies. This repository contains a simple deployment of governance policies for deploying OpenShift GitOps and Cloud Paks to a fleet of clusters.

## Hardware

Within the hardware support constraints (see below) of each Cloud Pak, the Argo CD applications in this repository were tested with OpenShift clusters running in `amd64` and `ppc64le` processor architecture.

The repository content is enabled to run in the `s390x` architecture but has not been directly tested with it.

### Hardware support constraints

- The API Connect component of Cloud Pak for Integration is not supported in the `ppc64le` processor architecture.
- The R Studio component of Cloud Pak for Data is only supported in `amd64` processors.
- The AI Manager component of Cloud Pak for AIOps is only supported in `amd64` processors.
- Cloud Pak for Security is only supported in `amd64` processors.

This section's list of exclusions and support is not exhaustive. Always consult the respective Cloud Pak documentation to determine if a component is supported in a given hardware platform.

## Storage

The instructions in this repository assume the user already has an OpenShift cluster with storage capable of RWO and RWX access mode.
The Argo CD `Application` resources have pre-synchronization hooks that will attempt auto-detection of the storage in the cluster from common storage providers, in decreasing order of precedence:
The Argo CD `Application` resources have pre-synchronization hooks that will attempt auto-detection of the storage in the cluster from common storage providers in decreasing order of precedence:

| Precedence | OpenShift platform | Storage | Storage classes |
| ---------- | ------------ | ------- | --------------- |
| Highest | All | OpenShift Data Foundation | `ocs-storagecluster-ceph-rbd` (RWO) and `ocs-storagecluster-cephfs` (RWX). |
| High | All | Rook Ceph | `rook-ceph-block` (RWO) and `rook-cephfs` (RWX) |
| Medium | All | NetApp ONTAP (Trident driver) | `trident-csi` (RWO and RWX.) Note about RWO access mode: the synchronization hooks give preference to a storage class named `trident-block-csi` if it can find one in the cluster. |
| Medium | All | NetApp ONTAP (Trident driver) | `trident-csi` (RWO and RWX.) Note about RWO access mode: the synchronization hooks prefer a storage class named `trident-block-csi` if they can find one in the cluster. |
| Low | IBM Cloud (ROKS classic) | IBM Cloud File Storage | `ibmc-block-gold` (RWO) and `ibmc-file-gold-gid` (RWX) |
| Low | AWS (self-managed and ROSA) | Elastic Block Store and Elastic File System | First storage class with provisioner `ebs.csi.aws.com` (RWO) and first storage class with provisioner `efs.csi.aws.com` (RWX) |
| Low | Azure (self-managed and ARO) | Azure Disk (classic) and Azure File System | `managed-premium` (RWO) and `azure-file` (RWX) |
Expand Down

0 comments on commit 76ff6b7

Please sign in to comment.