CNV#33978-1: ROSA support and general storage update

[ousleyp]

Version(s): 4.14+

Issue: CNV-33978

Link to docs preview: https://67365--docspreview.netlify.app/openshift-enterprise/latest/virt/install/preparing-cluster-for-virt#virt-aws-bm_preparing-cluster-for-virt

QE review:

• QE has approved this change.
• seeking additional approval from @orenc1 before merge

Additional information:

• Removing information about specific storage vendors per @ronensdeor - FYI @aglitke @dankenigsberg @kmajcher-rh
• Planning to open a second PR to add a release note (for CNV) re: ROSA support (Update: opened CNV#33978-2: updating RN for ROSA support #67398)
• Also planning to follow up with the ROSA team to add a release note there

[ocpdocs-previewbot]

🤖 Updated build preview is available at:
https://67365--docspreview.netlify.app

Build log: https://circleci.com/gh/ocpdocs-previewbot/openshift-docs/31830

[dbasunag]

/lgtm

[cbippley]

/label peer-review-in-progress
/remove-label peer-review-needed

[cbippley]

LGTM!

[cbippley]

/label peer-review-done
/remove-label peer-review-in-progress

[orenc1]

lgtm in general, but I would like to comment about:

• You can install the cluster by using installer-provisioned infrastructure, ensuring that you specify bare-metal node instance types in the install-config.yaml file. For example, you can use the c5n.metal type value for a machine based on x86_64 architecture.

I think we can mention that at least the worker nodes should be bare metal instances (e.g. c5n.metal), since they are running the VirtualMachines. At least two of them in order to support live migration.
There is no hard requirements for the master/control-plane nodes to be also bare metals.
This distinction has a considerable influence on the running costs of the AWS/ROSA cluster. Metal instance is 10x more expensive than a virtual EC2 that can serve as a master node.

[openshift-ci]

New changes are detected. LGTM label has been removed.

[ousleyp]

lgtm in general, but I would like to comment about:

• You can install the cluster by using installer-provisioned infrastructure, ensuring that you specify bare-metal node instance types in the install-config.yaml file. For example, you can use the c5n.metal type value for a machine based on x86_64 architecture.

I think we can mention that at least the worker nodes should be bare metal instances (e.g. c5n.metal), since they are running the VirtualMachines. At least two of them in order to support live migration. There is no hard requirements for the master/control-plane nodes to be also bare metals. This distinction has a considerable influence on the running costs of the AWS/ROSA cluster. Metal instance is 10x more expensive than a virtual EC2 that can serve as a master node.

Changed to:

* You can install the cluster by using installer-provisioned infrastructure, ensuring that you specify bare-metal instance types for the worker nodes by editing the `install-config.yaml` file. For example, you can use the `c5n.metal` type value for a machine based on x86_64 architecture.

[ousleyp]

/cherrypick enterprise-4.15

[ousleyp]

/cherrypick enterprise-4.14

[openshift-cherrypick-robot]

@ousleyp: new pull request created: #67422

In response to this:

/cherrypick enterprise-4.15

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[openshift-cherrypick-robot]

@ousleyp: new pull request created: #67423

In response to this:

/cherrypick enterprise-4.14

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.