Update nova04delta docs

Signed-off-by: Bohdan Dobrelia <[email protected]>

Author: bogdando

dt/nova/nova04delta/README.md

@@ -0,0 +1,12 @@
+# Deployed Topology - Nova/nova04delta
+
+If you are looking for information on how to deploy the nova04delta based DT, then
+please see the 
+[README](../../../examples/dt/nova/nova04delta/README.md) in the examples
+directory.
+
+This directory ,`dt/nova/nova04delta/`, exists so that the
+[kustomization.yaml](../../../examples/dt/nova/nova04delta/edpm/nodeset/kustomization.yaml)
+in the examples directory of nova04delta topology, reference it by path as a
+component. It's contents are likely uninteresting unless you want to understand
+how kustomize was implemented in this repository.

examples/dt/nova/nova04delta/README.md

@@ -29,14 +29,16 @@ The following parameters are crucial for host-level configuration:
 
 ## Nova Configuration
 
-# FIXME: this looks only used in PCI groups in placement? The same pci_passthrough:alias should be used in both cases it seems?
-Nova control plane is configured for requesting PCI devices from Placement
-through `resources:VGPU=X` flavor extra specs.
-That is a contrary to the legacy mode where PCI devices used to be requested through
-`pci_passthrough:alias` flavor extra specs.
+A count of `X` PCI devices may be requested through `"pci_passthrough:alias"="nvidia_a2:X"` flavor extra specs:
+```
+$ openstack --os-compute-api=2.86 flavor set --property "pci_passthrough:alias"="nvidia_a2:1" device_passthrough
+```
 
 ### Control Plane (`examples/dt/nova/nova04delta/control-plane/service-values.yaml`)
 
+See [README.md](control-plane/README.md) for deployment instructions.
+There are most essential configuration values to define:
+
 *   `[pci]alias`: Creates an alias for a specific GPU type. This allows users to request a GPU by a friendly name (e.g., `nvidia_a2`) when creating a VM. This configuration should match the configuration found on the compute nodes.
     ```yaml
     nova:
@@ -53,6 +55,9 @@ That is a contrary to the legacy mode where PCI devices used to be requested thr
 
 ### Compute Node (`examples/dt/nova/nova04delta/edpm/nodeset/values.yaml`)
 
+See [dataplane section](data-plane.md) for deployment instructions.
+There are most essential configuration values to define:
+
 *   `[pci]report_in_placement`: Required for PCI in placement to work.
 *   `[pci]device_spec`: Whitelists the physical GPUs that are available for passthrough. You must create a `device_spec` entry for each physical GPU you want to make available. For example:
     ```yaml

examples/dt/nova/nova04delta/data-plane.md

@@ -2,27 +2,30 @@
 
 ## Assumptions
 
-- The [control plane](control-plane.md) has been created and successfully deployed
+- The [control plane](control-plane/README.md) has been created and successfully deployed
 
 ## Initialize
 
 Switch to the "openstack" namespace
 ```
 oc project openstack
 ```
-Change to the nfv/sriov/edpm directory
+Change to the nova/nova04delta directory
 ```
-cd architecture/examples/va/nfv/sriov/edpm
+cd architecture/examples/va/nova/nova04delta/edpm
 ```
-Edit the [nodeset/values.yaml](nodeset/values.yaml) and [deployment/values.yaml](deployment/values.yaml) files to suit 
+Edit the [nodeset/values.yaml](nodeset/values.yaml), [nodeset2/values.yaml](nodeset/values.yaml)
+and [deployment/values.yaml](deployment/values.yaml) files to suit
 your environment.
 ```
 vi nodeset/values.yaml
+vi nodeset2/values.yaml
 vi deployment/values.yaml
 ```
-Generate the dataplane nodeset CR.
+Generate the dataplane nodeset CRs.
 ```
 kustomize build nodeset > dataplane-nodeset.yaml
+kustomize build nodeset2 > dataplane-nodeset2.yaml
 ```
 Generate the dataplane deployment CR.
 ```
@@ -33,6 +36,7 @@ kustomize build deployment > dataplane-deployment.yaml
 Create the nodeset CR
 ```
 oc apply -f dataplane-nodeset.yaml
+oc apply -f dataplane-nodeset2.yaml
 ```
 Wait for dataplane nodeset setup to finish
 ```
@@ -48,3 +52,32 @@ Wait for dataplane deployment to finish
 ```
 oc wait osdpns openstack-edpm --for condition=Ready --timeout=40m
 ```
+
+## Generate yamls necessary to finialize Nvidia GPU installation
+After the Nvidia drivers have been blacklisted on the EDPM nodes, the computes need
+to be rebooted. Below explains the necessary steps to apply a reboot on the necessary nodesets.
+
+Change to the nova/nova04delta/edpm-post-driver directory
+```
+cd architecture/examples/dt/nova/nova04delta/edpm-post-driver/
+```
+Edit the [deployment/values.yaml](edpm-post-driver/deployment/values.yaml) files to suit
+your environment.
+```
+vi deployment/values.yaml
+```
+Generate the dataplane deployment CR.
+```
+kustomize build deployment > dataplane-post-driver-deployment.yaml
+```
+
+## Create CRs to finalize GPU installation
+Start the deployment
+```
+oc apply -f dataplane-post-driver-deployment.yaml
+```
+
+Wait for dataplane deployment to finish
+```
+oc wait osdpd edpm-deployment-post-driver --for condition=Ready --timeout=20m
+```