📖 Document CRD changes from v1alpha4 to v1alpha5

docs/book/src/SUMMARY.md

@@ -7,5 +7,7 @@
     - [external cloud provider](./topics/external-cloud-provider.md)
     - [move from bootstrap](./topics/mover.md)
     - [trouble shooting](./topics/troubleshooting.md)
+    - [CRD Changes](./topics/crd-changes/index.md)
+        - [v1alpha4 to v1alpha5](./topics/crd-changes/v1alpha4-to-v1alpha5.md)
 - [Development](./development/development.md)
 - [Hacking CI](./development/ci.md)

docs/book/src/topics/crd-changes/index.md

@@ -0,0 +1,5 @@
+# CRD Changes
+
+### Conversions
+
+CAPO is able to automatically convert your old resources into new API versions.

docs/book/src/topics/crd-changes/v1alpha4-to-v1alpha5.md

@@ -0,0 +1,123 @@
+<!-- START doctoc generated TOC please keep comment here to allow auto update -->
+<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
+**Table of Contents**  *generated with [DocToc](https://github.com/thlorenz/doctoc)*
+
+- [v1alpha4 compared to v1alpha5](#v1alpha4-compared-to-v1alpha5)
+  - [Conversion from v1alpha4 to v1alpha5](#conversion-from-v1alpha4-to-v1alpha5)
+  - [API Changes](#api-changes)
+    - [`OpenStackCluster`](#openstackcluster)
+      - [Managed API LoadBalancer](#managed-api-loadbalancer)
+      - [Major Changes to Ports and Networks](#major-changes-to-ports-and-networks)
+    - [`OpenStackMachine`](#openstackmachine)
+      - [Rename of `status.error{Reason,Message}` to `status.failure{Reason,Message}`](#rename-of-statuserrorreasonmessage-to-statusfailurereasonmessage)
+      - [Changes to `rootVolume`](#changes-to-rootvolume)
+
+<!-- END doctoc generated TOC please keep comment here to allow auto update -->
+
+# v1alpha4 compared to v1alpha5
+
+## Migration
+
+
+All users are encouraged to migrate their usage of the CAPO CRDs from older versions to `v1alpha5`. This includes yaml files and source code. As CAPO implements automatic conversions between the CRD versions, this migration can happen after installing the new CAPO release.
+
+## API Changes
+
+This only documents backwards incompatible changes. Fields that were added to v1alpha5 are not listed here.
+
+### `OpenStackCluster`
+
+#### Managed API LoadBalancer
+
+The fields related to the managed API LoadBalancer were moved into a seperate object:
+
+```yaml
+apiVersion: infrastructure.cluster.x-k8s.io/v1alpha4
+kind: OpenStackCluster
+spec:
+  managedAPIServerLoadBalancer: true
+  apiServerLoadBalancerAdditionalPorts: [443]
+```
+
+becomes:
+
+```yaml
+apiVersion: infrastructure.cluster.x-k8s.io/v1alpha5
+kind: OpenStackCluster
+spec:
+  apiServerLoadBalancer:
+    enabled: true
+    additionalPorts: [443]
+```
+
+### `OpenStackMachine`
+
+#### Major Changes to Ports and Networks
+
+When using Ports it is now possible to specify network and subnet by filter instead of just ID. As a consequence, the relevant ID fields are now moved into the new filter specifications:
+
+```yaml
+ports:
+  - networkId: d-e-a-d
+    fixedIPs:
+      - subnetId: b-e-e-f
+```
+
+becomes:
+
+```yaml
+ports:
+  - network:
+      id: d-e-a-d
+    fixedIPs:
+      subnet:
+        id: b-e-e-f
+```
+
+Networks are now deprecated. With one exception, all functionality of Networks is now available for Ports. Consequently, Networks will be removed from the API in a future release.
+
+The ability of a Network to add multiple ports with a single directive will not be added to Ports. When moving to Ports, all ports must be added explicitly. Specifically, when evaluating the network or subnet filter of a Network, if there are multiple matches we will add all of these to the server. By contrast we raise an error if the network or subnet filter of a Port does not return exactly one result.
+
+`tenantId` was previously a synonym for `projectId` in both network and subnet filters. This has now been removed. Use `projectId` instead.
+
+The following fields are removed from network and subnet filters without replacement:
+
+- status
+- adminStateUp
+- shared
+- marker
+- limit
+- sortKey
+- sortDir
+- subnetPoolId
+
+#### Rename of `status.error{Reason,Message}` to `status.failure{Reason,Message}`
+
+The actual fields were previously already renamed, but we still used the `error` prefix in JSON. This was done to align with CAPI, where these fields were [renamed in v1alpha3](https://cluster-api.sigs.k8s.io/developer/providers/v1alpha2-to-v1alpha3.html#external-objects-will-need-to-rename-statuserrorreason-and-statuserrormessage).
+
+```yaml
+apiVersion: infrastructure.cluster.x-k8s.io/v1alpha4
+kind: OpenStackMachine
+status:
+  errorReason: UpdateError
+  errorMessage: Something when wrong
+```
+
+becomes:
+
+```yaml
+apiVersion: infrastructure.cluster.x-k8s.io/v1alpha5
+kind: OpenStackMachine
+status:
+  failureReason: UpdateError
+  failureMessage: Something when wrong
+```
+
+#### Changes to `rootVolume`
+
+The following fields were removed without replacement:
+
+- `rootVolume.deviceType`
+- `rootVolume.sourceType`
+
+Additionally, `rootVolume.sourceUUID` has been replaced by using `ImageUUID` or `Image` from the OpenStackMachine as appropriate.