You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: development-docs/RELEASE.md
+13-12Lines changed: 13 additions & 12 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -144,11 +144,11 @@ In order to provide the bundle for a new release, you can use the previous one a
144
144
Create a folder for the new release by copying the previous one and make the following changes:
145
145
146
146
* if releasing a new minor or major version (rather than fix), change the `metadata/annotations.yaml` to update the second channel listed next to `operators.operatorframework.io.bundle.channels.v1` to the new release version range (e.g. `strimzi-0.45.x`).
147
-
* copy the CRDs and the Cluster Roles YAML to the `manifests` folder by taking them from the `install/cluster-operator` folder (within the Strimzi repo).
147
+
* copy the CRDs, the Cluster Roles YAML (excluding the `strimzi-cluster-operator` roles) and the operator `ConfigMap` to the `manifests` folder by taking them from the `install/cluster-operator` folder (within the Strimzi repo).
148
148
* take the `strimzi-cluster-operator.v<VERSION>.clusterserviceversion.yaml` CSV file (by using the new release as `<VERSION>`) in order to update the following:
149
-
* `createAt` field with date/time creation of the current CSV file.
150
-
* `metadata.annotations.alm-examples-metadata` section by using the examples from the `examples` folder (within the Strimzi repo).
149
+
* `alm-examples-metadata` and `alm-examples` sections by using the examples from the `examples` folder (within the Strimzi repo).
151
150
* `containerImage` field with the new operator image (using the SHA).
151
+
* `createdAt` field with date/time creation of the current CSV file.
152
152
* `name` field by setting the new version in the operator name.
153
153
* `customresourcedefinitions.owned` section with the CRDs descriptions, from the `install/cluster-operator` folder (within the Strimzi repo).
154
154
* `description` section with all the Strimzi operator information already used for the release on GitHub.
@@ -180,7 +180,7 @@ In this section, the following steps show how to:
* docker, podman or buildah (the following steps will use docker as an example)
183
+
* docker, podman or buildah (the following steps allow you to configure the one to use via the `DOCKER_CMD` env var)
184
184
185
185
*Note*
186
186
For further details about the following steps, you can also refer to the official [Operator Lifecycle Manager documentation](https://olm.operatorframework.io/docs/tasks/).
@@ -193,7 +193,8 @@ Inside the bundle directory (i.e. `operators/strimzi-kafka-operator/0.45.0`), ex
193
193
```shell
194
194
export OPERATOR_VERSION=0.45.0
195
195
export DOCKER_REGISTRY=quay.io
196
-
export DOCKER_USER=ppatierno
196
+
export DOCKER_ORG=ppatierno
197
+
export DOCKER_CMD=docker
197
198
```
198
199
199
200
Run the following command in order to generate a `bundle.Dockerfile` representing the operator bundle image.
@@ -224,14 +225,14 @@ COPY metadata /metadata/
224
225
Run the following command to build the operator bundle image and push it to a repository:
You can test operator upgrades by starting from an existing catalog and then building a new catalog with a new operator version/bundle.
322
323
In this case, the new catalog image is pulled by Kubernetes/OpenShift.
323
-
It happens automatically if you are using a specific tag for the catalog image, for example going from `$DOCKER_REGISTRY/$DOCKER_USER/olm-catalog:1.0` to `$DOCKER_REGISTRY/$DOCKER_USER/olm-catalog:1.1`.
324
+
It happens automatically if you are using a specific tag for the catalog image, for example going from `$DOCKER_REGISTRY/$DOCKER_ORG/olm-catalog:1.0` to `$DOCKER_REGISTRY/$DOCKER_ORG/olm-catalog:1.1`.
324
325
If you are using the `latest` tag instead, you have to force pulling the new catalog image and one way is to kill the pod running the catalog.
0 commit comments