Bug 1876535

Description of problem:

Instantiation of templates gives no positive (on success) or negative (on instantiation failure) feedback to user
For example, if a template's rendered resource contains a typo and cannot be applied, or if creation of a resource is not allowed

Connect to the console

Version-Release number of selected component (if applicable):
4.4
4.5

How reproducible:


Steps to Reproduce:

0. All under the "default" project (but it's the same under other projects)

1. Create the following templates:
~~~
cat <<'EOF' | oc apply -f -
apiVersion: template.openshift.io/v1
kind: Template
metadata:
  name: deployment-template-working
  annotations:
    description: "Create a new deployment from image"
objects:
- apiVersion: apps/v1
  kind: Deployment
  metadata:
    name: ${IMAGE_NAME}-deployment
    labels:
      app: ${IMAGE_NAME}-deployment
  spec:
    replicas: ${{DEPLOYMENT_SCALE}}
    selector:
      matchLabels:
        app: ${IMAGE_NAME}-pod
    template:
      metadata:
        labels:
          app: ${IMAGE_NAME}-pod
      spec:
        containers:
        - name: ${IMAGE_NAME}
          image: ${IMAGE_NAME}
          command:
            - sleep
            - infinity
          imagePullPolicy: IfNotPresent
parameters:
- description: Image name - sets image to pull and also sets the deployment and pod name
  required: true
  name: IMAGE_NAME
- description: Image name - sets image to pull and also sets the deployment and pod name
  from: '[1-3]{1}'
  generate: expression
  name: DEPLOYMENT_SCALE
EOF

cat <<'EOF' | oc apply -f -
apiVersion: template.openshift.io/v1
kind: Template
metadata:
  name: deployment-template-failing
  annotations:
    description: "Create a new deployment from image"
objects:
- apiVersion: apps/v1
  kind: Deployment
  metadata:
    name: ${IMAGE_NAME}-deployment
    labels:
      app: ${IMAGE_NAME}-deployment
  spec:
    replicas: ${DEPLOYMENT_SCALE}
    selector:
      matchLabels:
        app: ${IMAGE_NAME}-pod
    template:
      metadata:
        labels:
          app: ${IMAGE_NAME}-pod
      spec:
        containers:
        - name: ${IMAGE_NAME}
          image: ${IMAGE_NAME}
          command:
            - sleep
            - infinity
          imagePullPolicy: IfNotPresent
parameters:
- description: Image name - sets image to pull and also sets the deployment and pod name
  required: true
  name: IMAGE_NAME
- description: Image name - sets image to pull and also sets the deployment and pod name
  from: '[1-3]{1}'
  generate: expression
  name: DEPLOYMENT_SCALE
EOF

cat <<'EOF' | oc apply -f -
apiVersion: template.openshift.io/v1
kind: Template
metadata:
  name: namespace-template
  annotations:
    description: "Create a new namespace"
objects:
- apiVersion: project.openshift.io/v1
  description: "${SERVICE_NAME}"
  displayName: "${SERVICE_NAME}"
  kind: ProjectRequest
  metadata:
    name: ${SERVICE_NAME}
parameters:
- description: Service name - will set a namespace from it - must be the same than your gitlab repository
  from: '[A-Za-z\_\-]{60}'
  generate: expression
  name: SERVICE_NAME
EOF
~~~

~~~
[akaris@linux aws-env]$ oc get templates
NAME                          DESCRIPTION                          PARAMETERS        OBJECTS
deployment-template-failing   Create a new deployment from image   2 (1 blank)       1
deployment-template-working   Create a new deployment from image   2 (1 blank)       1
namespace-template            Create a new namespace               1 (1 generated)   1
~~~

2. Instantiate the templates via CLI:
~~~
[akaris@linux aws-env]$ oc process deployment-template-working IMAGE_NAME=ubuntu DEPLOYMENT_SCALE=2  | oc create -f -
deployment.apps/ubuntu-deployment created
[akaris@linux aws-env]$ oc process deployment-template-failing IMAGE_NAME=fedora DEPLOYMENT_SCALE=2  | oc create -f -
Error from server (BadRequest): Deployment in version "v1" cannot be handled as a Deployment: v1.Deployment.Spec: v1.DeploymentSpec.Replicas: readUint32: unexpected character: �, error found in #10 byte of ...|eplicas":"2","select|..., bigger context ...|yment","namespace":"default"},"spec":{"replicas":"2","selector":{"matchLabels":{"app":"fedora-pod"}}|...
[akaris@linux aws-env]$ oc process namespace-template SERVICE_NAME=test | oc apply -f -
project.project.openshift.io/test created
~~~

Verify:
~~~
[akaris@linux aws-env]$ oc projects | grep test
    test
[akaris@linux aws-env]$ 
[akaris@linux aws-env]$ 
[akaris@linux aws-env]$ oc delete project test
project.project.openshift.io "test" deleted
~~~

~~~
[akaris@linux aws-env]$ oc delete project test
project.project.openshift.io "test" deleted
[akaris@linux aws-env]$ oc get deployments
NAME                READY   UP-TO-DATE   AVAILABLE   AGE
ubuntu-deployment   2/2     2            2           2m27s
[akaris@linux aws-env]$ oc delete deployment ubuntu-deployment
deployment.apps "ubuntu-deployment" deleted
[akaris@linux aws-env]$ 
~~~

Up to here, everything as expected.

3. Go to the developer console, "Add": under $URL/add/ns/default
   Select "From Catalog": under $URL/catalog/ns/default
   Filter for "namespace-template": under $URL/catalog/ns/default?keyword=namespace-template
   Click on the item and select "Instantiate template": under $URL/catalog/instantiate-template?template=namespace-template&template-ns=default&preselected-ns=default
   Under "SERVIE_NAME", type "test" and then click "Create"
   You will be dropped back to "Topology" with  message:
---
Topology
No workloads found

To add content to your project, create an application, component or service using one of these options.
---
   No new namespace will be created. No message to the user about why this did not work.

   Repeat the steps for deployment-template-working: under $URL/catalog/ns/default?keyword=deployment-template-working
   Click on "Instantiate template": under $URL/catalog/instantiate-template?template=deployment-template-working&template-ns=default&preselected-ns=default
   Under IMAGE_NAME: "ubuntu" and under DEPLOYMENT_SCALE: 2
   Click "Create". You will be dropped at the "Deployment" view under: $URL/topology/ns/default/graph
   And the deployment is created

   Repeat the steps for deployment-template-failing: under $URL/catalog/ns/default?keyword=deployment-template-failing
   Click on "Instantiate template": under $URL/catalog/instantiate-template?template=deployment-template-failing&template-ns=default&preselected-ns=default
   Under IMAGE_NAME: "fedora" and under DEPLOYMENT_SCALE: 2
   Click "Create". You will be dropped at the "Deployment" view under: $URL/topology/ns/default/graph
   The deployment will not be created, nothing happens

Actual results:

See above


Expected results:

An error message.  For deployment-template-failing, the same message as from the CLI should be reported to the end user via the GUI
~~~
akaris@linux aws-env]$ oc process deployment-template-failing IMAGE_NAME=fedora DEPLOYMENT_SCALE=2  | oc create -f -
Error from server (BadRequest): Deployment in version "v1" cannot be handled as a Deployment: v1.Deployment.Spec: v1.DeploymentSpec.Replicas: readUint32: unexpected character: �, error found in #10 byte of ...|eplicas":"2","select|..., bigger context ...|yment","namespace":"default"},"spec":{"replicas":"2","selector":{"matchLabels":{"app":"fedora-pod"}}|...
~~~

For the namespace-template, we must relay an error message to the user, too.

Additional info: