Various refactoring and fixes

.github/workflows/sync-release-branches.yml

@@ -27,7 +27,7 @@ jobs:
           set -x;
           git show -s;
           git fetch origin --unshallow;
-          for branch in release-{4.6,4.7,4.8,4.9};
+          for branch in release-{4.5,4.6,4.7,4.8,4.9};
           do
             git checkout $branch;
             git show -s;

roles/capture_environment/tasks/main.yml

@@ -4,7 +4,6 @@
     oc version -o json
        | jq --raw-output '.openshiftVersion'
        > {{ artifact_extra_logs_dir }}/ocp.version
-  register: ocp_full_version
 
 - name: Store OpenShift YAML version
   shell:

roles/cluster_scaleup/tasks/main.yml

@@ -11,8 +11,7 @@
   when: cluster_has_machinetype.stdout != ""
   block:
     - name: Cluster has running machines with the requested machine type
-      debug:
-        msg: "Cluster has running machines with type '{{ cluster_scaleup_machine_instance_type }}', no need to scale-up"
+      debug: msg="Cluster has running machines with type '{{ cluster_scaleup_machine_instance_type }}', no need to scale-up"
 
     - name: Cluster already scaled-up, end the play
       # Warning: this will prevent using multiple roles side by side ...

roles/cluster_scaleup/tasks/scaleup.yml

@@ -83,5 +83,4 @@
     failed_when: false
 
   - name: Fail because of the cluster scale-up failed
-    fail:
-      msg: Failing because of cluster scale-up failed
+    fail: msg="Failing because of cluster scale-up failed"

roles/entitlement_test_wait/tasks/main.yml

@@ -47,8 +47,7 @@
     when: entitlement_inspect_on_failure == 'yes'
 
   - name: Failed because the entitlement test Pod did not succeed.
-    fail:
-      msg: Failed because the entitlement test Pod did not succeed.
+    fail: msg="Failed because the entitlement test Pod did not succeed."
 
   always:
   - name: Get the test Pod logs

roles/gpu_operator_deploy_custom_commit/tasks/build.yml

@@ -61,12 +61,11 @@
     failed_when: false
 
   - name: Suggest how to retry the build
-    debug:
-      msg: "Suggestion: retry with 'oc new-build bc/helper-image-builder -n gpu-operator-ci-utils'"
+    debug: msg="To retry the build, run 'oc new-build bc/helper-image-builder -n gpu-operator-ci-utils'"
 
   - name: Fail because the custom operator image failed to build
-    fail:
-      msg: Fail because the custom operator image failed to build
+    fail: msg="Fail because the custom operator image failed to build"
+
 
 - name: Apply the operator image builder script manifest
   command: oc apply -f "{{ gpu_operator_image_builder_script }}"
@@ -77,8 +76,8 @@
       register: builder_secret
   rescue:
     - name: Failed to find the builder-dockercfg secret
-      fail:
-        msg: "No builder-dockercfg secret in the 'gpu-operator-ci' namespace ..."
+      fail: msg="No builder-dockercfg secret in the 'gpu-operator-ci' namespace ..."
+
 
 - name: Delete the operator image builder pod, if any
   command: oc delete -f "{{ gpu_operator_image_builder_pod }}" --ignore-not-found=true
@@ -92,8 +91,7 @@
   when: has_operator_image.rc != 0
   block:
   - name: The operator image does not exist, build it
-    debug:
-      msg: The operator image does not exist, build it
+    debug: msg="The operator image does not exist, build it"
 
   - name: Apply the operator image builder pod manifest
     shell:
@@ -119,9 +117,7 @@
 
     - name: Fail if the operator image failed to be built
       when: "'Failed' in wait_image_builder_pod.stdout or 'Error' in wait_image_builder_pod.stdout"
-      fail:
-        msg: The operator image failed to build
-
+      fail: msg="The operator image failed to build"
 
     rescue:
     - name: Get info about the operator image build failure (debug)
@@ -133,5 +129,4 @@
       failed_when: false
 
     - name: Fail because the operator image failed to build
-      fail:
-        msg: Fail because the operator image failed to build
+      fail: msg="Fail because the operator image failed to build"

roles/gpu_operator_deploy_custom_commit/tasks/deploy.yml

@@ -17,6 +17,17 @@
                 "{{ gpu_operator_image_tag }}"
 
   rescue:
+  - name: "Get the state of the GPU operator image (debug: ImagePullBackOff)"
+    command: oc get imagestreamtag/ci-artifacts:gpu-operator-ci-image -n gpu-operator-ci
+    failed_when: false
+
+  - name: "Get the name of the GPU Operator pod image (debug: ImagePullBackOff)"
+    command:
+      oc get pods
+         -n gpu-operator-ci
+         -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.containers[*]}{.image}{" "}{end}{end}'
+    failed_when: false
+
   - name: Get the state of the GPU operator pod (debug)
     command: oc get pods -n gpu-operator-ci
     failed_when: false
@@ -27,5 +38,4 @@
     failed_when: false
 
   - name: Fail because the GPU Operator custom commit could not be deployed
-    fail:
-      msg: Fail because the GPU Operator custom commit could not be deployed
+    fail: msg="Fail because the GPU Operator custom commit could not be deployed"

roles/gpu_operator_deploy_from_operatorhub/tasks/main.yml

@@ -34,8 +34,7 @@
   when: gpu_operator_operatorhub_version != ''
 
 - name: "Create the OperatorHub subscription for {{ gpu_operator_csv_name }}"
-  debug:
-    msg: "{{ gpu_operator_csv_name }}"
+  debug: msg="{{ gpu_operator_csv_name }}"
 
 - name: "Create the OperatorHub subscription for {{ gpu_operator_csv_name }}"
   shell:
@@ -85,12 +84,10 @@
     failed_when: false
 
   - name: "Indicate where the Catalog-operator logs have been saved"
-    debug:
-      msg: "The logs of Catalog Operator have been saved in {{ artifact_extra_logs_dir }}/catalog_operator.log"
+    debug: msg="The logs of Catalog Operator have been saved in {{ artifact_extra_logs_dir }}/catalog_operator.log"
 
   - name: Failed because the GPU Operator could not be install from the Catalog Operator
-    fail:
-      msg: Failed because the GPU Operator could not be install from the Catalog Operator
+    fail: msg="Failed because the GPU Operator could not be install from the Catalog Operator"
 
 - name: Create a temporary file for the GPU Operator clusterpolicy
   ansible.builtin.tempfile:
@@ -143,8 +140,7 @@
     failed_when: false
 
   - name: Failing because the ClusterPolicy CR cannot be created
-    fail:
-      msg: Failed because the ClusterPolicy CR cannot be created
+    fail: msg="Failed because the ClusterPolicy CR cannot be created"
 
 - block:
   - name: Check if the namespace has the openshift.io/cluster-monitoring label

roles/gpu_operator_wait_deployment/tasks/main.yml

@@ -26,8 +26,7 @@
 - name: Ensure that nvidia-device-plugin-validation Pod has ran successfully
   command:
     oc get pods
-      --field-selector=metadata.name=nvidia-device-plugin-validation
-      --field-selector=status.phase=Succeeded
+      --field-selector=metadata.name=nvidia-device-plugin-validation,status.phase=Succeeded
       -n gpu-operator-resources
       -oname --no-headers
   register: has_deviceplugin_validation_pod
@@ -53,8 +52,7 @@
     failed_when: false
 
   - name: The GFD did not label the nodes
-    fail:
-      msg: The GFD did not label the nodes
+    fail: msg="The GFD did not label the nodes"
 
 - block:
   - name: Wait for the nvidia-dcgm-exporter Pod to respond appropriately
@@ -103,5 +101,4 @@
     failed_when: false
 
   - name: The DCGM does not correctly expose the GPU metrics
-    fail:
-      msg: The DCGM does not correctly expose the GPU metrics
+    fail: msg="The DCGM does not correctly expose the GPU metrics"

roles/local-ci_deploy/tasks/main.yml

@@ -56,5 +56,4 @@
       | oc apply -f-
 
 - name: "Finished: the local CI execution has been launched."
-  debug:
-    msg: oc get pod/ci-artifacts -n ci-artifacts; oc logs -f pod/ci-artifacts -n ci-artifacts
+  debug: msg="oc get pod/ci-artifacts -n ci-artifacts; oc logs -f pod/ci-artifacts -n ci-artifacts"

roles/nfd_deploy/tasks/main.yml

@@ -38,5 +38,4 @@
       failed_when: false
 
     - name: Failed when creating the NFD NodeFeatureDiscovery CR
-      fail:
-        msg: "{{ apply_nfd_cr }}"
+      fail: msg="{{ apply_nfd_cr }}"

roles/nfd_has_labels/tasks/main.yml

@@ -8,10 +8,7 @@
   rescue:
   - name: Check if NFD CR exists (debug)
     command: oc get NodeFeatureDiscovery -A
-    register: has_nfd_cr
-    failed_when: has_nfd_cr.stdout == ""
-    ignore_errors: true
+    failed_when: false
 
   - name: Fail because nodes do not have NFD labels
-    fail:
-      msg: Fail because nodes do not have NFD labels
+    fail: msg="Fail because nodes do not have NFD labels"

roles/nfd_test_wait_gpu/tasks/main.yml

@@ -28,11 +28,10 @@
     shell: oc get nodes --show-labels --selector='!node-role.kubernetes.io/master' | tr , '\n'
 
   - name: Failing because no GPU node showed up
-    fail:
-      msg: Failed because no GPU node showed up
+    fail: msg="Failed because no GPU node showed up"
+
     when: nfd_wait_gpu_nodes == 'yes'
 
   - name: Failing because no GPU node is available
-    fail:
-      msg: Failed because no GPU node is available
+    fail: msg="Failed because no GPU node is available"
     when: nfd_wait_gpu_nodes != 'yes'

toolbox/_common.sh

@@ -44,7 +44,7 @@ mkdir -p "$(dirname "${ANSIBLE_LOG_PATH}")"
 # Ansible caching directory
 
 if [ -z "${ANSIBLE_CACHE_PLUGIN_CONNECTION:-}" ]; then
-    export ANSIBLE_CACHE_PLUGIN_CONNECTION="${ARTIFACT_EXTRA_LOGS_DIR}/ansible_facts"
+    export ANSIBLE_CACHE_PLUGIN_CONNECTION="${ARTIFACT_DIR}/ansible_facts"
 fi
 echo "Using '${ANSIBLE_CACHE_PLUGIN_CONNECTION}' to store ansible facts."
 mkdir -p "${ANSIBLE_CACHE_PLUGIN_CONNECTION}"