Mark shiftstack test failures via test_stage_unstable_flag

Testing stages write test_stage_unstable_flag on failure via rescue
blocks. Soft verification checks keep using stage_unstable_flag without
triggering job failure.

Remove end-of-playbook fail from ocp_testing; ci-framework shiftstack
role fails after artifact rsync when test_stage_unstable_flag is present.
Pair with ci-framework-jobs failure-message: unstable for yellow Zuul UI.

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: tusharjadhav3302

collection/stages/roles/conformance_test/tasks/run_conformance_test.yml

@@ -104,7 +104,7 @@
     - name: Mark the openshift tests as UNSTABLE
       ansible.builtin.include_role:
         name: tools_stage_results
-        tasks_from: mark_stage_unstable.yml
+        tasks_from: mark_test_stage_unstable.yml
       vars:
         unstable_msg: >-
           The openshift test Suite failed.

collection/stages/roles/cpms_test/tasks/run_cpms_test.yml

@@ -28,7 +28,7 @@
     - name: Mark the CPMS {{ cpms_tests_type }} tests as UNSTABLE
       ansible.builtin.include_role:
         name: tools_stage_results
-        tasks_from: mark_stage_unstable.yml
+        tasks_from: mark_test_stage_unstable.yml
       vars:
         unstable_msg: >-
           The {{ cpms_test_name }} {{ cpms_tests_type }} test suite failed.

collection/stages/roles/csi_tests/tasks/run_csi_test.yml

@@ -28,7 +28,7 @@
     - name: Mark the openshift tests as UNSTABLE
       ansible.builtin.include_role:
         name: tools_stage_results
-        tasks_from: mark_stage_unstable.yml
+        tasks_from: mark_test_stage_unstable.yml
       vars:
         unstable_msg: >-
           The openshift test Suite failed.

collection/stages/roles/egressip_tests/tasks/run_egressip_tests.yml

@@ -93,7 +93,7 @@
     - name: Mark the egressip tests as UNSTABLE
       ansible.builtin.include_role:
         name: tools_stage_results
-        tasks_from: mark_stage_unstable.yml
+        tasks_from: mark_test_stage_unstable.yml
       vars:
         unstable_msg: >-
           The EgressIP test suite failed.

collection/stages/roles/openstack_test/tasks/run_openstack_test.yml

@@ -207,7 +207,7 @@
     - name: Mark the openshift tests as UNSTABLE
       ansible.builtin.include_role:
         name: tools_stage_results
-        tasks_from: mark_stage_unstable.yml
+        tasks_from: mark_test_stage_unstable.yml
       vars:
         unstable_msg: >-
           The openshift test Suite failed.

collection/stages/roles/verification/defaults/main.yml

@@ -5,7 +5,7 @@ cinder_csi_project_name: cinder-test # project for testing cinder-csi sc provide
 manila_project_name: manila-test # project for testing manila sc provider
 demo_project_name: demo # project for testing demo app
 verify_existing_namespace: false
-soft_checks: true # run checks that won't stop the playbook but will generate the jenkins message to mark the run as unstable.
+soft_checks: true # run checks that won't stop the playbook but will mark the run as UNSTABLE via stage_unstable_flag.
 alerts_to_ignore:
   # - 'Watchdog' # Should not be considered as severity is none
   - 'NodeClockNotSynchronising'

collection/tools/roles/tools_stage_results/tasks/mark_test_stage_unstable.yml

@@ -0,0 +1,17 @@
+---
+- name: Set the test stage unstable var as true
+  ansible.builtin.set_fact:
+    test_stage_unstable: true
+    unstable_warning: "Warning! {{ unstable_msg }}"
+
+- name: Print the test stage UNSTABLE message
+  ansible.builtin.debug:
+    msg: "{{ unstable_warning }}"
+
+- name: Mark the test stage result as UNSTABLE
+  ansible.builtin.lineinfile:
+    path: "{{ test_stage_unstable_flag }}"
+    line: "{{ unstable_warning }}"
+    create: yes
+    mode: u=rw,g=rw,o=r
+  delegate_to: localhost

configs/global.yml

@@ -32,9 +32,14 @@ report_dir: "{{ artifacts_dir }}/test_results"
 
 # Sets the Ansible Controller Node machine's HOME environment variable
 controller_home_dir: "{{ lookup('ansible.builtin.env', 'HOME') }}"
-# The resources_file and stage_unstable_flag are in the Ansible control node's HOME environment variable
+# The resources_file and stage_unstable flags are on the Ansible control node's HOME.
+# test_stage_unstable_flag is written only by testing-stage rescues. The ci-framework
+# shiftstack role fails the Zuul job when this file is present in rsynced artifacts;
+# pair with failure-message: unstable on monolithic shiftstack Zuul jobs for yellow UI.
+# stage_unstable_flag is used by soft verification/install workarounds (no end-of-run fail).
 resources_file: "{{ controller_home_dir }}/artifacts/resources.yml"
 stage_unstable_flag: "{{ controller_home_dir }}/artifacts/stage_unstable_flag"
+test_stage_unstable_flag: "{{ controller_home_dir }}/artifacts/test_stage_unstable_flag"
 
 admin_cloud: "default"
 admin_env_file: "{{ osp_config_dir }}/{{ admin_cloud }}rc"

playbooks/ocp_testing.yaml

@@ -10,6 +10,11 @@
         path: "{{ stage_unstable_flag }}"
         state: absent
 
+    - name: Cleanup workspace flags - Remove the test_stage_unstable_flag if it exists
+      ansible.builtin.file:
+        path: "{{ test_stage_unstable_flag }}"
+        state: absent
+
     - name: Update ocp_deployment_topology with overrides if defined
       # Apply ocp_deployment_topology_override on top of the base topology
       # The base configuration comes from `configs/global.yml`