Merge pull request #6422 from john-westcott-iv/tower_job_wait_update

Initial cut at tower_job_wait conversion

Reviewed-by: https://github.com/apps/softwarefactory-project-zuul

awx_collection/plugins/modules/tower_job_wait.py

@@ -28,24 +28,33 @@ options:
         - ID of the job to monitor.
       required: True
       type: int
+    interval:
+      description:
+        - The interval in sections, to request an update from Tower.
+        - For backwards compatability if unset this will be set to the average of min and max intervals
+      required: False
+      default: 1
+      type: float
     min_interval:
       description:
         - Minimum interval in seconds, to request an update from Tower.
-      default: 1
+        - deprecated, use interval instead
       type: float
     max_interval:
       description:
         - Maximum interval in seconds, to request an update from Tower.
-      default: 30
+        - deprecated, use interval instead
       type: float
     timeout:
       description:
         - Maximum time in seconds to wait for a job to finish.
       type: int
-
-requirements:
-- ansible-tower-cli >= 3.0.2
-
+    tower_oauthtoken:
+      description:
+        - The Tower OAuth token to use.
+      required: False
+      type: str
+      version_added: "3.7"
 extends_documentation_fragment: awx.awx.auth
 '''
 
@@ -90,78 +99,94 @@ status:
 '''
 
 
-from ..module_utils.ansible_tower import TowerModule, tower_auth_config, tower_check_mode
-from ansible.module_utils.six import PY2
-from ansible.module_utils.six.moves import cStringIO as StringIO
-from codecs import getwriter
+from ..module_utils.tower_api import TowerModule
+import time
 
 
-try:
-    import tower_cli
-    import tower_cli.exceptions as exc
+def check_job(module, job_url):
+    response = module.get_endpoint(job_url)
+    if response['status_code'] != 200:
+        module.fail_json(msg="Unable to read job from Tower {0}: {1}".format(response['status_code'], module.extract_errors_from_response(response)))
 
-    from tower_cli.conf import settings
-except ImportError:
-    pass
+    # Since we were successful, extract the fields we want to return
+    for k in ('id', 'status', 'elapsed', 'started', 'finished'):
+        module.json_output[k] = response['json'].get(k)
+
+    # And finally return the payload
+    return response['json']
 
 
 def main():
+    # Any additional arguments that are not fields of the item can be added here
     argument_spec = dict(
         job_id=dict(type='int', required=True),
         timeout=dict(type='int'),
-        min_interval=dict(type='float', default=1),
-        max_interval=dict(type='float', default=30),
+        min_interval=dict(type='float'),
+        max_interval=dict(type='float'),
+        interval=dict(type='float', default=1),
     )
 
-    module = TowerModule(
-        argument_spec,
-        supports_check_mode=True
-    )
+    # Create a module for ourselves
+    module = TowerModule(argument_spec=argument_spec, supports_check_mode=True)
 
-    json_output = {}
-    fail_json = None
+    # Extract our parameters
+    job_id = module.params.get('job_id')
+    timeout = module.params.get('timeout')
+    min_interval = module.params.get('min_interval')
+    max_interval = module.params.get('max_interval')
+    interval = module.params.get('interval')
 
-    tower_auth = tower_auth_config(module)
-    with settings.runtime_values(**tower_auth):
-        tower_check_mode(module)
-        job = tower_cli.get_resource('job')
-        params = module.params.copy()
+    if min_interval is not None or max_interval is not None:
+        # We can't tell if we got the default or if someone actually set this to 1.
+        # For now if we find 1 and had a min or max then we will do the average logic.
+        if interval == 1:
+            if not min_interval:
+                min_interval = 1
+            if not max_interval:
+                max_interval = 30
+            interval = abs((min_interval + max_interval) / 2)
+        module.deprecate(
+            msg="Min and max interval have been deprecated, please use interval instead; interval will be set to {0}".format(interval),
+            version="3.7"
+        )
 
-        # tower-cli gets very noisy when monitoring.
-        # We pass in our our outfile to suppress the out during our monitor call.
-        if PY2:
-            outfile = getwriter('utf-8')(StringIO())
-        else:
-            outfile = StringIO()
-        params['outfile'] = outfile
+    # Attempt to look up job based on the provided id
+    job = module.get_one('jobs', **{
+        'data': {
+            'id': job_id,
+        }
+    })
 
-        job_id = params.get('job_id')
-        try:
-            result = job.monitor(job_id, **params)
-        except exc.Timeout:
-            result = job.status(job_id)
-            result['id'] = job_id
-            json_output['msg'] = 'Timeout waiting for job to finish.'
-            json_output['timeout'] = True
-        except exc.NotFound as excinfo:
-            fail_json = dict(msg='Unable to wait, no job_id {0} found: {1}'.format(job_id, excinfo), changed=False)
-        except exc.JobFailure as excinfo:
-            fail_json = dict(msg='Job with id={0} failed, error: {1}'.format(job_id, excinfo))
-            fail_json['success'] = False
-            result = job.get(job_id)
-            for k in ('id', 'status', 'elapsed', 'started', 'finished'):
-                fail_json[k] = result.get(k)
-        except (exc.ConnectionError, exc.BadRequest, exc.AuthError) as excinfo:
-            fail_json = dict(msg='Unable to wait for job: {0}'.format(excinfo), changed=False)
+    if job is None:
+        module.fail_json(msg='Unable to wait on job {0}; that ID does not exist in Tower.'.format(job_id))
 
-    if fail_json is not None:
-        module.fail_json(**fail_json)
+    job_url = job['url']
 
-    json_output['success'] = True
-    for k in ('id', 'status', 'elapsed', 'started', 'finished'):
-        json_output[k] = result.get(k)
+    # Grab our start time to compare against for the timeout
+    start = time.time()
 
-    module.exit_json(**json_output)
+    # Get the initial job status from Tower, this will exit if there are any issues with the HTTP call
+    result = check_job(module, job_url)
+
+    # Loop while the job is not yet completed
+    while not result['finished']:
+        # If we are past our time out fail with a message
+        if timeout and timeout < time.time() - start:
+            module.json_output['msg'] = "Monitoring aborted due to timeout"
+            module.fail_json(**module.json_output)
+
+        # Put the process to sleep for our interval
+        time.sleep(interval)
+
+        # Check the job again
+        result = check_job(module, job_url)
+
+    # If the job has failed, we want to raise an Exception for that so we get a non-zero response.
+    if result['failed']:
+        module.json_output['msg'] = 'Job with id {0} failed'.format(job_id)
+        module.fail_json(**module.json_output)
+
+    module.exit_json(**module.json_output)
 
 
 if __name__ == '__main__':

awx_collection/test/awx/test_job.py

@@ -18,7 +18,7 @@ def test_job_wait_successful(run_module, admin_user):
     assert result.pop('started', '')[:10] == str(job.started)[:10]
     assert result == {
         "status": "successful",
-        "success": True,
+        "changed": False,
         "elapsed": str(job.elapsed),
         "id": job.id
     }
@@ -36,10 +36,10 @@ def test_job_wait_failed(run_module, admin_user):
     assert result == {
         "status": "failed",
         "failed": True,
-        "success": False,
+        "changed": False,
         "elapsed": str(job.elapsed),
         "id": job.id,
-        "msg": "Job with id=1 failed, error: Job failed."
+        "msg": "Job with id 1 failed"
     }
 
 
@@ -50,7 +50,6 @@ def test_job_wait_not_found(run_module, admin_user):
     ), admin_user)
     result.pop('invocation', None)
     assert result == {
-        "changed": False,
         "failed": True,
-        "msg": "Unable to wait, no job_id 42 found: The requested object could not be found."
+        "msg": "Unable to wait on job 42; that ID does not exist in Tower."
     }

awx_collection/tests/integration/targets/tower_job_wait/tasks/main.yml

@@ -1,18 +1,50 @@
 ---
-- name: Launch a Job Template
-  tower_job_launch:
-    job_template: "Demo Job Template"
-  register: job
+- name: Generate random string for template and project
+  set_fact:
+    jt_name: "AWX-Collection-tests-tower_job_wait-long_running-{{ lookup('password', '/dev/null chars=ascii_letters length=16') }}"
+    proj_name: "AWX-Collection-tests-tower_job_wait-long_running-{{ lookup('password', '/dev/null chars=ascii_letters length=16') }}"
+
+- name: Assure that the demo project exists
+  tower_project:
+    name: "{{ proj_name }}"
+    scm_type: 'git'
+    scm_url: 'https://github.com/ansible/test-playbooks.git'
+    scm_update_on_launch: true
+    organization: Default
+
+- name: Create a job template
+  tower_job_template:
+    name: "{{ jt_name }}"
+    playbook: "sleep.yml"
+    job_type: run
+    project: "{{ proj_name }}"
+    inventory: "Demo Inventory"
+
+- name: Check deprecation warnings
+  tower_job_wait:
+    min_interval: 10
+    max_interval: 20
+    job_id: "99999999"
+  register: result
+  ignore_errors: true
 
 - assert:
     that:
-      - "job is changed"
-      - "job.status == 'pending'"
+      - "'Min and max interval have been deprecated, please use interval instead; interval will be set to 15'"
 
-- name: Wait for the Job to finish
+- name: Validate that interval superceeds min/max
   tower_job_wait:
-    job_id: "{{ job.id }}"
-    timeout: 60
+    min_interval: 10
+    max_interval: 20
+    interval: 12
+    job_id: "99999999"
+  register: result
+  ignore_errors: true
+
+- assert:
+    that:
+      - "result.msg =='Unable to wait on job 99999999; that ID does not exist in Tower.' or
+        'min and max interval have been depricated, please use interval instead, interval will be set to 12'"
 
 - name: Check module fails with correct msg
   tower_job_wait:
@@ -22,4 +54,82 @@
 
 - assert:
     that:
-      - "result.msg =='Unable to wait, no job_id 99999999 found: The requested object could not be found.'"
+      - result is failed
+      - "result.msg =='Unable to wait, no job_id 99999999 found: The requested object could not be found.' or
+        'Unable to wait on job 99999999; that ID does not exist in Tower.'"
+
+- name: Launch Demo Job Template (take happy path)
+  tower_job_launch:
+    job_template: "Demo Job Template"
+  register: job
+
+- assert:
+    that:
+      - job is changed
+
+- name: Wait for the Job to finish
+  tower_job_wait:
+    job_id: "{{ job.id }}"
+  register: wait_results
+
+# Make sure it worked and that we have some data in our results
+- assert:
+    that:
+      - wait_results is successful
+      - "'elapsed' in wait_results"
+      - "'id' in wait_results"
+
+- name: Launch a long running job
+  tower_job_launch:
+    job_template: "{{ jt_name }}"
+  register: job
+
+- assert:
+    that:
+      - job is changed
+
+- name: Timeout waiting for the job to complete
+  tower_job_wait:
+    job_id: "{{ job.id }}"
+    timeout: 5
+  ignore_errors: true
+  register: wait_results
+
+# Make sure that we failed and that we have some data in our results
+- assert:
+    that:
+      - "wait_results.msg == 'Monitoring aborted due to timeout' or 'Timeout waiting for job to finish.'"
+      - "'id' in wait_results"
+
+- name: Async cancel the long running job
+  tower_job_cancel:
+    job_id: "{{ job.id }}"
+  async: 3600
+  poll: 0
+
+- name: Wait for the job to exit on cancel
+  tower_job_wait:
+    job_id: "{{ job.id }}"
+  register: wait_results
+  ignore_errors: true
+
+- assert:
+    that:
+      - wait_results is failed
+      - 'wait_results.status == "canceled"'
+      - "wait_results.msg == 'Job with id {{ job.id }} failed' or 'Job with id={{ job.id }} failed, error: Job failed.'"
+
+- name: Delete the job template
+  tower_job_template:
+    name: "{{ jt_name }}"
+    playbook: "sleep.yml"
+    job_type: run
+    project: "{{ proj_name }}"
+    inventory: "Demo Inventory"
+    state: absent
+
+- name: Delete the project
+  tower_project:
+    name: "{{ proj_name }}"
+    organization: Default
+    state: absent