fix ecs_cluster integration test (#1145) (#1153)

[PR #1145/35f4f592 backport][stable-2] fix ecs_cluster integration test

This is a backport of PR #1145 as merged into main (35f4f59).
SUMMARY

ecs_cluster: make ecs_cluster integration test work again - as it is

bugs I've hit and must be fixed to complete this challenge

ecs_taskdefinition: fix change detection of changing launch_type parameter
ecs_service:

compare of task_definition never works and results always in a changed task
change detect of health_check_grace_period_seconds was never implemented, but tested and failing, after the task_definition is compared correctly







ref: #1142
ISSUE TYPE


Bugfix Pull Request

COMPONENT NAME
ecs_taskdefinition
ecs_service
ADDITIONAL INFORMATION
basically the existing test tasks are not changed. just sorted and removed what was marked as fixme because it's simple not possible (changing network settings of a created service).

Reviewed-by: Markus Bergholz <[email protected]>

changelogs/fragments/0000-ecs_taskdefinition_fix.yml

@@ -0,0 +1,4 @@
+bugfixes:
+  - ecs_taskdefinition - fix broken change detect of ``launch_type`` parameter (https://github.com/ansible-collections/community.aws/pull/1145).
+  - ecs_service - fix broken compare of ``task_definition`` that results always in a changed task (https://github.com/ansible-collections/community.aws/pull/1145).
+  - ecs_service - add missing change detect of ``health_check_grace_period_seconds`` parameter (https://github.com/ansible-collections/community.aws/pull/1145).

plugins/modules/ecs_service.py

@@ -549,7 +549,14 @@ def describe_service(self, cluster_name, service_name):
         raise Exception("Unknown problem describing service %s." % service_name)
 
     def is_matching_service(self, expected, existing):
-        if expected['task_definition'] != existing['taskDefinition']:
+        # aws returns the arn of the task definition
+        #   arn:aws:ecs:eu-central-1:123456789:task-definition/ansible-fargate-nginx:3
+        # but the user is just entering
+        #   ansible-fargate-nginx:3
+        if expected['task_definition'] != existing['taskDefinition'].split('/')[-1]:
+            return False
+
+        if expected.get('health_check_grace_period_seconds') != existing.get('healthCheckGracePeriodSeconds'):
             return False
 
         if (expected['load_balancers'] or []) != existing['loadBalancers']:

plugins/modules/ecs_taskdefinition.py

@@ -956,7 +956,7 @@ def _task_definition_matches(requested_volumes, requested_containers, requested_
                 if requested_task_role_arn != td.get('taskRoleArn', ""):
                     return None
 
-                if requested_launch_type is not None and requested_launch_type not in td.get('compatibilities', []):
+                if requested_launch_type is not None and requested_launch_type not in td.get('requiresCompatibilities', []):
                     return None
 
                 existing_volumes = td.get('volumes', []) or []