Ecs service add features (#50059)

* Support UpdateService forceNewDeployment in ecs_service module

* Fixes for review

* Add force_new_deployment option to ecs_service.py

cherrypicks changes from via/ansible
Adds tests for pull request #42518
fixes backwards compatability with boto<1.8.4

* WIP commit so I don't have to stash

* WIP commit for healthcheck grace period

* WIP commit; ecs_module handles service registries

* Fix bad check for desired_count

* Add scheduling strategy test, comment out service registry test

* Fix names in ecs_cluster role main task.

* move full test run back to the end

* Change botocore version for full test to support scheduling strategy

* fix bug with desired_count==0 in amazon/ecs_service

* Fix changed checking for scheduling strategy DAEMON in ecs_service

* Pass testS

* Fix some unhelpful comments

* Add changelog for ecs_service

test/integration/targets/ecs_cluster/playbooks/roles/ecs_cluster/tasks/main.yml

@@ -479,6 +479,95 @@
           - "update_ecs_service_with_vpc.service.networkConfiguration.awsvpcConfiguration.subnets|length == 2"
           - "update_ecs_service_with_vpc.service.networkConfiguration.awsvpcConfiguration.securityGroups|length == 1"
 
+    - name: create ecs_service using health_check_grace_period_seconds
+      ecs_service:
+        name: "{{ ecs_service_name }}-mft"
+        cluster: "{{ ecs_cluster_name }}"
+        load_balancers:
+          - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+            containerName: "{{ ecs_task_name }}"
+            containerPort: "{{ ecs_task_container_port }}"
+        task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+        scheduling_strategy: "REPLICA"
+        health_check_grace_period_seconds: 10
+        desired_count: 1
+        state: present
+        <<: *aws_connection_info
+      register: ecs_service_creation_hcgp
+
+
+    - name: health_check_grace_period_seconds sets HealthChecGracePeriodSeconds
+      assert:
+        that:
+          - ecs_service_creation_hcgp.changed
+          - "{{ecs_service_creation_hcgp.service.healthCheckGracePeriodSeconds}} == 10"
+
+    - name: update ecs_service using health_check_grace_period_seconds
+      ecs_service:
+        name: "{{ ecs_service_name }}-mft"
+        cluster: "{{ ecs_cluster_name }}"
+        load_balancers:
+          - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+            containerName: "{{ ecs_task_name }}"
+            containerPort: "{{ ecs_task_container_port }}"
+        task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+        desired_count: 1
+        health_check_grace_period_seconds: 30
+        state: present
+        <<: *aws_connection_info
+      register: ecs_service_creation_hcgp2
+      ignore_errors: no
+
+    - name: check that module returns success
+      assert:
+        that:
+          - ecs_service_creation_hcgp2.changed
+          - "{{ecs_service_creation_hcgp2.service.healthCheckGracePeriodSeconds}} == 30"
+
+# until ansible supports service registries, this test can't run.
+#   - name: update ecs_service using service_registries
+#     ecs_service:
+#       name: "{{ ecs_service_name }}-service-registries"
+#       cluster: "{{ ecs_cluster_name }}"
+#       load_balancers:
+#         - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+#           containerName: "{{ ecs_task_name }}"
+#           containerPort: "{{ ecs_task_container_port }}"
+#       service_registries:
+#         - containerName: "{{ ecs_task_name }}"
+#           containerPort: "{{ ecs_task_container_port }}"
+#           ### TODO: Figure out how to get a service registry ARN without a service registry module.
+#           registryArn: "{{ ecs_task_service_registry_arn }}"
+#       task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+#       desired_count: 1
+#       state: present
+#       <<: *aws_connection_info
+#     register: ecs_service_creation_sr
+#     ignore_errors: yes
+
+#   - name: dump sr output
+#     debug: var=ecs_service_creation_sr
+
+#   - name: check that module returns success
+#     assert:
+#       that:
+#         - ecs_service_creation_sr.changed
+
+    - name: update ecs_service using REPLICA scheduling_strategy
+      ecs_service:
+        name: "{{ ecs_service_name }}-replica"
+        cluster: "{{ ecs_cluster_name }}"
+        load_balancers:
+          - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+            containerName: "{{ ecs_task_name }}"
+            containerPort: "{{ ecs_task_container_port }}"
+        scheduling_strategy: "REPLICA"
+        task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+        desired_count: 1
+        state: present
+        <<: *aws_connection_info
+      register: ecs_service_creation_replica
+
     - name: obtain facts for all ECS services in the cluster
       ecs_service_facts:
         cluster: "{{ ecs_cluster_name }}"
@@ -728,6 +817,56 @@
       ignore_errors: yes
       register: ecs_service_scale_down
 
+    - name: scale down multifunction-test service
+      ecs_service:
+        name: "{{ ecs_service_name }}-mft"
+        cluster: "{{ ecs_cluster_name }}"
+        state: present
+        load_balancers:
+          - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+            containerName: "{{ ecs_task_name }}"
+            containerPort: "{{ ecs_task_container_port }}"
+        task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+        desired_count: 0
+        <<: *aws_connection_info
+      ignore_errors: yes
+      register: ecs_service_scale_down
+
+
+
+    - name: scale down scheduling_strategy service
+      ecs_service:
+        name: "{{ ecs_service_name }}-replica"
+        cluster: "{{ ecs_cluster_name }}"
+        state: present
+        load_balancers:
+          - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+            containerName: "{{ ecs_task_name }}"
+            containerPort: "{{ ecs_task_container_port }}"
+        task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+        desired_count: 0
+        <<: *aws_connection_info
+      ignore_errors: yes
+      register: ecs_service_scale_down
+
+
+# until ansible supports service registries, the test for it can't run and this
+# scale down is not needed
+#   - name: scale down service_registries service
+#     ecs_service:
+#       name: "{{ ecs_service_name }}-service-registries"
+#       cluster: "{{ ecs_cluster_name }}"
+#       state: present
+#       load_balancers:
+#         - targetGroupArn: "{{ elb_target_group_instance.target_group_arn }}"
+#           containerName: "{{ ecs_task_name }}"
+#           containerPort: "{{ ecs_task_container_port }}"
+#       task_definition: "{{ ecs_task_name }}:{{ ecs_task_definition.taskdefinition.revision }}"
+#       desired_count: 0
+#       <<: *aws_connection_info
+#     ignore_errors: yes
+#     register: ecs_service_scale_down
+
     - name: scale down Fargate ECS service
       ecs_service:
         state: present
@@ -761,6 +900,32 @@
         <<: *aws_connection_info
       ignore_errors: yes
 
+    - name: remove mft ecs service
+      ecs_service:
+        state: absent
+        cluster: "{{ ecs_cluster_name }}"
+        name: "{{ ecs_service_name }}-mft"
+        <<: *aws_connection_info
+      ignore_errors: yes
+
+    - name: remove scheduling_strategy ecs service
+      ecs_service:
+        state: absent
+        cluster: "{{ ecs_cluster_name }}"
+        name: "{{ ecs_service_name }}-replica"
+        <<: *aws_connection_info
+      ignore_errors: yes
+
+# until ansible supports service registries, the test for it can't run and this
+# removal is not needed
+#   - name: remove service_registries ecs service
+#     ecs_service:
+#       state: absent
+#       cluster: "{{ ecs_cluster_name }}"
+#       name: "{{ ecs_service_name }}-service-registries"
+#       <<: *aws_connection_info
+#     ignore_errors: yes
+
     - name: remove fargate ECS service 
       ecs_service:
         state: absent
@@ -769,7 +934,7 @@
         <<: *aws_connection_info
       ignore_errors: yes
       register: ecs_fargate_service_network_with_awsvpc
- 
+
     - name: remove ecs task definition
       ecs_taskdefinition:
         containers: "{{ ecs_task_containers }}"

test/integration/targets/ecs_cluster/runme.sh

@@ -12,6 +12,8 @@ trap 'rm -rf "${MYTMPDIR}"' EXIT
 # but for the python3 tests we need virtualenv to use python3
 PYTHON=${ANSIBLE_TEST_PYTHON_INTERPRETER:-python}
 
+
+
 # Test graceful failure for older versions of botocore
 virtualenv --system-site-packages --python "${PYTHON}" "${MYTMPDIR}/botocore-1.7.40"
 source "${MYTMPDIR}/botocore-1.7.40/bin/activate"
@@ -42,5 +44,5 @@ ansible-playbook -i ../../inventory -e @../../integration_config.yml -e @../../c
 # Run full test suite
 virtualenv --system-site-packages --python "${PYTHON}" "${MYTMPDIR}/botocore-recent"
 source "${MYTMPDIR}/botocore-recent/bin/activate"
-$PYTHON -m pip install 'botocore>=1.8.4' boto3
+$PYTHON -m pip install 'botocore>=1.10.37' boto3 # version 1.10.37 for scheduling strategy
 ansible-playbook -i ../../inventory -e @../../integration_config.yml -e @../../cloud-config-aws.yml -v playbooks/full_test.yml "$@"