Handle successful resource update retries correctly

CHANGELOG.md

@@ -17,6 +17,7 @@ ENHANCEMENTS:
 * Update the version of `super-linter` used in the `build_validation_develop` workflow to 8.7.0 ([#4957](https://github.com/microsoft/AzureTRE/issues/4957))
 
 BUG FIXES:
+* Fix API resource update retries so a successful retry returns the patched resource to the original caller. ([#5025](https://github.com/microsoft/AzureTRE/pull/5025))
 * Fix workspace deletion when backup is enabled for the base, unrestricted and airlock-import-review workspaces by adding a `delete_backups_on_uninstall` flag and a pre-teardown backup cleanup (`remove_backup.sh`) that stops protection and either deletes or retains the Recovery Services Vault, so deletion works with Azure secure-by-default soft delete ([#4962](https://github.com/microsoft/AzureTRE/issues/4962))
 * Fix Nexus shared service security: fetch admin password from Key Vault at runtime via managed identity (IMDS) instead of embedding it in the VM Run Command script content. Fix `deploy_nexus_container.sh` short-circuit path to fail loudly if the container does not start. (`sonatype-nexus` 3.10.0) ([#4983](https://github.com/microsoft/AzureTRE/pull/4983))
 * Fix UI TypeScript deprecation warning by updating `moduleResolution` to `bundler` in `tsconfig.json`. ([#4968](https://github.com/microsoft/AzureTRE/issues/4968))

api_app/_version.py

@@ -1 +1 @@
-__version__ = "0.26.0"
+__version__ = "0.26.1"

api_app/service_bus/helpers.py

@@ -130,7 +130,7 @@ async def try_update_with_retries(num_retries: int, attempt_count: int, resource
     except CosmosAccessConditionFailedError as e:
         logger.warning(f"Etag mismatch for {resource_to_update_id}. Retrying.")
         if attempt_count < num_retries:
-            await try_update_with_retries(
+            return await try_update_with_retries(
                 num_retries=num_retries,
                 attempt_count=(attempt_count + 1),
                 resource_repo=resource_repo,

api_app/tests_ma/test_service_bus/test_resource_request_sender.py

@@ -148,6 +148,49 @@ async def test_multi_step_document_sends_first_step(
     )
 
 
+@patch("service_bus.resource_request_sender.ResourceHistoryRepository.create")
+@patch("service_bus.resource_request_sender.ResourceRepository.create")
+@patch("service_bus.resource_request_sender.ResourceTemplateRepository.create")
+async def test_multi_step_document_retries_then_succeeds(
+    resource_template_repo,
+    resource_repo,
+    resource_history_repo,
+    basic_shared_service,
+    basic_shared_service_template,
+    test_user,
+    multi_step_resource_template,
+    primary_resource
+):
+
+    resource_repo.get_resource_by_id.return_value = basic_shared_service
+    resource_template_repo.get_template_by_name_and_version.return_value = (
+        basic_shared_service_template
+    )
+    updated_resource = basic_shared_service.copy(update={"etag": "updated-etag"})
+    resource_repo.patch_resource.side_effect = [
+        CosmosAccessConditionFailedError(),
+        (updated_resource, basic_shared_service_template),
+    ]
+
+    result = await try_update_with_retries(
+        num_retries=5,
+        attempt_count=0,
+        resource_repo=resource_repo,
+        resource_template_repo=resource_template_repo,
+        user=test_user,
+        resource_to_update_id="resource-id",
+        template_step=multi_step_resource_template.pipeline.install[0],
+        resource_history_repo=resource_history_repo,
+        primary_resource=primary_resource,
+        primary_parent_workspace=None,
+        primary_parent_workspace_svc=None
+    )
+
+    assert result == updated_resource
+    assert len(resource_repo.patch_resource.mock_calls) == 2
+    assert len(resource_repo.get_resource_by_id.mock_calls) == 2
+
+
 @patch("service_bus.resource_request_sender.ResourceHistoryRepository.create")
 @patch("service_bus.resource_request_sender.ResourceRepository.create")
 @patch("service_bus.resource_request_sender.ResourceTemplateRepository.create")