acc: follow the base's per-command recording switch

The DMS tests now enable recording through the env var per command rather than
by editing databricks.yml, so the tests that turn it off had to do the same.
failed-delete drains the reads print_state.py makes against the service, which
would otherwise be left for the harness to diff.

Co-authored-by: Isaac <no-reply@databricks.com>

Author: shreyas-goenka

acceptance/bundle/dms/existing-state/output.txt

@@ -76,7 +76,7 @@ Resources: 1 created, 0 changed, 0 deleted, 0 unchanged
 === So it is accepted even with the deployment gone. The file's state is a no-op while recording, so the service having nothing means the job is created again under a fresh deployment
 >>> MSYS_NO_PATHCONV=1 [CLI] workspace delete /Workspace/Users/[USERNAME]/.bundle/dms-existing-state-[UNIQUE_NAME]/default/state/resources.deployment.json
 
->>> [CLI] bundle deploy
+>>> DATABRICKS_BUNDLE_RECORD_DEPLOYMENT_HISTORY=true [CLI] bundle deploy
 Uploading bundle files to /Workspace/Users/[USERNAME]/.bundle/dms-existing-state-[UNIQUE_NAME]/default/files...
 Created jobs.one
 Files: 2 uploaded, 0 deleted

acceptance/bundle/dms/existing-state/script

@@ -26,5 +26,5 @@ trace jq '{state_version, features}' .databricks/bundle/default/resources.json
 
 title "So it is accepted even with the deployment gone. The file's state is a no-op while recording, so the service having nothing means the job is created again under a fresh deployment"
 trace MSYS_NO_PATHCONV=1 $CLI workspace delete "/Workspace/Users/${CURRENT_USER_NAME}/.bundle/dms-existing-state-${UNIQUE_NAME}/default/state/resources.deployment.json"
-trace $CLI bundle deploy
+trace DATABRICKS_BUNDLE_RECORD_DEPLOYMENT_HISTORY=true $CLI bundle deploy
 trace print_requests.py --dms //api/2.0/bundle --oneline

acceptance/bundle/dms/failed-delete/output.txt

@@ -126,5 +126,24 @@ API message: Fault injected by test.
  "features": {
   "record_deployment_history": {}
  },
- "state": {}
+ "state": {
+  "resources.jobs.stuck": {
+   "__id__": "[NUMID]",
+   "state": {
+    "deployment": {
+     "deployment_id": "[NUMID]",
+     "kind": "BUNDLE",
+     "metadata_file_path": "/Workspace/Users/[USERNAME]/.bundle/dms-failed-delete-[UNIQUE_NAME]/default/state/metadata.json",
+     "version_id": "1"
+    },
+    "edit_mode": "UI_LOCKED",
+    "format": "MULTI_TASK",
+    "max_concurrent_runs": 1,
+    "name": "stuck",
+    "queue": {
+     "enabled": true
+    }
+   }
+  }
+ }
 }

acceptance/bundle/dms/failed-delete/script

@@ -10,3 +10,7 @@ trace fault.py "POST /api/2.2/jobs/delete" 400 0 1 INVALID_PARAMETER_VALUE
 trace musterr $CLI bundle destroy --auto-approve
 trace print_requests.py --dms //api/2.0/bundle
 trace print_state.py
+
+# print_state.py asks the service for the resources while recording, so drain those reads instead
+# of leaving them behind for the harness to diff.
+print_requests.py //api/2.0 > /dev/null

acceptance/bundle/dms/requires-recording/out.test.toml

@@ -14,14 +14,14 @@ Resources: 1 created, 0 changed, 0 deleted, 0 unchanged
 === Turning recording off is refused: the service holds this deployment, so deploying without recording would leave it describing resources that have moved on
 >>> update_file.py databricks.yml record_deployment_history: true record_deployment_history: false
 
->>> musterr [CLI] bundle deploy
+>>> DATABRICKS_BUNDLE_RECORD_DEPLOYMENT_HISTORY= [CLI] bundle deploy
 Error: this deployment was previously deployed with experimental.record_deployment_history, so the feature cannot be turned off now
 
 Set it back to true to deploy or destroy this bundle
 
 
 === Destroy is refused for the same reason, which is why the error says to put the setting back
->>> musterr [CLI] bundle destroy --auto-approve
+>>> DATABRICKS_BUNDLE_RECORD_DEPLOYMENT_HISTORY= [CLI] bundle destroy --auto-approve
 Error: this deployment was previously deployed with experimental.record_deployment_history, so the feature cannot be turned off now
 
 Set it back to true to deploy or destroy this bundle

acceptance/bundle/dms/requires-recording/output.txt

@@ -6,10 +6,10 @@ trace jq '.features' .databricks/bundle/default/resources.json
 
 title "Turning recording off is refused: the service holds this deployment, so deploying without recording would leave it describing resources that have moved on"
 trace update_file.py databricks.yml "record_deployment_history: true" "record_deployment_history: false"
-trace musterr $CLI bundle deploy
+musterr trace DATABRICKS_BUNDLE_RECORD_DEPLOYMENT_HISTORY= $CLI bundle deploy
 
 title "Destroy is refused for the same reason, which is why the error says to put the setting back"
-trace musterr $CLI bundle destroy --auto-approve
+musterr trace DATABRICKS_BUNDLE_RECORD_DEPLOYMENT_HISTORY= $CLI bundle destroy --auto-approve
 
 title "With recording back on, deploy and destroy work again"
 trace update_file.py databricks.yml "record_deployment_history: false" "record_deployment_history: true"