make cluster creation/deletion async by default (#16185)

src/lightning_app/cli/cmd_clusters.py

@@ -109,7 +109,7 @@ class AWSClusterManager:
         region: str = "us-east-1",
         external_id: str = None,
         edit_before_creation: bool = False,
-        do_async: bool = False,
+        do_async: bool = True,
     ) -> None:
         """request Lightning AI BYOC compute cluster creation.
 
@@ -192,7 +192,7 @@ class AWSClusterManager:
         console = Console()
         console.print(clusters.as_table())
 
-    def delete(self, cluster_id: str, force: bool = False, do_async: bool = False) -> None:
+    def delete(self, cluster_id: str, force: bool = False, do_async: bool = True) -> None:
         if force:
             click.echo(
                 """

src/lightning_app/cli/lightning_cli_create.py

@@ -50,13 +50,13 @@ def create() -> None:
     help="Edit the cluster specs before submitting them to the API server.",
 )
 @click.option(
-    "--async",
-    "do_async",
+    "--sync",
+    "do_sync",
     type=bool,
     required=False,
     default=False,
     is_flag=True,
-    help="This flag makes the CLI return immediately and lets the cluster creation happen in the background.",
+    help="This flag makes the CLI wait until cluster creation completes.",
 )
 def create_cluster(
     cluster_id: str,
@@ -66,7 +66,7 @@ def create_cluster(
     provider: str,
     edit_before_creation: bool,
     enable_performance: bool,
-    do_async: bool,
+    do_sync: bool,
     **kwargs: Any,
 ) -> None:
     """Create a Lightning AI BYOC compute cluster with your cloud provider credentials."""
@@ -81,7 +81,7 @@ def create_cluster(
         external_id=external_id,
         edit_before_creation=edit_before_creation,
         cost_savings=not enable_performance,
-        do_async=do_async,
+        do_async=not do_sync,
     )
 
 

src/lightning_app/cli/lightning_cli_delete.py

@@ -20,15 +20,15 @@ def delete() -> None:
 @delete.command("cluster")
 @click.argument("cluster", type=str)
 @click.option(
-    "--async",
-    "do_async",
+    "--sync",
+    "do_sync",
     type=bool,
     required=False,
     default=False,
     is_flag=True,
-    help="This flag makes the CLI return immediately and lets the cluster deletion happen in the background",
+    help="This flag makes the CLI wait until cluster deletion completes.",
 )
-def delete_cluster(cluster: str, force: bool = False, do_async: bool = False) -> None:
+def delete_cluster(cluster: str, force: bool = False, do_sync: bool = False) -> None:
     """Delete a Lightning AI BYOC cluster and all associated cloud provider resources.
 
     Deleting a cluster also deletes all apps that were started on the cluster.
@@ -44,7 +44,7 @@ def delete_cluster(cluster: str, force: bool = False, do_async: bool = False) ->
     VPC components, etc. are irreversibly deleted and cannot be recovered!
     """
     cluster_manager = AWSClusterManager()
-    cluster_manager.delete(cluster_id=cluster, force=force, do_async=do_async)
+    cluster_manager.delete(cluster_id=cluster, force=force, do_async=not do_sync)
 
 
 def _find_cluster_for_user(app_name: str, cluster_id: Optional[str]) -> str:

tests/tests_app/cli/test_cli.py

@@ -88,6 +88,7 @@ def test_create_cluster(create_command: mock.MagicMock):
             "dummy",
             "--role-arn",
             "arn:aws:iam::1234567890:role/lai-byoc",
+            "--sync",
         ],
     )
 
@@ -124,7 +125,7 @@ def test_list_clusters(list_command: mock.MagicMock):
 @mock.patch("lightning_app.cli.cmd_clusters.AWSClusterManager.delete")
 def test_delete_cluster(delete: mock.MagicMock):
     runner = CliRunner()
-    runner.invoke(delete_cluster, ["test-7"])
+    runner.invoke(delete_cluster, ["test-7", "--sync"])
 
     delete.assert_called_once_with(cluster_id="test-7", force=False, do_async=False)