hi

Signed-off-by: Judy Ng <[email protected]>
judysng · Mar 26, 2024 · 66a4346 · 66a4346
1 parent b7761cf
commit 66a4346
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 7 deletions.
diff --git a/tests/integration-tests/configs/scaling_stress_test.yaml b/tests/integration-tests/configs/scaling_stress_test.yaml
@@ -2,7 +2,7 @@ test-suites:
   performance_tests:
     test_scaling.py::test_scaling_stress_test:
       dimensions:
-        - regions: [ "euw1-az3" ]
-          instances: [ "p3.2xlarge" ]
+        - regions: [ "us-east-1" ]
+          instances: [ "c5.large" ]
           oss: [ "alinux2" ]
           schedulers: [ "slurm" ]
diff --git a/tests/integration-tests/tests/common/scaling_common.py b/tests/integration-tests/tests/common/scaling_common.py
@@ -78,8 +78,8 @@ def get_bootstrap_errors(remote_command_executor: RemoteCommandExecutor, cluster
             Filters=[{"Name": "private-ip-address", "Values": [ip_address]}]
         )["Reservations"][0]["Instances"][0]["InstanceId"]
         logging.info(f"Instance {instance_id} had bootstrap errors. Check the logs for details.")
-        compute_node_log = client.get_console_output(InstanceId=instance_id)["Output"]
-        with open(f"bootstrap_errors/{ip_address}-{cluster_name}-{instance_id}-bootstrap-error.txt", "w") as file:
+        compute_node_log = client.get_console_output(InstanceId=instance_id, Latest=True)["Output"]
+        with open(f"bootstrap_errors/{region}-{cluster_name}-{instance_id}-bootstrap-error.txt", "w") as file:
             file.write(compute_node_log)
 
 

diff --git a/tests/integration-tests/tests/performance_tests/test_scaling.py b/tests/integration-tests/tests/performance_tests/test_scaling.py
@@ -157,7 +157,7 @@ def test_scaling_stress_test(
 
 
 @pytest.mark.usefixtures("scheduler")
-@pytest.mark.parametrize("scaling_strategy", ["all-or-nothing", "best-effort"])
+@pytest.mark.parametrize("scaling_strategy", ["all-or-nothing"])
 def test_static_scaling_stress_test(
     test_datadir,
     instance,
@@ -287,8 +287,8 @@ def _scale_up_and_down(
     # Get the compute node logs for bootstrap errors if compute nodes did not scale up to scaling target within time
     if scaling_target not in compute_nodes_time_series_up:
         get_bootstrap_errors(remote_command_executor, cluster.name, region)
-        raise Exception(f"Cluster did not scale up to {scaling_target} nodes."
-                        f"Check the compute node logs for any bootstrap errors.")
+        raise Exception(f"Cluster did not scale up to {scaling_target} nodes. "
+                        f"Check the compute node logs for any bootstrap errors in the test artifacts.")
 
     # Extract scale up duration and timestamp from the monitoring metrics collected above
     _, scale_up_time_ec2 = _get_scaling_time(ec2_capacity_time_series_up, timestamps, scaling_target, start_time)