start cluster checks thread without delay (#1131)

Co-authored-by: Alexandra Belousov <sashabelousovrh@Alexandras-MacBook-Pro.local>
run-house · Aug 11, 2024 · a6a4981 · a6a4981
1 parent e7692ed
commit a6a4981
Show file tree

Hide file tree

Showing 4 changed files with 1 addition and 6 deletions.
diff --git a/runhouse/constants.py b/runhouse/constants.py
@@ -84,5 +84,4 @@
 SERVER_LOGS_FILE_NAME = "server.log"
 DEFAULT_SURFACED_LOG_LENGTH = 20
 # Constants for schedulers
-SCHEDULERS_DELAY = 2 * MINUTE
 INCREASED_INTERVAL = 1 * HOUR
diff --git a/runhouse/servers/cluster_servlet.py b/runhouse/servers/cluster_servlet.py
@@ -15,7 +15,6 @@
     DEFAULT_STATUS_CHECK_INTERVAL,
     INCREASED_INTERVAL,
     INCREASED_STATUS_CHECK_INTERVAL,
-    SCHEDULERS_DELAY,
     SERVER_LOGFILE,
     SERVER_LOGS_FILE_NAME,
 )
@@ -240,8 +239,7 @@ async def save_status_metrics_to_den(
     async def aperiodic_cluster_checks(self):
         """Periodically check the status of the cluster, gather metrics about the cluster's utilization & memory,
         and save it to Den."""
-        # Delay the start of post_status_thread, so we'll finish the cluster startup properly
-        await asyncio.sleep(SCHEDULERS_DELAY)
+
         cluster_config = await self.aget_cluster_config()
         interval_size = cluster_config.get(
             "status_check_interval", DEFAULT_STATUS_CHECK_INTERVAL

diff --git a/tests/test_resources/test_clusters/test_cluster.py b/tests/test_resources/test_clusters/test_cluster.py
@@ -765,7 +765,6 @@ def test_status_scheduler_basic_flow(self, cluster):
         cluster.save()
         # the scheduler start running in a delay of 1 min, so the cluster startup will finish properly.
         # Therefore, the test needs to sleep for a while.
-        time.sleep(60)
         cluster_logs = cluster.run([f"cat {SERVER_LOGFILE_PATH}"], stream_logs=False)[
             0
         ][1]

diff --git a/tests/test_resources/test_clusters/test_on_demand_cluster.py b/tests/test_resources/test_clusters/test_on_demand_cluster.py
@@ -280,7 +280,6 @@ def test_status_cluster_rh_daemon_stopped(self, cluster):
     @pytest.mark.level("minimal")
     def test_logs_surfacing_scheduler_basic_flow(self, cluster):
 
-        time.sleep(120)
         cluster_uri = rh.globals.rns_client.format_rns_address(cluster.rns_address)
         headers = rh.globals.rns_client.request_headers()
         api_server_url = rh.globals.rns_client.api_server_url