tweaks

sky/backends/backend_utils.py

@@ -123,7 +123,7 @@
 # request) and (the instances appearing on the cloud).
 # See https://github.com/skypilot-org/skypilot/issues/4431.
 _LAUNCH_DOUBLE_CHECK_WINDOW = 60
-_LAUNCH_DOUBLE_CHECK_DELAY = 2
+_LAUNCH_DOUBLE_CHECK_DELAY = 1
 
 # Include the fields that will be used for generating tags that distinguishes
 # the cluster in ray, to avoid the stopped cluster being discarded due to
@@ -1803,13 +1803,12 @@ def run_ray_status_to_check_ray_cluster_healthy() -> bool:
             logger.debug(
                 f'Refreshing status ({cluster_name!r}) failed to get IPs.')
         except RuntimeError as e:
-            logger.debug(str(e))
+            logger.debug(common_utils.format_exception(e))
         except Exception as e:  # pylint: disable=broad-except
             # This can be raised by `external_ssh_ports()`, due to the
             # underlying call to kubernetes API.
-            logger.debug(
-                f'Refreshing status ({cluster_name!r}) failed: '
-                f'{common_utils.format_exception(e, use_bracket=True)}')
+            logger.debug(f'Refreshing status ({cluster_name!r}) failed: ',
+                         exc_info=e)
         return False
 
     # Determining if the cluster is healthy (UP):

sky/backends/cloud_vm_ray_backend.py

@@ -2375,15 +2375,17 @@ def get_command_runners(self,
                 zip(ip_list, port_list), **ssh_credentials)
             return runners
         if self.cached_cluster_info is None:
-            # We have `or self.cached_external_ips is None` here, because
+            # We have `and self.cached_external_ips is None` here, because
             # when a cluster's cloud is just upgraded to the new provsioner,
             # although it has the cached_external_ips, the cached_cluster_info
             # can be None. We need to update it here, even when force_cached is
             # set to True.
             # TODO: We can remove `self.cached_external_ips is None` after
             # version 0.8.0.
-            assert not force_cached or self.cached_external_ips is not None, (
-                force_cached, self.cached_external_ips)
+            if force_cached and self.cached_external_ips is None:
+                raise RuntimeError(
+                    'Tried to use cached cluster info, but it\'s missing for '
+                    f'cluster "{self.cluster_name}"')
             self._update_cluster_info()
         assert self.cached_cluster_info is not None, self
         runners = provision_lib.get_command_runners(