skypilot-nightly 1.0.0.dev20240828__tar.gz → 1.0.0.dev20240830__tar.gz

{skypilot_nightly-1.0.0.dev20240828/skypilot_nightly.egg-info → skypilot_nightly-1.0.0.dev20240830}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: skypilot-nightly
-Version: 1.0.0.dev20240828
+Version: 1.0.0.dev20240830
 Summary: SkyPilot: An intercloud broker for the clouds
 Author: SkyPilot Team
 License: Apache 2.0

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/__init__.py

@@ -5,7 +5,7 @@ from typing import Optional
 import urllib.request
 
 # Replaced with the current commit when building the wheels.
-_SKYPILOT_COMMIT_SHA = 'c7ee6a22c2fbb912bbe8f83b2931062e6b4e4490'
+_SKYPILOT_COMMIT_SHA = '50f68d2093cbe9dc7da6d53c3c17c45e0b97b84c'
 
 
 def _get_git_commit():
@@ -35,7 +35,7 @@ def _get_git_commit():
 
 
 __commit__ = _get_git_commit()
-__version__ = '1.0.0.dev20240828'
+__version__ = '1.0.0.dev20240830'
 __root_dir__ = os.path.dirname(os.path.abspath(__file__))
 
 

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/backends/backend_utils.py

@@ -882,6 +882,11 @@ def write_cluster_config(
         f'open(os.path.expanduser("{constants.SKY_REMOTE_RAY_PORT_FILE}"), "w", encoding="utf-8"))\''
     )
 
+    # We disable conda auto-activation if the user has specified a docker image
+    # to use, which is likely to already have a conda environment activated.
+    conda_auto_activate = ('true' if to_provision.extract_docker_image() is None
+                           else 'false')
+
     # Use a tmp file path to avoid incomplete YAML file being re-used in the
     # future.
     tmp_yaml_path = yaml_path + '.tmp'
@@ -916,10 +921,11 @@ def write_cluster_config(
                 'specific_reservations': specific_reservations,
 
                 # Conda setup
-                'conda_installation_commands':
-                    constants.CONDA_INSTALLATION_COMMANDS,
                 # We should not use `.format`, as it contains '{}' as the bash
                 # syntax.
+                'conda_installation_commands':
+                    constants.CONDA_INSTALLATION_COMMANDS.replace(
+                        '{conda_auto_activate}', conda_auto_activate),
                 'ray_skypilot_installation_commands':
                     (constants.RAY_SKYPILOT_INSTALLATION_COMMANDS.replace(
                         '{sky_wheel_hash}',

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/backends/cloud_vm_ray_backend.py

@@ -82,7 +82,7 @@ _NODES_LAUNCHING_PROGRESS_TIMEOUT = {
     clouds.AWS: 90,
     clouds.Azure: 90,
     clouds.GCP: 240,
-    clouds.Lambda: 150,
+    clouds.Lambda: 300,
     clouds.IBM: 160,
     clouds.OCI: 300,
     clouds.Paperspace: 600,
@@ -1933,7 +1933,7 @@ class RetryingVmProvisioner(object):
         while True:
             if (isinstance(to_provision.cloud, clouds.Azure) and
                     to_provision.accelerators is not None and
-                    'A10' in to_provision.accelerators):
+                    'A10' in to_provision.accelerators and prev_handle is None):
                 logger.warning(f'{style.BRIGHT}{fore.YELLOW}Trying to launch '
                                'an A10 cluster on Azure. This may take ~20 '
                                'minutes due to driver installation.'

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/clouds/service_catalog/common.py

@@ -58,7 +58,9 @@ class InstanceTypeInfo(NamedTuple):
 
 
 def get_catalog_path(filename: str) -> str:
-    return os.path.join(_ABSOLUTE_VERSIONED_CATALOG_DIR, filename)
+    catalog_path = os.path.join(_ABSOLUTE_VERSIONED_CATALOG_DIR, filename)
+    os.makedirs(os.path.dirname(catalog_path), exist_ok=True)
+    return catalog_path
 
 
 def is_catalog_modified(filename: str) -> bool:
@@ -225,7 +227,7 @@ def read_catalog(filename: str,
                         with open(meta_path + '.md5', 'w',
                                   encoding='utf-8') as f:
                             f.write(hashlib.md5(r.text.encode()).hexdigest())
-                logger.info(f'Updated {cloud} catalog.')
+                logger.debug(f'Updated {cloud} catalog {filename}.')
 
     return LazyDataFrame(catalog_path, update_func=_update_catalog)
 

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/provision/kubernetes/utils.py

@@ -110,8 +110,9 @@ def get_gke_accelerator_name(accelerator: str) -> str:
     if accelerator == 'H100':
         # H100 is named as H100-80GB in GKE.
         accelerator = 'H100-80GB'
-    if accelerator in ('A100-80GB', 'L4', 'H100-80GB'):
-        # A100-80GB, L4 and H100-80GB have a different name pattern.
+    if accelerator in ('A100-80GB', 'L4', 'H100-80GB', 'H100-MEGA-80GB'):
+        # A100-80GB, L4, H100-80GB and H100-MEGA-80GB
+        # have a different name pattern.
         return 'nvidia-{}'.format(accelerator.lower())
     else:
         return 'nvidia-tesla-{}'.format(accelerator.lower())
@@ -194,13 +195,10 @@ class GKELabelFormatter(GPULabelFormatter):
             return value.replace('nvidia-tesla-', '').upper()
         elif value.startswith('nvidia-'):
             acc = value.replace('nvidia-', '').upper()
-            if acc in ['H100-80GB', 'H100-MEGA-80GB']:
-                # H100 is named H100-80GB or H100-MEGA-80GB in GKE,
-                # where the latter has improved bandwidth.
-                # See a3-mega instances on GCP.
-                # TODO: we do not distinguish the two GPUs for simplicity,
-                # but we can evaluate whether we should distinguish
-                # them based on users' requests.
+            if acc == 'H100-80GB':
+                # H100 can be either H100-80GB or H100-MEGA-80GB in GKE
+                # we map H100 ---> H100-80GB and keep H100-MEGA-80GB
+                # to distinguish between a3-high and a3-mega instances
                 return 'H100'
             return acc
         else:

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/provision/runpod/utils.py

@@ -77,7 +77,11 @@ def list_instances() -> Dict[str, Dict[str, Any]]:
         info['name'] = instance['name']
         info['port2endpoint'] = {}
 
-        if instance['desiredStatus'] == 'RUNNING' and instance.get('runtime'):
+        # Sometimes when the cluster is in the process of being created,
+        # the `port` field in the runtime is None and we need to check for it.
+        if (instance['desiredStatus'] == 'RUNNING' and
+                instance.get('runtime') and
+                instance.get('runtime').get('ports')):
             for port in instance['runtime']['ports']:
                 if port['isIpPublic']:
                     if port['privatePort'] == 22:

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830}/sky/skylet/constants.py

@@ -135,8 +135,9 @@ CONDA_INSTALLATION_COMMANDS = (
     # true.
     '{ bash Miniconda3-Linux-x86_64.sh -b; '
     'eval "$(~/miniconda3/bin/conda shell.bash hook)" && conda init && '
-    'conda config --set auto_activate_base true && '
-    f'conda activate base; }}; }}; '
+    # Caller should replace {conda_auto_activate} with either true or false.
+    'conda config --set auto_activate_base {conda_auto_activate} && '
+    'conda activate base; }; }; '
     'grep "# >>> conda initialize >>>" ~/.bashrc || '
     '{ conda init && source ~/.bashrc; };'
     # If Python version is larger then equal to 3.12, create a new conda env
@@ -145,7 +146,7 @@ CONDA_INSTALLATION_COMMANDS = (
     # costly to create a new conda env, and venv should be a lightweight and
     # faster alternative when the python version satisfies the requirement.
     '[[ $(python3 --version | cut -d " " -f 2 | cut -d "." -f 2) -ge 12 ]] && '
-    f'echo "Creating conda env with Python 3.10" && '
+    'echo "Creating conda env with Python 3.10" && '
     f'conda create -y -n {SKY_REMOTE_PYTHON_ENV_NAME} python=3.10 && '
     f'conda activate {SKY_REMOTE_PYTHON_ENV_NAME};'
     # Create a separate conda environment for SkyPilot dependencies.

{skypilot_nightly-1.0.0.dev20240828 → skypilot_nightly-1.0.0.dev20240830/skypilot_nightly.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: skypilot-nightly
-Version: 1.0.0.dev20240828
+Version: 1.0.0.dev20240830
 Summary: SkyPilot: An intercloud broker for the clouds
 Author: SkyPilot Team
 License: Apache 2.0