skypilot-nightly 1.0.0.dev20241123__tar.gz → 1.0.0.dev20241124__tar.gz

{skypilot_nightly-1.0.0.dev20241123/skypilot_nightly.egg-info → skypilot_nightly-1.0.0.dev20241124}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: skypilot-nightly
-Version: 1.0.0.dev20241123
+Version: 1.0.0.dev20241124
 Summary: SkyPilot: An intercloud broker for the clouds
 Author: SkyPilot Team
 License: Apache 2.0

{skypilot_nightly-1.0.0.dev20241123 → skypilot_nightly-1.0.0.dev20241124}/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 = '0b20d568ee1af454bfec3e50ff62d239f976e52d'
+_SKYPILOT_COMMIT_SHA = '44625e0e62b70e052072851129507a06ff60636d'
 
 
 def _get_git_commit():
@@ -35,7 +35,7 @@ def _get_git_commit():
 
 
 __commit__ = _get_git_commit()
-__version__ = '1.0.0.dev20241123'
+__version__ = '1.0.0.dev20241124'
 __root_dir__ = os.path.dirname(os.path.abspath(__file__))
 
 

{skypilot_nightly-1.0.0.dev20241123 → skypilot_nightly-1.0.0.dev20241124}/sky/clouds/service_catalog/kubernetes_catalog.py

@@ -65,9 +65,14 @@ def list_accelerators(
     # TODO(romilb): We should consider putting a lru_cache() with TTL to
     #   avoid multiple calls to kubernetes API in a short period of time (e.g.,
     #   from the optimizer).
-    return list_accelerators_realtime(gpus_only, name_filter, region_filter,
-                                      quantity_filter, case_sensitive,
-                                      all_regions, require_price)[0]
+    return _list_accelerators(gpus_only,
+                              name_filter,
+                              region_filter,
+                              quantity_filter,
+                              case_sensitive,
+                              all_regions,
+                              require_price,
+                              realtime=False)[0]
 
 
 def list_accelerators_realtime(
@@ -78,10 +83,36 @@ def list_accelerators_realtime(
     case_sensitive: bool = True,
     all_regions: bool = False,
     require_price: bool = True
+) -> Tuple[Dict[str, List[common.InstanceTypeInfo]], Dict[str, int], Dict[str,
+                                                                          int]]:
+    return _list_accelerators(gpus_only,
+                              name_filter,
+                              region_filter,
+                              quantity_filter,
+                              case_sensitive,
+                              all_regions,
+                              require_price,
+                              realtime=True)
+
+
+def _list_accelerators(
+    gpus_only: bool,
+    name_filter: Optional[str],
+    region_filter: Optional[str],
+    quantity_filter: Optional[int],
+    case_sensitive: bool = True,
+    all_regions: bool = False,
+    require_price: bool = True,
+    realtime: bool = False
 ) -> Tuple[Dict[str, List[common.InstanceTypeInfo]], Dict[str, int], Dict[str,
                                                                           int]]:
     """List accelerators in the Kubernetes cluster.
 
+    If realtime is True, the function will query the cluster to fetch real-time
+    GPU usage, which is returned in total_accelerators_available. Note that
+    this may require an expensive list_pod_for_all_namespaces call, which
+    requires cluster-wide pod read permissions.
+
     If the user does not have sufficient permissions to list pods in all
     namespaces, the function will return free GPUs as -1.
     """
@@ -115,18 +146,20 @@ def list_accelerators_realtime(
     accelerators_qtys: Set[Tuple[str, int]] = set()
     keys = lf.get_label_keys()
     nodes = kubernetes_utils.get_kubernetes_nodes(context)
-    # Get the pods to get the real-time GPU usage
-    try:
-        pods = kubernetes_utils.get_all_pods_in_kubernetes_cluster(context)
-    except kubernetes.api_exception() as e:
-        if e.status == 403:
-            logger.warning('Failed to get pods in the Kubernetes cluster '
-                           '(forbidden). Please check if your account has '
-                           'necessary permissions to list pods. Realtime GPU '
-                           'availability information may be incorrect.')
-            pods = None
-        else:
-            raise
+    pods = None
+    if realtime:
+        # Get the pods to get the real-time GPU usage
+        try:
+            pods = kubernetes_utils.get_all_pods_in_kubernetes_cluster(context)
+        except kubernetes.api_exception() as e:
+            if e.status == 403:
+                logger.warning(
+                    'Failed to get pods in the Kubernetes cluster '
+                    '(forbidden). Please check if your account has '
+                    'necessary permissions to list pods. Realtime GPU '
+                    'availability information may be incorrect.')
+            else:
+                raise
     # Total number of GPUs in the cluster
     total_accelerators_capacity: Dict[str, int] = {}
     # Total number of GPUs currently available in the cluster

{skypilot_nightly-1.0.0.dev20241123 → skypilot_nightly-1.0.0.dev20241124}/sky/templates/kubernetes-ray.yml.j2

@@ -327,6 +327,13 @@ available_node_types:
           command: ["/bin/bash", "-c", "--"]
           args: 
             - |
+              # For backwards compatibility, we put a marker file in the pod
+              # to indicate that the pod is running with the changes introduced 
+              # in project nimbus: https://github.com/skypilot-org/skypilot/pull/4393
+              # TODO: Remove this marker file and it's usage in setup_commands
+              # after v0.10.0 release.
+              touch /tmp/skypilot_is_nimbus
+              
               # Helper function to conditionally use sudo
               # TODO(zhwu): consolidate the two prefix_cmd and sudo replacements
               prefix_cmd() { if [ $(id -u) -ne 0 ]; then echo "sudo"; else echo ""; fi; }
@@ -575,9 +582,12 @@ setup_commands:
     STEPS=("apt-ssh-setup" "runtime-setup" "env-setup")
     start_epoch=$(date +%s);
     echo "=== Logs for asynchronous ray and skypilot installation ===";
-    [ -f /tmp/ray_skypilot_installation_complete ] && cat /tmp/${STEPS[1]}.log || 
-    { tail -f -n +1 /tmp/${STEPS[1]}.log & TAIL_PID=$!; echo "Tail PID: $TAIL_PID"; until [ -f /tmp/ray_skypilot_installation_complete ]; do sleep 0.5; done; kill $TAIL_PID || true; };
-    [ -f /tmp/${STEPS[1]}.failed ] && { echo "Error: ${STEPS[1]} failed. Exiting."; exit 1; } || true;
+    if [ -f /tmp/skypilot_is_nimbus ]; then
+      echo "=== Logs for asynchronous ray and skypilot installation ===";
+      [ -f /tmp/ray_skypilot_installation_complete ] && cat /tmp/${STEPS[1]}.log || 
+      { tail -f -n +1 /tmp/${STEPS[1]}.log & TAIL_PID=$!; echo "Tail PID: $TAIL_PID"; until [ -f /tmp/ray_skypilot_installation_complete ]; do sleep 0.5; done; kill $TAIL_PID || true; };
+      [ -f /tmp/${STEPS[1]}.failed ] && { echo "Error: ${STEPS[1]} failed. Exiting."; exit 1; } || true;
+    fi
     end_epoch=$(date +%s);
     echo "=== Ray and skypilot dependencies installation completed in $(($end_epoch - $start_epoch)) secs ===";
     start_epoch=$(date +%s);

{skypilot_nightly-1.0.0.dev20241123 → skypilot_nightly-1.0.0.dev20241124/skypilot_nightly.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: skypilot-nightly
-Version: 1.0.0.dev20241123
+Version: 1.0.0.dev20241124
 Summary: SkyPilot: An intercloud broker for the clouds
 Author: SkyPilot Team
 License: Apache 2.0