RHOAIENG-8098 - ClusterConfiguration should support tolerations

src/codeflare_sdk/common/utils/unit_test_support.py

@@ -22,6 +22,7 @@
 import yaml
 from pathlib import Path
 from kubernetes import client
+from kubernetes.client import V1Toleration
 from unittest.mock import patch
 
 parent = Path(__file__).resolve().parents[4]  # project directory
@@ -427,8 +428,18 @@ def create_cluster_all_config_params(mocker, cluster_name, is_appwrapper) -> Clu
         head_memory_requests=12,
         head_memory_limits=16,
         head_extended_resource_requests={"nvidia.com/gpu": 1, "intel.com/gpu": 2},
+        head_tolerations=[
+            V1Toleration(
+                key="key1", operator="Equal", value="value1", effect="NoSchedule"
+            )
+        ],
         worker_cpu_requests=4,
         worker_cpu_limits=8,
+        worker_tolerations=[
+            V1Toleration(
+                key="key2", operator="Equal", value="value2", effect="NoSchedule"
+            )
+        ],
         num_workers=10,
         worker_memory_requests=12,
         worker_memory_limits=16,

src/codeflare_sdk/ray/cluster/build_ray_cluster.py

@@ -16,7 +16,7 @@
     This sub-module exists primarily to be used internally by the Cluster object
     (in the cluster sub-module) for RayCluster/AppWrapper generation.
 """
-from typing import Union, Tuple, Dict
+from typing import List, Union, Tuple, Dict
 from ...common import _kube_api_error_handling
 from ...common.kubernetes_cluster import get_api_client, config_check
 from kubernetes.client.exceptions import ApiException
@@ -40,6 +40,7 @@
     V1PodTemplateSpec,
     V1PodSpec,
     V1LocalObjectReference,
+    V1Toleration,
 )
 
 import yaml
@@ -139,7 +140,11 @@ def build_ray_cluster(cluster: "codeflare_sdk.ray.cluster.Cluster"):
                     "resources": head_resources,
                 },
                 "template": {
-                    "spec": get_pod_spec(cluster, [get_head_container_spec(cluster)])
+                    "spec": get_pod_spec(
+                        cluster,
+                        [get_head_container_spec(cluster)],
+                        cluster.config.head_tolerations,
+                    )
                 },
             },
             "workerGroupSpecs": [
@@ -154,7 +159,11 @@ def build_ray_cluster(cluster: "codeflare_sdk.ray.cluster.Cluster"):
                         "resources": worker_resources,
                     },
                     "template": V1PodTemplateSpec(
-                        spec=get_pod_spec(cluster, [get_worker_container_spec(cluster)])
+                        spec=get_pod_spec(
+                            cluster,
+                            [get_worker_container_spec(cluster)],
+                            cluster.config.worker_tolerations,
+                        )
                     ),
                 }
             ],
@@ -243,14 +252,21 @@ def update_image(image) -> str:
     return image
 
 
-def get_pod_spec(cluster: "codeflare_sdk.ray.cluster.Cluster", containers):
+def get_pod_spec(
+    cluster: "codeflare_sdk.ray.cluster.Cluster",
+    containers: List,
+    tolerations: List[V1Toleration],
+) -> V1PodSpec:
     """
     The get_pod_spec() function generates a V1PodSpec for the head/worker containers
     """
+
     pod_spec = V1PodSpec(
         containers=containers,
         volumes=generate_custom_storage(cluster.config.volumes, VOLUMES),
+        tolerations=tolerations or None,
     )
+
     if cluster.config.image_pull_secrets != []:
         pod_spec.image_pull_secrets = generate_image_pull_secrets(cluster)
 

src/codeflare_sdk/ray/cluster/config.py

@@ -22,7 +22,7 @@
 import warnings
 from dataclasses import dataclass, field, fields
 from typing import Dict, List, Optional, Union, get_args, get_origin
-from kubernetes.client import V1Volume, V1VolumeMount
+from kubernetes.client import V1Toleration, V1Volume, V1VolumeMount
 
 dir = pathlib.Path(__file__).parent.parent.resolve()
 
@@ -58,6 +58,8 @@ class ClusterConfiguration:
             The number of GPUs to allocate to the head node. (Deprecated, use head_extended_resource_requests)
         head_extended_resource_requests:
             A dictionary of extended resource requests for the head node. ex: {"nvidia.com/gpu": 1}
+        head_tolerations:
+            List of tolerations for head nodes.
         min_cpus:
             The minimum number of CPUs to allocate to each worker.
         max_cpus:
@@ -70,6 +72,8 @@ class ClusterConfiguration:
             The maximum amount of memory to allocate to each worker.
         num_gpus:
             The number of GPUs to allocate to each worker. (Deprecated, use worker_extended_resource_requests)
+        worker_tolerations:
+            List of tolerations for worker nodes.
         appwrapper:
             A boolean indicating whether to use an AppWrapper.
         envs:
@@ -110,6 +114,7 @@ class ClusterConfiguration:
     head_extended_resource_requests: Dict[str, Union[str, int]] = field(
         default_factory=dict
     )
+    head_tolerations: Optional[List[V1Toleration]] = None
     worker_cpu_requests: Union[int, str] = 1
     worker_cpu_limits: Union[int, str] = 1
     min_cpus: Optional[Union[int, str]] = None  # Deprecating
@@ -120,6 +125,7 @@ class ClusterConfiguration:
     min_memory: Optional[Union[int, str]] = None  # Deprecating
     max_memory: Optional[Union[int, str]] = None  # Deprecating
     num_gpus: Optional[int] = None  # Deprecating
+    worker_tolerations: Optional[List[V1Toleration]] = None
     appwrapper: bool = False
     envs: Dict[str, str] = field(default_factory=dict)
     image: str = ""
@@ -272,7 +278,10 @@ def check_type(value, expected_type):
             if origin_type is Union:
                 return any(check_type(value, union_type) for union_type in args)
             if origin_type is list:
-                return all(check_type(elem, args[0]) for elem in value)
+                if value is not None:
+                    return all(check_type(elem, args[0]) for elem in (value or []))
+                else:
+                    return True
             if origin_type is dict:
                 return all(
                     check_type(k, args[0]) and check_type(v, args[1])

tests/test_cluster_yamls/appwrapper/unit-test-all-params.yaml

@@ -99,6 +99,11 @@ spec:
               imagePullSecrets:
               - name: secret1
               - name: secret2
+              tolerations:
+              - effect: NoSchedule
+                key: key1
+                operator: Equal
+                value: value1
               volumes:
               - emptyDir:
                   sizeLimit: 500Gi
@@ -185,6 +190,11 @@ spec:
               imagePullSecrets:
               - name: secret1
               - name: secret2
+              tolerations:
+              - effect: NoSchedule
+                key: key2
+                operator: Equal
+                value: value2
               volumes:
               - emptyDir:
                   sizeLimit: 500Gi

tests/test_cluster_yamls/ray/unit-test-all-params.yaml

@@ -90,6 +90,11 @@ spec:
         imagePullSecrets:
         - name: secret1
         - name: secret2
+        tolerations:
+        - effect: NoSchedule
+          key: key1
+          operator: Equal
+          value: value1
         volumes:
         - emptyDir:
             sizeLimit: 500Gi
@@ -176,6 +181,11 @@ spec:
         imagePullSecrets:
         - name: secret1
         - name: secret2
+        tolerations:
+        - effect: NoSchedule
+          key: key2
+          operator: Equal
+          value: value2
         volumes:
         - emptyDir:
             sizeLimit: 500Gi