You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
"description": "ComponentRequirements represents the requirements for a specific component.",
21181
+
"type": "object",
21182
+
"required": [
21183
+
"replicas"
21184
+
],
21185
+
"properties": {
21186
+
"name": {
21187
+
"description": "Name of this component. It is required when the resource contains multiple components to ensure proper identification, and must also be unique within the same resource.",
21188
+
"type": "string"
21189
+
},
21190
+
"replicaRequirements": {
21191
+
"description": "ReplicaRequirements represents the requirements required by each replica for this component.",
"description": "Components represents the requirements of multiple pod templates of the referencing resource. It is designed to support workloads that consist of multiple pod templates, such as distributed training jobs (e.g., PyTorch, TensorFlow) and big data workloads (e.g., FlinkDeployment), where each workload is composed of more than one pod template. It is also capable of representing single-component workloads, such as Deployment.\n\nNote: This field is intended to replace the legacy ReplicaRequirements and Replicas fields above. It is only populated when the MultiplePodTemplatesScheduling feature gate is enabled.",
"description": "ConflictResolution declares how potential conflict should be handled when a resource that is being propagated already exists in the target cluster.\n\nIt defaults to \"Abort\" which means stop propagating to avoid unexpected overwrites. The \"Overwrite\" might be useful when migrating legacy cluster resources to Karmada, in which case conflict is predictable and can be instructed to Karmada take over the resource by overwriting.",
0 commit comments