ai-stack/manifests/ollama/values.yaml

36 lines
529 B
YAML

service:
type: ClusterIP
# Enable GPU support
ollama:
gpu:
enabled: true
type: nvidia
number: 1
resources:
requests:
cpu: 2
memory: 4Gi
limits:
cpu: 4
memory: 8Gi
# Schedule on GPU worker node
nodeSelector:
kubernetes.io/hostname: k3s-worker-5-gpu
# Tolerate GPU node taint
tolerations:
- key: gpu
operator: Exists
effect: NoSchedule
- key: nvidia.com/gpu
operator: Exists
effect: NoSchedule
persistentVolume:
enabled: true
size: 50Gi
storageClass: nfs-client