# services/ai-llm/deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: ollama
  namespace: ai
spec:
  replicas: 1
  revisionHistoryLimit: 2
  selector:
    matchLabels:
      app: ollama
  template:
    metadata:
      labels:
        app: ollama
    spec:
      nodeSelector:
        kubernetes.io/hostname: titan-24
      runtimeClassName: nvidia
      volumes:
        - name: models
          persistentVolumeClaim:
            claimName: ollama-models
      initContainers:
        - name: warm-model
          image: ollama/ollama:latest
          env:
            - name: OLLAMA_HOST
              value: 0.0.0.0
            - name: NVIDIA_VISIBLE_DEVICES
              value: all
            - name: NVIDIA_DRIVER_CAPABILITIES
              value: compute,utility
            - name: OLLAMA_MODELS
              value: /root/.ollama
            - name: OLLAMA_MODEL
              value: qwen2.5-coder:7b-instruct-q4_0
          command:
            - /bin/sh
            - -c
            - |
              set -e
              ollama serve >/tmp/ollama.log 2>&1 &
              sleep 6
              ollama pull "${OLLAMA_MODEL}"
              pkill ollama || true
          volumeMounts:
            - name: models
              mountPath: /root/.ollama
          resources:
            requests:
              cpu: 250m
              memory: 1Gi
              nvidia.com/gpu: 1
            limits:
              nvidia.com/gpu: 1
      containers:
        - name: ollama
          image: ollama/ollama:latest
          imagePullPolicy: IfNotPresent
          ports:
            - name: http
              containerPort: 11434
          env:
            - name: OLLAMA_HOST
              value: 0.0.0.0
            - name: OLLAMA_KEEP_ALIVE
              value: 6h
            - name: OLLAMA_MODELS
              value: /root/.ollama
            - name: NVIDIA_VISIBLE_DEVICES
              value: all
            - name: NVIDIA_DRIVER_CAPABILITIES
              value: compute,utility
          volumeMounts:
            - name: models
              mountPath: /root/.ollama
          resources:
            requests:
              cpu: "2"
              memory: 8Gi
              nvidia.com/gpu: 1
            limits:
              cpu: "4"
              memory: 12Gi
              nvidia.com/gpu: 1