DaemonSet

# Fluent Bit 로그 수집 DaemonSet 예제
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: fluent-bit
  namespace: logging
  labels:
    app: fluent-bit
spec:
  selector:
    matchLabels:
      app: fluent-bit
  template:
    metadata:
      labels:
        app: fluent-bit
    spec:
      serviceAccountName: fluent-bit
      containers:
      - name: fluent-bit
        image: fluent/fluent-bit:2.2
        ports:
        - containerPort: 2020
          name: metrics
        volumeMounts:
        - name: varlog
          mountPath: /var/log
          readOnly: true
        - name: containers
          mountPath: /var/lib/docker/containers
          readOnly: true
        - name: config
          mountPath: /fluent-bit/etc/
        resources:
          requests:
            cpu: 50m
            memory: 64Mi
          limits:
            cpu: 200m
            memory: 128Mi
      volumes:
      - name: varlog
        hostPath:
          path: /var/log
      - name: containers
        hostPath:
          path: /var/lib/docker/containers
      - name: config
        configMap:
          name: fluent-bit-config
      tolerations:
      - operator: Exists  # 모든 taint 허용

# DaemonSet 목록 조회
kubectl get daemonsets -n kube-system

# DaemonSet 상세 정보 (각 노드별 Pod 상태 확인)
kubectl describe daemonset fluent-bit -n logging

# DaemonSet이 배포된 노드별 Pod 확인
kubectl get pods -n logging -o wide -l app=fluent-bit

# 노드 수와 DaemonSet Pod 수 비교
kubectl get nodes --no-headers | wc -l
kubectl get pods -n logging -l app=fluent-bit --no-headers | wc -l

# 특정 노드의 DaemonSet Pod 로그 확인
kubectl logs -n logging daemonset/fluent-bit --tail=100

# DaemonSet 롤링 업데이트 상태 확인
kubectl rollout status daemonset/fluent-bit -n logging

# DaemonSet 이미지 업데이트
kubectl set image daemonset/fluent-bit \
  fluent-bit=fluent/fluent-bit:2.3 -n logging

# 롤백
kubectl rollout undo daemonset/fluent-bit -n logging

# DaemonSet 삭제
kubectl delete daemonset fluent-bit -n logging

# GPU 노드 전용 NVIDIA Device Plugin DaemonSet
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: nvidia-device-plugin
  namespace: kube-system
spec:
  selector:
    matchLabels:
      name: nvidia-device-plugin
  updateStrategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 1  # 한 번에 하나씩만 업데이트
  template:
    metadata:
      labels:
        name: nvidia-device-plugin
    spec:
      # GPU 노드에만 배포
      nodeSelector:
        accelerator: nvidia-gpu

      # Control Plane 노드 제외
      tolerations:
      - key: nvidia.com/gpu
        operator: Exists
        effect: NoSchedule

      # 시스템 컴포넌트로 우선순위 설정
      priorityClassName: system-node-critical

      containers:
      - name: nvidia-device-plugin
        image: nvcr.io/nvidia/k8s-device-plugin:v0.14.3
        securityContext:
          privileged: true  # GPU 접근에 필요
        volumeMounts:
        - name: device-plugin
          mountPath: /var/lib/kubelet/device-plugins
        env:
        - name: FAIL_ON_INIT_ERROR
          value: "false"  # 초기화 실패 시에도 계속 실행

      volumes:
      - name: device-plugin
        hostPath:
          path: /var/lib/kubelet/device-plugins

      # Init Container로 사전 조건 확인
      initContainers:
      - name: check-nvidia-driver
        image: nvidia/cuda:12.0-base
        command: ['nvidia-smi']