请教下没有启用 HPA， deployment 什么情况下会自动将 replica 设置为 0

V2EX = way to explore

V2EX 是一个关于分享和探索的地方

现在注册

已注册用户请登录

这是一个创建于 962 天前的主题，其中的信息可能已经有所发展或是发生改变。

最近碰到个问题，使用 deployment 运行 nexus3 跑一段时间后 replicas 会被设置为 0 ，也没启用 HPA ，restartPolicy 配置的是 Always ，查了 kube-apiserver 的 audit 日志也不是人为操作的，查看 nexus3 的日志也没有报错。

deployment yaml:

kind: Deployment
apiVersion: apps/v1
metadata:
  name: service-nexus3-deployment
  namespace: service
  annotations:
    deployment.kubernetes.io/revision: '6'
spec:
  replicas: 1
  selector:
    matchLabels:
      app: service-nexus3
      envronment: test
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: service-nexus3
        envronment: test
      annotations:
        kubesphere.io/restartedAt: '2022-02-16T01:11:44.479Z'
    spec:
      volumes:
        - name: service-nexus3-volume
          persistentVolumeClaim:
            claimName: service-nexus3-pvc
        - name: docker-proxy
          configMap:
            name: docker-proxy
            defaultMode: 493
      containers:
        - name: nexus3
          # 用的阿里的镜像仓库，删了仓库名
          image: 'registry.cn-hangzhou.aliyuncs.com/nexus3-latest'
          ports:
            - name: tcp8081
              containerPort: 8081
              protocol: TCP
          resources:
            limits:
              cpu: '4'
              memory: 8Gi
            requests:
              cpu: 500m
              memory: 1Gi
          volumeMounts:
            - name: service-nexus3-volume
              mountPath: /data/server/nexus3/
          terminationMessagePath: /dev/termination-log
          terminationMessagePolicy: File
          imagePullPolicy: Always
        - name: docker-proxy
          # 用的阿里的镜像仓库，删了仓库名
          image: 'registry.cn-hangzhou.aliyuncs.com/nginx-latest'
          ports:
            - name: tcp80
              containerPort: 80
              protocol: TCP
          resources:
            limits:
              cpu: '2'
              memory: 4Gi
            requests:
              cpu: 500m
              memory: 1Gi
          volumeMounts:
            - name: docker-proxy
              mountPath: /usr/local/nginx/conf/vhosts/
          terminationMessagePath: /dev/termination-log
          terminationMessagePolicy: File
          imagePullPolicy: Always
      restartPolicy: Always
      terminationGracePeriodSeconds: 30
      dnsPolicy: ClusterFirst
      nodeSelector:
        disktype: raid1
      securityContext: {}
      imagePullSecrets:
        - name: registrysecret
      schedulerName: default-scheduler
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 0
      maxSurge: 1
  revisionHistoryLimit: 10
  progressDeadlineSeconds: 600

HPA:

# kubectl get hpa -A
No resources found

deployment describe：

...
...
Events:
  Type    Reason             Age                From                   Message
  ----    ------             ----               ----                   -------
  Normal  ScalingReplicaSet  34m (x2 over 38h)  deployment-controller  Scaled down replica set service-nexus3-deployment-57995fcd76 to 0

kube controller 日志：

# kubectl logs kube-controller-manager-k8s-130 -n kube-system|grep nexus
I0509 10:49:11.687356       1 event.go:281] Event(v1.ObjectReference{Kind:"Deployment", Namespace:"service", Name:"service-nexus3-deployment", UID:"e0c4abba-bbe5-4c19-9853-de63ee571124", APIVersion:"apps/v1", ResourceVersion:"126342143", FieldPath:""}): type: 'Normal' reason: 'ScalingReplicaSet' Scaled down replica set service-nexus3-deployment-57995fcd76 to 0
I0509 10:49:11.701642       1 event.go:281] Event(v1.ObjectReference{Kind:"ReplicaSet", Namespace:"service", Name:"service-nexus3-deployment-57995fcd76", UID:"9f96fdf1-1e20-4c83-ad18-1b3640d52493", APIVersion:"apps/v1", ResourceVersion:"126342151", FieldPath:""}): type: 'Normal' reason: 'SuccessfulDelete' Deleted pod: service-nexus3-deployment-57995fcd76-t6bhx

kube-apiserver audit 相关日志: