0

我有一个 K8 集群，需要将我的 pod 与 cinder 存储选项相关联。我尝试了两种选择，但都失败了。任何人都可以阐明正在发生的事情吗？

选项1： 在 OpenStack 中手动创建卷并将其集成到我的 yaml 文件中。Kubectl describe on pod 显示以下错误

Error: Volumes:
  jenkins-volume:
    Type:      Cinder (a Persistent Disk resource in OpenStack)
    VolumeID:  09405897-8477-4479-9730-843a80f88302
    FSType:    ext4
    ReadOnly:  false
  default-token-x76pk:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-x76pk
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type     Reason                 Age   From                     Message
  ----     ------                 ----  ----                     -------
  Normal   Scheduled              2m    default-scheduler        Successfully assigned mongo1-6dfcc8fb88-rzh88 to k8slave1
  Normal   SuccessfulMountVolume  2m    kubelet, k8slave1        MountVolume.SetUp succeeded for volume "default-token-x76pk"
  Warning  FailedAttachVolume     1m    attachdetach-controller  AttachVolume.Attach failed for volume "jenkins-volume" : Volume "09405897-8477-4479-9730-843a80f88302" failed to be attached within the alloted time
  Warning  FailedMount            24s   kubelet, k8slave1        Unable to mount volumes for pod "mongo1-6dfcc8fb88-rzh88_db(6cf912cf-c238-11e8-8224-fa163e01527a)": timeout expired waiting for volumes to attach or mount for pod "db"/"mongo1-6dfcc8fb88-rzh88". list of unmounted volumes=[jenkins-volume]. list of unattached volumes=[jenkins-volume default-token-x76pk]

YAML 文件：

apiVersion: apps/v1
kind: Deployment
metadata:
  name: mongo1
 spec:
  replicas: 1 
  selector:
    matchLabels:
      app: mongo1
  template:
    metadata:
      labels:
        app: mongo1
    spec:
     containers:
      - name: mongo1
        image: mongo:3.5
        volumeMounts:
        - name: jenkins-volume
          mountPath: /data/db 
        ports:
         - containerPort: 27017
     volumes:
      - name: jenkins-volume
        cinder:
         volumeID: 09405897-8477-4479-9730-843a80f88302
         fsType: ext4

Option2： 新建一个存储类，新建一个PV，PVC。这会在 OpenStack 中创建一个新卷，它也会出现在我的 K8 集群中。

将声明关联到 yaml 文件：

root@K8Masternew cinder]# kubectl get pvc

NAME            STATUS    VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   AGE
claim1-volume   Bound     pvc-2bb2a16a-c23d-11e8-8224-fa163e01527a   2Gi        RWO            test           29m


Error:
Events:
  Type     Reason            Age               From               Message
  ----     ------            ----              ----               -------
  Warning  FailedScheduling  9s (x6 over 24s)  default-scheduler  0/3 nodes are available: 1 node(s) had taints that the pod didn't tolerate, 2 node(s) had no available volume zone.

YAML 文件：

apiVersion: apps/v1
kind: Deployment
metadata:
  name: mongo1
spec:
  replicas: 1 
  selector:
    matchLabels:
      app: mongo1
  template:
    metadata:
      labels:
        app: mongo1
    spec:
     containers:
      - name: mongo1
        image: mongo:3.5
        volumeMounts:
        - name: jenkins-volume
          mountPath: /data-db 
        ports:
         - containerPort: 27017
     volumes:
      - name: jenkins-volume
        persistentVolumeClaim:
          claimName: claim1-volum

kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: test
provisioner: kubernetes.io/cinder
parameters:
  availability: nova
************

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: claim1-volume
spec:
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 2Gi
  storageClassName: test
4

1 回答 1

0

看起来您的所有或部分节点上可能有一些污点。Taints在 的输出上寻找kubectl describe node <node-name>。所以如果输出是这样的：

Taints:             node-role.kubernetes.io/slave:NoSchedule

您可以在您的 pod 定义中指定如下内容：

tolerations:
- key: "node-role.kubernetes.io/slave"
  operator: "Equal"
  value: "NoSchedule"
  effect: "NoSchedule"
- key: "node-role.kubernetes.io/slave"
  operator: "Equal"
  value: "NoSchedule"
  effect: "NoExecute"

您可以做的另一件事是完全删除所有污点：

kubectl taint nodes <node-name> node-role.kubernetes.io/slave:NoSchedule-
kubectl taint nodes <node-name> node-role.kubernetes.io/slave:NoExecute-
于 2018-09-27T17:14:36.673 回答