hello云胜

技术与生活

0%

发表于

volume的节点亲和性冲突问题解决

image-20231106155341771

describe一下pod看看

1
Warning  FailedScheduling  <unknown>  default-scheduler  0/5 nodes are available: 1 node(s) had taints that the pod didn't tolerate, 1 node(s) were unschedulable, 3 node(s) had volume node affinity conflict.

node上有volume的节点亲和性冲突。

那么去看看看这个pv有什么清河县的配置

1
kubectl -n kubesphere-monitoring-system get pod prometheus-k8s-0 -o yaml

看到pvc是这个

1
2
3
4
volumes:
 - name: prometheus-k8s-db
   persistentVolumeClaim:
     claimName: prometheus-k8s-db-prometheus-k8s-0

进而找到pv，查看pv的信息

1
kubectl get pv pvc-15224634-6105-4716-a4b7-5e2fd6689db8 -oyaml

肯定是有nodeAffinity信息的

1
2
3
4
5
6
7
8
nodeAffinity:
   required:
     nodeSelectorTerms:
     - matchExpressions:
       - key: kubernetes.io/hostname
         operator: In
         values:
         - str-paas-k8s-0-node-2

嗯，设置了只能调度到str-paas-k8s-0-node-2这台机器上，

而之前因为内存过高，将这台机器加了污点，禁止调度