我有一个 3 节点 kube 控制网络设置,其中、etcd
和正在运行。当我检查组件的状态时,它显示 etcd 不健康,但 etcd 报告说它是健康的kube-apiserver
kube-control-manager
kube-scheduler
> kubectl get componentstatus --kubeconfig .kube/admin.kubeconfig
Warning: v1 ComponentStatus is deprecated in v1.19+
NAME STATUS MESSAGE ERROR
etcd-0 Unhealthy HTTP probe failed with statuscode: 404
etcd-2 Unhealthy HTTP probe failed with statuscode: 404
etcd-1 Unhealthy HTTP probe failed with statuscode: 404
scheduler Healthy ok
controller-manager Healthy ok
当我检查 etcd 的状态时,我得到了以下结果
> etcdctl endpoint health --write-out=table
+------------------------+--------+------------+-------+
| ENDPOINT | HEALTH | TOOK | ERROR |
+------------------------+--------+------------+-------+
| http://10.0.1.164:2379 | true | 2.867446ms | |
| http://10.0.1.62:2379 | true | 3.305321ms | |
| http://10.0.1.176:2379 | true | 6.970254ms | |
+------------------------+--------+------------+-------+
我还能够从 etcd 中放入和检索项目,提供给 kube-apiserver 的端点也是正确的。我该如何调试这样的问题?