Kubernetes生产实践系列之七：通过etcd备份和恢复Kubernetes集群状态

一、前言

严格来讲，Kubernetes的所有组建都是无状态的，这些组建的状态包括各种后来部署的资源的状态都存储在etcd集群之中，所以通过备份etcd，可以在灾难情况下快速恢复集群和集群上的应用。

转载自https://blog.csdn.net/cloudvtec

二、进行集群备份

2.1 查看集群当前状态

kubectl get pods --all-namespaces -o wide
NAMESPACE     NAME                                       READY     STATUS    RESTARTS   AGE       IP              NODE
default       demo-deployment-7c687dbbfd-jxvnw           1/1       Running   0          17m       10.254.66.8     k8s-node-02
default       demo-deployment-7c687dbbfd-l4pzv           1/1       Running   0          17m       10.254.95.250   k8s-node-01
default       demo-deployment-7c687dbbfd-pjz9q           1/1       Running   0          17m       10.254.95.251   k8s-node-01
default       demo-deployment-7c687dbbfd-xzb6b           1/1       Running   0          17m       10.254.95.249   k8s-node-01
kube-system   calico-kube-controllers-64b4dd5f65-5r6sj   1/1       Running   0          17m       10.0.2.15       k8s-node-01
kube-system   calico-node-bdv8q                          2/2       Running   0          17m       10.0.2.15       k8s-node-01
kube-system   calico-node-v4d25                          2/2       Running   0          17m       10.0.2.15       k8s-node-02
kube-system   coredns-794cc4cddd-jt4jf                   1/1       Running   0          1m        10.254.66.15    k8s-node-02
kube-system   coredns-794cc4cddd-k5n88                   1/1       Running   0          1m        10.254.95.253   k8s-node-01

ETCDCTL_API=3 etcdctl --cert=/opt/etcd/ca/client.pem  --key=/opt/etcd/ca/client-key.pem --cacert=/opt/etcd/ca/ca.pem --endpoints=https://k8s-master-01:2379,https://k8s-master-02:2379,https://k8s-master-03:2379 get / --prefix --keys-only

2.2 进行etcd备份

 

etcd集群一般都有多个节点，只要选择其中一个进行备份就