etcd三节点，其中一个坏掉了的恢复办法

一、配置etcdctl环境变量

---

#其中证书实际路径和endpoints,以环境情况为准,查询方式

ps -ef | grep 'etcd-cafile'

ps -ef | grep 'etcd-servers'

export ETCDCTL_API=3

export ETCDCTL_CACERT=/etc/kubernetes/ssl/etcd/ca.crt

export ETCDCTL_CERT=/etc/kubernetes/ssl/etcd/etcdctl-etcd-client.crt

export ETCDCTL_KEY=/etc/kubernetes/ssl/etcd/etcdctl-etcd-client.key

export ETCDCTL_DIAL_TIMEOUT=3s

export ETCDCTL_ENDPOINTS=https://10.12.107.99:2379,https://10.12.107.100:2379,https://10.12.107.101:2379

---

二、确保有etcdctl命令，如果没有去官网下载二进制包安装即可

三、从集群中剔除故障节点(确保故障节点etcd服务已停止 systemctl stop etcd)

#集群成员信息

etcdctl -wtable member list

#集群主从同步情况

etcdctl -wtable endpoint status

etcdctl -wtable endpoint health

摘除一个节点

etcdctl member remove 成员id

四、清空故障节点etcd的数据目录

mv /var/lib/etcd /var/lib/etcd_bak

五、加入节点 (确保etcd启动参数etcd.env中--initial-cluster-state 为existing)

etcdctl member add 成员名称（主机名） --peer-urls=https://10.8.47.13:2380

六、启动etcd服务

systemctl start etcd