小知识，大挑战！本文正在参与“程序员必备小知识”创作活动

环境信息

• 3台CentOS7.8(172.168.1.232/234/235)
• 配置epel源

初始化集群

实施

1. 安装Etcd(这里采用Yum安装，如果对版本有要求访问Etcd官网下载对应二进制版本）

3台机器分别执行
[root@pgslave2 ~]# yum install -y etcd
[root@pgslave2 ~]# rpm -qa|grep etcd etcd-3.3.11-2.el7.centos.x86_64

2. 新建配置集群配置文件

[root@pgslave2 ~]# cat /etc/etcd/etcd.conf 
ETCD_DATA_DIR="/var/lib/etcd/"
ETCD_LISTEN_PEER_URLS="http://0.0.0.0:2380"
ETCD_LISTEN_CLIENT_URLS="http://localhost:2379,http://172.168.1.232:2379"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.168.1.232:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.168.1.232:2379"
ETCD_INITIAL_CLUSTER_TOKEN="hive-cluster"
ETCD_ADVERTISE_CLIENT_URLS="http://172.168.1.232:2379"
ETCD_ENABLE_V2="true"
ETCD_NAME="etcd0"
ETCD_INITIAL_CLUSTER="etcd1=http://172.168.1.234:2380,etcd2=http://172.168.1.235:2380,etcd0=http://172.168.1.232:2380"
ETCD_INITIAL_CLUSTER_STATE="new"

[root@pgslave1 ~]# cat /etc/etcd/etcd.conf 
ETCD_DATA_DIR="/var/lib/etcd/"
ETCD_LISTEN_PEER_URLS="http://0.0.0.0:2380"
ETCD_LISTEN_CLIENT_URLS="http://localhost:2379,http://172.168.1.234:2379"
ETCD_NAME="etcd1"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.168.1.234:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.168.1.234:2379"
ETCD_INITIAL_CLUSTER="etcd0=http://172.168.1.232:2380,etcd1=http://172.168.1.234:2380,etcd2=http://172.168.1.235:2380"
ETCD_INITIAL_CLUSTER_TOKEN="hive-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"
ETCD_ADVERTISE_CLIENT_URLS="http://172.168.1.234:2379"
ETCD_ENABLE_V2="true"

[root@pgmaster ~]# cat /etc/etcd/etcd.conf
ETCD_DATA_DIR="/var/lib/etcd/"
ETCD_LISTEN_PEER_URLS="http://0.0.0.0:2380"
ETCD_LISTEN_CLIENT_URLS="http://localhost:2379,http://172.168.1.235:2379"
ETCD_NAME="etcd2"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.168.1.235:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.168.1.235:2379"
ETCD_INITIAL_CLUSTER="etcd0=http://172.168.1.232:2380,etcd1=http://172.168.1.234:2380,etcd2=http://172.168.1.235:2380"
ETCD_INITIAL_CLUSTER_TOKEN="hive-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"
ETCD_ADVERTISE_CLIENT_URLS="http://172.168.1.235:2379"
ETCD_ENABLE_V2="true"

3. 分别启动3台etcd

systemctl start etcd

故障恢复

a. 恢复某个节点(3台只挂掉一台)

1. 获取集群列表

etcdctl member list

2. 删除不健康节点

etcdctl member remove c13845537406e22f

3. 删除故障节点数据目录

rm -rf  /var/lib/etcd/member

4. 重新添加节点至集群

etcdctl member add etcd2  https://192.168.216.242:2380

5. 启动故障节点etcd恢复

systemctl restart etcd