etcd集群安装和维护（V2）

yum install -y etcd

如果之前安装过etcd，停掉etcd服务并，清理持久化目录，默认目录在/var/lib/etcd/default.etcd。

停止方法：systemctl stop etcd

数据清理方法: mv /var/lib/etcd/default.etcd /var/lib/etcd/default.etcd.bak

 

1、修改配置文件：vi /etc/etcd/etcd.conf

配置文件如下，绿色部分IP需要修改，以172.28.73.54、172.28.73.40、172.28.16.153为例

172.28.73.54配置

ETCD_NAME="node1"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="http://172.28.73.54:2380"
ETCD_LISTEN_CLIENT_URLS="http://172.28.73.54:2379,http://127.0.0.1:2379"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.28.73.54:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.28.73.54:2379"
ETCD_INITIAL_CLUSTER="node1=http://172.28.73.54:2380,node2=http://172.28.73.40:2380,node3=http://172.28.16.153:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"

172.28.73.40配置

ETCD_NAME="node2"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="http://172.28.73.40:2380"
ETCD_LISTEN_CLIENT_URLS="http://172.28.73.40:2379,http://127.0.0.1:2379"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.28.73.40:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.28.73.40:2379"
ETCD_INITIAL_CLUSTER="node1=http://172.28.73.54:2380,node2=http://172.28.73.40:2380,node3=http://172.28.16.153:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"

172.28.16.53配置

ETCD_NAME="node3"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="http://172.28.16.153:2380"
ETCD_LISTEN_CLIENT_URLS="http://172.28.16.153:2379,http://127.0.0.1:2379"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.28.16.153:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.28.16.153:2379"
ETCD_INITIAL_CLUSTER="node1=http://172.28.73.54:2380,node2=http://172.28.73.40:2380,node3=http://172.28.16.153:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="new"

2、每台服务器修改etcd系统启动文件

vi /usr/lib/systemd/system/etcd.service

[Service]
Type=notify
WorkingDirectory=/var/lib/etcd/
EnvironmentFile=-/etc/etcd/etcd.conf
User=etcd
# set GOMAXPROCS to number of processors
ExecStart=/bin/bash -c "GOMAXPROCS=$(nproc) /usr/bin/etcd 
--name="${ETCD_NAME}" 
--data-dir="${ETCD_DATA_DIR}" 
--listen-peer-urls="${ETCD_LISTEN_PEER_URLS}" 
--listen-client-urls="${ETCD_LISTEN_CLIENT_URLS}" 
--initial-advertise-peer-urls="${ETCD_INITIAL_ADVERTISE_PEER_URLS}" 
--advertise-client-urls="${ETCD_ADVERTISE_CLIENT_URLS}" 
--initial-cluster="${ETCD_INITIAL_CLUSTER}"  
--initial-cluster-token="${ETCD_INITIAL_CLUSTER_TOKEN}" 
--initial-cluster-state="${ETCD_INITIAL_CLUSTER_STATE}""
Restart=on-failure
LimitNOFILE=65536

3、启动

依次启动etcd服务

systemclt start etcdctl

4、验证

查看成员

etcdctl member list

每个节点上执行查看健康状态

etcdctl cluster-health

在任意一个节点etcdctl  set  /test  "hello etcd"

在另一个几点 etcdctl  get  /test 能看到数据则正常

5、添加新节点

如在ip172.28.16.166上添加

配置文件

ETCD_NAME="node4"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_PEER_URLS="http://172.28.16.166:2380"
ETCD_LISTEN_CLIENT_URLS="http://172.28.16.166:2379,http://127.0.0.1:2379"
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://172.28.16.166:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://172.28.16.166:2379"
ETCD_INITIAL_CLUSTER="node1=http://172.28.73.54:2380,node2=http://172.28.73.40:2380,node3=http://172.28.16.153:2380,node4=http://172.28.16.166:2380"
ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_INITIAL_CLUSTER_STATE="existing"　

启动文件修改同上

启动节点systemctl start etcd

在另外任意节点上执行

etcdctl member add node4  http://172.28.16.166:2380 

etcdctl member list 查看节点信息和状态

如果出现添加失败，则清理node4的存储目录并重启node4的etcd服务

修改另外3个节点的 ETCD_INITIAL_CLUSTER ，添加node4信息，重启etcd服务

6、移除节点

先查出节点id      etcdctl member list
删除             etcdctl member remove 节点id

修改另外3个节点的 ETCD_INITIAL_CLUSTER ，去掉node4信息，重启etcd服务　

 7、持久化数据备份和恢复

 etcd v2 和 v3 的数据不能混合存放，默认是v2。以下是v2的备份和恢复方法，v3参考 etcd集群备份和数据恢复

7.1备份

//预置测试数据,非必须
etcdctl set /test test
etcdctl backup --data-dir /var/lib/etcd/default.etcd/ -backup-dir /tmp/etcd_backup
tar -cvf backup.etcd.tar.gz /tmp/etcd_backup

 backup.etcd.tar.gz 为备份的数据

7.2恢复

将backup.etcd.tar.gz copy到要恢复的集群任意一个服务器上，集群默认配置为1中的配置，将第一个恢复的服务器的ETCD_INITIAL_CLUSTER_STATE设置为new，其他两个设置为exsiting

tar -xvf backup.etcd.tar.gz
rm -rf /var/lib/etcd/*
mv tmp/etcd_backup/ /var/lib/etcd/default.etcd/

7.3数据验证和修改节点信息

etcd -data-dir="/var/lib/etcd/default.etcd" --name="node1" --force-new-cluster
开启一个新的窗口
etcdctl get /test

查看节点id 

etcdctl member list

 修改peerURL

curl http://localhost:2379/v2/members/71bf939a1c01 -XPUT -H "Content-Type:application/json" -d '{"peerURLs":["http://172.28.73.54:2380"]}'

etcdctl member list

 第一个窗口ctrl c,关闭当前节点

7.4启动当前节点

chown etcd:etcd /var/lib/etcd -R
systemctl daemon-reload
systemctl start etcd

7.5添加节点（ip为第二个服务器ip）

etcdctl member add node2 http://172.28.73.40:2380
etcdctl member list

不管。

7.6启动第二个节点

更新/etc/etcd/etcd.conf的配置

ETCD_INITIAL_CLUSTER="node1=http://172.28.73.54:2380,node2=http://172.28.73.40:2380"

ETCD_INITIAL_CLUSTER_STATE="existing"

启动

rm -rf /var/lib/etcd/*
systemctl daemon-reload
systemctl start etcd

7.7启动第三个节点

etcdctl member add node3 http://172.28.16.153:2380

同7.6

做好数据验证，如没问题，将第二个节点的ETCD_INITIAL_CLUSTER更新为3个节点的，然后重启。

8、常见错误

无法启动：

方法1、清理data-dir再重启

方法2、查看相关目录权限是否和/usr/lib/systemd/system/etcd.service中的一直，如不一致，则修改，如默认为etcd，经过迁移后相关目录权限变成了root导致无法启动

chown etcd:etcd /var/lib/etcd -R
systemctl daemon-reload
systemctl start etcd