软件包下载地址：http://galeracluster.com/downloads/
安装前需卸载原有mysql，删除相关文件夹；卸载Postfix,这个可能跟MySQL配置有冲突: yum remove postfix –y
先检测一下有没有：yum list installed | grep mysql 卸载： yum -y remove mysql-libs.x86_64
卸载系统自带的mariadb
[root@node0 ~]# rpm -qa| grep mariadb
mariadb-libs-5.5.41-2.el7_0.x86_64
[root@node0 ~]# rpm -e --nodeps mariadb-libs-5.5.41-2.el7_0.x86_64
建议先使用阿里云 yum 源安装基础依赖
yum -y install gcc gcc-c++ openssl openssl-devel lsof socat perl boost-devel rsync jemalloc libaio libaio-devel
1、所需安装包如下：
[root@node0 ~]# ll
-rw-r--r-- 1 root root 12867764 12月 21 10:41 galera-3-25.3.25-2.el7.x86_64.rpm
-rw-r--r-- 1 root root 44852 12月 21 10:41 mysql-wsrep-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 25405632 12月 21 10:41 mysql-wsrep-client-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 283836 12月 21 10:41 mysql-wsrep-common-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 3908032 12月 21 10:41 mysql-wsrep-devel-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 2278988 12月 21 10:41 mysql-wsrep-libs-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 2148744 12月 21 10:41 mysql-wsrep-libs-compat-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 143030144 12月 21 10:41 mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm
-rw-r--r-- 1 root root 26199024 12月 21 10:41 mysql-wsrep-test-5.7-5.7.24-25.16.el7.x86_64.rpm
开始安装（安装过程中，若出现提示需要安装的依赖，就先把所需依赖安装了，然后继续安装软件）
[root@node0 ~]# rpm -ivh mysql-wsrep-common-5.7-5.7.24-25.16.el7.x86_64.rpm
[root@node0 ~]# rpm -ivh mysql-wsrep-libs-5.7-5.7.24-25.16.el7.x86_64.rpm
[root@node0 ~]# rpm -ivh mysql-wsrep-client-5.7-5.7.24-25.16.el7.x86_64.rpm
安装到mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm时发需要依赖，用yum -y install lsof && yum -y install socat && yum -y install perl && yum -y install net-tools安装下依赖，然后继续安装软件包
[root@node0 ~]# rpm -ivh mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm
警告：mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm: 头V4 RSA/SHA512 Signature, 密钥 ID bc19ddba: NOKEY
错误：依赖检测失败：
lsof 被 mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64 需要
net-tools 被 mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64 需要
socat 被 mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64 需要
[root@node0 ~]# yum -y install lsof && yum -y install socat && yum -y install perl && yum -y install net-tools
依赖被安装以后，安装mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm成功了；继续安装···
[root@node0 ~]# rpm -ivh mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm
警告：mysql-wsrep-server-5.7-5.7.24-25.16.el7.x86_64.rpm: 头V4 RSA/SHA512 Signature, 密钥 ID bc19ddba: NOKEY
准备中... ################################# [100%]
正在升级/安装...
1:mysql-wsrep-server-5.7-5.7.24-25.################################# [100%]
[root@node0 ~]# rpm -ivh mysql-wsrep-libs-compat-5.7-5.7.24-25.16.el7.x86_64.rpm
[root@node0 ~]# rpm -ivh galera-3-25.3.25-2.el7.x86_64.rpm
好，安装完毕！接下来我们就开始编写配置文件了：vi /etc/my.cnf 将里面的内容全部注释掉，然后输入： !includedir /etc/my.cnf.d/ 。接下来创建wsrep.cnf文件： vi /etc/my.cnf.d/wsrep.cnf 然后输入根据自己的需求输入配置信息：
# For advice on how to change settings please see
# http://dev.mysql.com/doc/refman/5.7/en/server-configuration-defaults.html
# *** DO NOT EDIT THIS FILE. It's a template which will be copied to the
# *** default location during install, and will be replaced if you
# *** upgrade to a newer version of MySQL.
[mysqld]
# Remove leading # and set to the amount of RAM for the most important data
# cache in MySQL. Start at 70% of total RAM for dedicated server, else 10%.
# innodb_buffer_pool_size = 128M
# Remove leading # to turn on a very important data integrity option: logging
# changes to the binary log between backups.
# log_bin
# These are commonly set, remove the # and set as required.
# basedir = .....
# datadir = .....
# port = .....
# server_id = .....
# socket = .....
log_timestamps=SYSTEM #这个是我们自己加的，防止日志时间和系统时间不一样
port=3306
server_id=11 #MySQL服务器的ID，必须是唯一的，集群各个节点也不同
explicit_defaults_for_timestamp=true
basedir=/usr/share/mysql/
datadir=/usr/local/mysql/data
socket=/usr/local/mysql/data/mysql.sock
pid_file=/run/mysqld/mysqld.pid
log_error=/var/log/mysql.error
wsrep_cluster_name='cs_cluster' #galera集群的名字，必须是统一的
wsrep-provider=/usr/lib64/galera-3/libgalera_smm.so #wsrep提供者,必须配置(.so文件的路径在哪，就配置成哪,一般安装好后都是在这个目录下)
wsrep_node_name = node1 #wsrep节点的ID，必须是唯一的，集群各个节点也不同
wsrep_cluster_address=gcomm:// #192.168.190.132,192.168.190.133,192.168.190.134（第一次启动前把地址都注释掉，之后再启动节点时再次添加） #集群中的其他节点地址，可以使用主机名或IP
wsrep_node_address='192.168.190.132' #本机节点地址，可以使用主机名或IP
wsrep_provider_options ="gmcast.listen_addr=tcp://192.168.190.132:4567" #指定wsrep启动端口号,4567为默认值
wsrep_sst_donor='node1,node2,node3' #一个逗号分割的节点串作为状态转移源，比如wsrep_sst_donor=node1,node2,node3 如果node1可用，用node2,如果node2不可用，用node3，最后的逗号表明让提供商自己选择一个最优的。
wsrep_sst_method=rsync #集群同步方式，我的系统没有可以用yum安装一下这个远程连接 yum -y install rsync
wsrep_sst_auth=test:123456 #集群同步的用户名密码
slow_query_log=on
[client]
default-character-set=utf8
socket=/usr/local/mysql/data/mysql.sock
[mysql]
default-character-set=utf8
socket=/usr/local/mysql/data/mysql.sock
[mysqldump]
max_allowed_packet = 512M
[mysqld_safe]
malloc-lib=/usr/lib64/libjemalloc.so.1 #这个我的系统里也没有可以用yum安装一下 yum -y install jemalloc 如果获取不到的话，下载一个数据源 rpm -ivh
编辑好配置文件就可以初始化数据库了：
[root@node0 ~]# mysqld --initialize --user=mysql
mysqld: Can't create directory '/usr/local/mysql/data/' (Errcode: 2 - No such file or directory)
2018-12-21T14:11:27.076527+08:00 0 [ERROR] Can't find error-message file '/usr/local/mysql/share/mysql/errmsg.sys'. Check error-message file location and 'lc-messages-dir' configuration directive.
2018-12-21T14:11:27.082187+08:00 0 [ERROR] Aborting
出现了两个错误，第一个是无法创建我们制定的文件目录是权限问题mkdir /usr/local/mysql && chown mysql /usr/local/mysql && chgrp mysql /usr/local/mysql 这样就好了，创建一个目录更改一下权限。第二个是在/usr/local/mysql/share/mysql/errmsg.sys目录下找不到文件，这和我们之前配置文件中basedir=/usr/local/mysql有关系，因为basedir指定是数据库的bin，lib之类的文件，我们rpm安装，这些文件默认的路径是在/usr/share/mysql下面，所以basedir这个参数可以不指定也可以，系统会自己找，或者直接指定到/usr/share/mysql/这个目录下面。所以可以将basedir注释掉或者basedir=/usr/ 再次初始化：
[root@node0 ~]# mysqld --initialize --user=mysql
2018-12-21T14:15:16.922554+08:00 0 [ERROR] Could not open file '/var/log/mysql.error' for error logging: Permission denied
2018-12-21T14:15:16.922696+08:00 0 [ERROR] Aborting
还是权限问题，创建一个更改下属组，然后就可以了touch /var/log/mysql.error && chown mysql /var/log/mysql.error && chgrp mysql /var/log/mysql.error。初始化后会有一个随机密码在错误日志中我们找一下：
[root@node0 ~]# touch /var/log/mysql.error && chown mysql /var/log/mysql.error && chgrp mysql /var/log/mysql.error
[root@node0 ~]# mysqld --initialize --user=mysql
[root@node0 ~]# grep 'temporary password' /var/log/mysql.error
2018-12-21T14:17:34.557895+08:00 1 [Note] A temporary password is generated for root@localhost: W#_jrIo<U9tG
然后就启动数据库了，我们先启动第一个节点，并查看数据库启动状态，如果数据库启动失败，看是否防火墙没有关闭等问题，可以事先执行setenforce 0 && systemctl stop firewalld关闭防火墙，再重新输入命令启动一下。
[root@node0 ~]# /usr/sbin/mysqld --wsrep-new-cluster --user=mysql &
[1] 6660
[root@node0 ~]# systemctl status mysqld
● mysqld.service - MySQL Server
Loaded: loaded (/usr/lib/systemd/system/mysqld.service; enabled; vendor preset: disabled)
Active: inactive (dead)
Docs: man:mysqld(8)
http://dev.mysql.com/doc/refman/en/using-systemd.html
[1]+ 退出 1 /usr/sbin/mysqld --wsrep-new-cluster --user=mysql
数据库没有启动成功，将vi /etc/my.cnf.d/wsrep.cnf 第一个启动节点的配置文件中的wsrep_cluster_address=gcomm:// 设置为这样。并且使用数据库启动命令：service mysqld start 再次查看数据库启动状态：
[root@node0 sbin]# systemctl status mysqld
● mysqld.service - MySQL Server
Loaded: loaded (/usr/lib/systemd/system/mysqld.service; enabled; vendor preset: disabled)
Active: activating (start-pre) since 五 2018-12-21 15:14:03 CST; 20s ago
Docs: man:mysqld(8)
http://dev.mysql.com/doc/refman/en/using-systemd.html
Control: 7474 (mysqld_pre_syst)
CGroup: /system.slice/mysqld.service
└─control
├─7474 /bin/bash /usr/bin/mysqld_pre_systemd --pre
└─7491 /usr/sbin/mysqld --datadir=/usr/local/mysql/data --user=mysql --wsrep_recover
12月 21 15:14:03 node0 systemd[1]: mysqld.service holdoff time over, scheduling restart.
12月 21 15:14:03 node0 systemd[1]: Starting MySQL Server...
数据库会出现如下提示，根据失败提示查看数据库状态
[root@localhost ~]# service mysqld start
Redirecting to /bin/systemctl start mysqld.service
Job for mysqld.service failed because the control process exited with error code. See "systemctl status mysqld.service" and "journalctl -xe" for details.
[root@localhost ~]# systemctl status mysqld.service
● mysqld.service - MySQL Server
Loaded: loaded (/usr/lib/systemd/system/mysqld.service; enabled; vendor preset: disabled)
Active: activating (start-pre) since 二 2018-12-25 10:32:22 CST; 13s ago
Docs: man:mysqld(8)
http://dev.mysql.com/doc/refman/en/using-systemd.html
Control: 3808 (mysqld_pre_syst)
CGroup: /system.slice/mysqld.service
└─control
├─3808 /bin/bash /usr/bin/mysqld_pre_systemd --pre
└─3825 /usr/sbin/mysqld --datadir=/usr/local/mysql/data --user=mysql --wsrep_recover
12月 25 10:32:22 localhost.localdomain systemd[1]: mysqld.service holdoff time over, scheduling rest
12月 25 10:32:22 localhost.localdomain systemd[1]: Starting MySQL Server...
[root@localhost ~]# journalctl -xe
# --------------以下是 journalctl -xe 最后几行内容---------------------------------
12月 25 10:32:21 localhost.localdomain systemd[1]: mysqld.service: control process exited, code=exit
12月 25 10:32:21 localhost.localdomain systemd[1]: Failed to start MySQL Server.
-- Subject: Unit mysqld.service has failed
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit mysqld.service has failed.
--
-- The result is failed.
12月 25 10:32:21 localhost.localdomain systemd[1]: Unit mysqld.service entered failed state.
12月 25 10:32:21 localhost.localdomain systemd[1]: mysqld.service failed.
12月 25 10:32:22 localhost.localdomain systemd[1]: mysqld.service holdoff time over, scheduling rest
12月 25 10:32:22 localhost.localdomain systemd[1]: Starting MySQL Server...
-- Subject: Unit mysqld.service has begun start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit mysqld.service has begun starting up.
最后会发现在journalctl -xe记录记录中，数据库实际是启动成功了。然后咱们再次运行systemctl status mysqld，启动成功
[root@localhost ~]# systemctl status mysqld
● mysqld.service - MySQL Server
Loaded: loaded (/usr/lib/systemd/system/mysqld.service; enabled; vendor preset: disabled)
Active: activating (start-pre) since 二 2018-12-25 10:34:04 CST; 1min 1s ago
Docs: man:mysqld(8)
http://dev.mysql.com/doc/refman/en/using-systemd.html
Control: 3848 (mysqld_pre_syst)
CGroup: /system.slice/mysqld.service
└─control
├─3848 /bin/bash /usr/bin/mysqld_pre_systemd --pre
└─3865 /usr/sbin/mysqld --datadir=/usr/local/mysql/data --user=mysql --wsrep_recover...
12月 25 10:34:04 localhost.localdomain systemd[1]: mysqld.service holdoff time over, schedulin...t.
12月 25 10:34:04 localhost.localdomain systemd[1]: Starting MySQL Server...
Hint: Some lines were ellipsized, use -l to show in full.
进入数据库，根据日志密码进入数据库后修改密码，添加同步数据的账户，也就是配置文件中我们设置的，更改密码后，我的节点二和三也变了，好像是同步过去的，不过咱们还需要在创建一下，要不然节点一挂掉后，其他用户无法使用了。再看一下节点状态：
[root@localhost ~]# mysql -uroot -p
Enter password:
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 6
Server version: 5.7.24-log
Copyright (c) 2000, 2018, Oracle and/or its affiliates. All rights reserved.
Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.
Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
mysql> ALTER USER 'root'@'localhost' IDENTIFIED BY '666666';
Query OK, 0 rows affected (0.00 sec)
mysql> GRANT ALL PRIVILEGES ON *.* TO 'test'@'%' IDENTIFIED BY '123456' WITH GRANT OPTION;FLUSH PRIVILEGES;
Query OK, 0 rows affected, 1 warning (0.00 sec)
Query OK, 0 rows affected (0.01 sec)
mysql> show status like "wsrep%";
+------------------------------+--------------------------------------+
| Variable_name | Value |
+------------------------------+--------------------------------------+
| wsrep_local_state_uuid | af4b7ecb-07ec-11e9-9d71-cfb0574baac9 |
| wsrep_protocol_version | 9 |
| wsrep_last_committed | 3 |
| wsrep_replicated | 3 |
| wsrep_replicated_bytes | 712 |
| wsrep_repl_keys | 3 |
| wsrep_repl_keys_bytes | 96 |
| wsrep_repl_data_bytes | 415 |
| wsrep_repl_other_bytes | 0 |
| wsrep_received | 2 |
| wsrep_received_bytes | 144 |
| wsrep_local_commits | 0 |
| wsrep_local_cert_failures | 0 |
| wsrep_local_replays | 0 |
| wsrep_local_send_queue | 0 |
| wsrep_local_send_queue_max | 1 |
| wsrep_local_send_queue_min | 0 |
| wsrep_local_send_queue_avg | 0.000000 |
| wsrep_local_recv_queue | 0 |
| wsrep_local_recv_queue_max | 2 |
| wsrep_local_recv_queue_min | 0 |
| wsrep_local_recv_queue_avg | 0.500000 |
| wsrep_local_cached_downto | 1 |
| wsrep_flow_control_paused_ns | 0 |
| wsrep_flow_control_paused | 0.000000 |
| wsrep_flow_control_sent | 0 |
| wsrep_flow_control_recv | 0 |
| wsrep_cert_deps_distance | 1.000000 |
| wsrep_apply_oooe | 0.000000 |
| wsrep_apply_oool | 0.000000 |
| wsrep_apply_window | 1.000000 |
| wsrep_commit_oooe | 0.000000 |
| wsrep_commit_oool | 0.000000 |
| wsrep_commit_window | 1.000000 |
| wsrep_local_state | 4 |
| wsrep_local_state_comment | Synced |
| wsrep_cert_index_size | 1 |
| wsrep_causal_reads | 0 |
| wsrep_cert_interval | 0.000000 |
| wsrep_open_transactions | 0 |
| wsrep_open_connections | 0 |
| wsrep_incoming_addresses | 192.168.190.132:3306 |
| wsrep_cluster_weight | 1 |
| wsrep_desync_count | 0 |
| wsrep_evs_delayed | |
| wsrep_evs_evict_list | |
| wsrep_evs_repl_latency | 0/0/0/0/0 |
| wsrep_evs_state | OPERATIONAL |
| wsrep_gcomm_uuid | af4a65e4-07ec-11e9-9aac-76a760227320 |
| wsrep_cluster_conf_id | 1 |
| wsrep_cluster_size | 1 |
| wsrep_cluster_state_uuid | af4b7ecb-07ec-11e9-9d71-cfb0574baac9 |
| wsrep_cluster_status | Primary |
| wsrep_connected | ON |
| wsrep_local_bf_aborts | 0 |
| wsrep_local_index | 0 |
| wsrep_provider_name | Galera |
| wsrep_provider_vendor | Codership Oy <[email protected]> |
| wsrep_provider_version | 3.25(rddf9876) |
| wsrep_ready | ON |
+------------------------------+--------------------------------------+
60 rows in set (0.03 sec)
mysql>
最后不要忘了将配置文件/etc/my.cnf.d/wsrep.cnf中节点的设置修改回来（设置成各节点的IP地址）wsrep_cluster_address=gcomm://192.168.190.132,192.168.190.133,192.168.190.134
【剩下两台机群如此配置即可，注意几点：】
1、初次启动不需要使用启动脚本：/usr/bin/mysqld_bootstrap，因为这个脚本中带有一个参数：--wsrep-new-cluster，代表新集群。
2、配置文件 wsrep.cnf 中的参数 wsrep_cluster_address，一开始就要配置好了，全部配置成指向主节点的IP地址，如：wsrep_cluster_address=gcomm://192.168.190.132:4567 （192.168.190.132为主节点IP地址 ）不能为空。
3、配置文件 wsrep.cnf 中的参数 wsrep_node_address主节点为wsrep_node_address=node1，其他的个节点依次为：wsrep_node_address=node2 ， wsrep_node_address=node3 ，wsrep_node_address=node...
4、配置文件 wsrep.cnf 中的参数server_id必须是唯一的，在配置各节点时，避免出现重复值！
5、类似配置文件 wsrep.cnf 中的 wsrep_node_address 这种参数要配置成自己当前及其所在的IP地址信息。
6、正式环境情况下，我们不要把防火墙关闭掉，把需要使用的端口开放给指定 IP 即可，更加安全。
都搭建成功后，可以看到，在任何一个节点上改动数据库（如创建数据库，改动数据表等操作），另外的节点都会有同步变化！