http://galeracluster.com/documentation-webpages/index.html主要关注点是数据一致性。 事务既可以应用于每个节点，也可以不全部应用。 所以，只要它们配置正确，数据库保持同步。
Galera复制插件不同于传统的MySQL复制，可以解决多个问题，包括多主写入冲突，复制滞后和主从不同步。
在典型的Galera集群实例中，应用程序可以写入集群中的任何节点，然后通过基于认证的复制将事务提交（RBR事件）应用于所有服务器。
使用组通信和事务排序技术，基于认证的复制是同步数据库复制的另一种方法
说明：Galera集群至少需要三个节点的服务器硬件。
node-12:10.71.11.12
node-13:10.71.11.13
node-14:10.71.11.14操作系统:
centos7内核版本:
3.10.0-693.21.1.el7.x86_64说明：为集群中的每个节点执行以下步骤。文档以node-12配置为例
1.编辑/etc/hosts文件，配置节点互相解析
[root@node-12 ~]# cat /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
10.71.11.12 node-12
10.71.11.13 node-13
10.71.11.14 node-142.关闭节点防火墙
[root@node-12 ~]# cat /etc/sysconfig/selinux |grep di
# disabled - No SELinux policy is loaded.
SELINUX=disabled
# minimum - Modification of targeted policy. Only selected processes are protected.3.配置Galera集群yum源
[root@node-12 ~]# cat /etc/yum.repos.d/galera.repo
[galera]
name = Galera
baseurl = http://releases.galeracluster.com/galera-3/centos/7/x86_64
gpgkey = http://releases.galeracluster.com/GPG-KEY-galeracluster.com
gpgcheck = 1
[mysql-wsrep]
name = MySQL-wsrep
baseurl = http://releases.galeracluster.com/mysql-wsrep-5.6/centos/7/x86_64
gpgkey = http://releases.galeracluster.com/GPG-KEY-galeracluster.com
gpgcheck = 14.更新yum源缓存
yum makecache5.安装服务
yum install mysql-wsrep-shared-5.6说明：如果节点上之前安装过mysql，执行上面命令可能会安装包冲突报错
6.以node-12为例，编辑/etc/my.cnf,添加如下配置
[root@node-12 ~]# cat /etc/my.cnf
[mysqld]
datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock
user=mysql
binlog_format=row
bind-address=10.71.11.12
default_storage_engine=innodb
innodb_autoinc_lock_mode=2
innodb_flush_log_at_trx_commit=0
innodb_buffer_pool_size=122M
wsrep_provider=/usr/lib64/galera-3/libgalera_smm.so
wsrep_provider_options="gcache.size=300M; gcache.page_size=300M"
wsrep_sst_method=rsync
wsrep_cluster_name=MyCluster
wsrep_cluster_address="gcomm://10.71.11.12,10.71.11.13,10.71.11.14"
wsrep_node_name=node-12
wsrep_node_address="10.71.11.12"
[mysql_safe]
log-error=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid
说明：其他另外node-13和node-14依据1-6步配置，需要修改/etc/my.cnf与每个节点对应
Galera Cluster的内存要求很难准确预测。 它使用的特定内存量可能会有很大差异，具体取决于给定节点接收的负载。 如果Galera Cluster尝试使用比节点更多的内存，则mysqld实例会崩溃。
保护节点系统崩溃的方法是确保服务有足够的swap 空间分区或者交换文件
检查系统可用交换空间
如果系统swap分区不满足环境需求，可以进行如下操作配置swap分区
说明：集群的中每个节点都要进行相关swap分区设置
1.根据需求在根目录下创建一个2G大小swap空文件
[root@node-11 ~]#fallocate -l 2048M /swapfile或者执行
[root@node-11 ~]#dd if=/dev/zero of=/swapfile bs=1M count=2048设置swap交换分区大小
2.设置swap文件权限
[root@node-11 ~]#chown 600 /swapfile
ll /swapfile
-rw-r--r-- 1 600 root 2147483648 Apr 17 09:13 /swapfile3.格式化swapfile
[root@node-11 ~]# mkswap /swapfile
Setting up swapspace version 1, size = 2097148 KiB
no label, UUID=6341a320-1804-4fe1-8c96-7c22fe270eeb4 .激活swapfile
[root@node-11 ~]# swapon /swapfile
swapon: /swapfile: insecure permissions 0644, 0600 suggested.
swapon: /swapfile: insecure file owner 600, 0 (root) suggested./swapfile none swap defaults 0 0
6.验证swap交换分区配置文件是否成功
[root@node-11 ~]# swapon --summary
Filename Type Size Used Priority
/swapfile file 2097148 0 -1说明：由于现在部署Galera集群的节点环境部署过其他环境，这里说下删除swap分区的步骤
[root@node-11 ~]# swapon --summary
Filename Type Size Used Priority
/openstack/swap.img file 4194300 0 0
/swapfile file 2097148 0 -1
[root@node-11 ~]# swapoff /openstack/swap.img ##首先停止swap分区
[root@node-11 ~]# rm -rf /openstack/swap.img ##删除swap分区文件
[root@node-11 ~]# swapon --summary
Filename Type Size Used Priority
/swapfile file 2097148 0 -1
说明：完成安装和配置Galera Cluster后，需要使用--wsrep-new-cluster选择一个节点上启动mysqld。 这将会初始化集群的核心组件。 在此之后启动的每个节点都将连接到该组件并开始复制
1.在node-12上使用参数--wsrep-new-cluster启动数据库服务
[root@node-12 mysql]# service mysql start --wsrep-new-cluster
数据库初始化报错解决办法，删除/var/lib/mysq/目录下的logfile
注意警告：只有在初始化主要组件时才使用--wsrep-new-cluster参数。 如果希望节点连接到现有群集时，不要执行这条命令。
2.在node-12节点启动数据库服务后，登录mysql数据库库，通过wsrep_cluster_size参数来判断启动是否成功。
SHOW STATUS LIKE ‘wsrep_cluster_size‘;
说明：参数wsrep_cluster_size数值代表目前galera集群当前节点在线数
3.向galera集群添加节点，分别在node-12和node-13上执行下面命令启动数据库服务
[root@node-13 mysql]# service mysql start
Starting MySQL.Logging to ‘/var/lib/mysql/node-13.err‘.
....... SUCCESS!
[root@node-14 yum.repos.d]# service mysql start
Starting MySQL.Logging to ‘/var/lib/mysql/node-14.err‘.
...... SUCCESS!
然后node-13或者node-14上登录mysql数据库，查看wsrep_cluster_size的数值
[root@node-14 yum.repos.d]# mysql -uroot -p
Enter password:
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 4
Server version: 5.6.39 MySQL Community Server (GPL), wsrep_25.22
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> SHOW STATUS LIKE ‘wsrep_cluster_size‘;
到此,galera集群搭建成功
说明：要测试Galera群集是否正常工作
1.在数据库客户端上，验证所有节点是否已相互连接
[root@node-12 ~]# mysql -uroot
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 4
Server version: 5.6.39 MySQL Community Server (GPL), wsrep_25.22
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> show status like ‘wsrep_%‘;
参数说明：
wsrep_local_state_comment参数中Synced表示该节点已经连接到集群并能正常运行
wsrep_cluster_size的数值3表示galera集群中有3个节点都在线
wsrep_ready中on表示此节点连接到集群并能够处理事务。
2.在node-12上创建一个表并插入数据
[root@node-12 mysql]# mysql
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 5
Server version: 5.6.39 MySQL Community Server (GPL), wsrep_25.22
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> CREATE DATABASE galeratest;
Query OK, 1 row affected (0.01 sec)
mysql> USE galeratest;
Database changed
mysql> CREATE TABLE test_table (id INT PRIMARY KEY AUTO_INCREMENT,msg TEXT ) ENGINE=InnoDB;
Query OK, 0 rows affected (0.00 sec)
mysql> INSERT INTO test_table (msg) VALUES ("Hello my dear cluster.");
Query OK, 1 row affected (0.00 sec)
mysql> INSERT INTO test_table (msg) VALUES ("Hello, again, cluster dear.");
Query OK, 1 row affected (0.00 sec)
[root@node-13 mysql]# mysql
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 4
Server version: 5.6.39 MySQL Community Server (GPL), wsrep_25.22Copyright (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> show databases;
+--------------------+
| Database |
+--------------------+
| information_schema |
| galeratest |
| mysql |
| performance_schema |
| test |
+--------------------+
5 rows in set (0.00 sec)
mysql> use galeratest;
Reading table information for completion of table and column names
You can turn off this feature to get a quicker startup with -A
Database changed
mysql> SELECT * FROM test_table;
+----+-----------------------------+
| id | msg |
+----+-----------------------------+
| 1 | Hello my dear cluster. |
| 4 | Hello, again, cluster dear. |
+----+-----------------------------+
2 rows in set (0.00 sec)
#### 三节点Galera集群故障模拟
1.模拟单个mysqld进程的崩溃，在node-14上kill掉mysqld进程[root@node-14 mysql]# service mysql status
SUCCESS! MySQL running (72075)
[root@node-14 mysql]# killall -9 mysqld
-bash: killall: command not found
安装killall命令[root@node-14 mysql]# yum install psmisc -y
[root@node-14 mysql]# killall -9 mysqld
[root@node-14mysql]# /usr/bin/mysqld_safe: line 183: 72075 Killed nohup /usr/sbin/mysqld --basedir=/usr --datadir=/var/lib/mysql --plugin-dir=/usr/lib64/mysql/plugin --user=mysql --log-error=node-14.err --pid-file=/var/lib/mysql/node-14.pid --socket=/var/lib/mysql/mysql.sock --wsrep_start_position=00000000-0000-0000-0000-000000000000:-1 < /dev/null > /dev/null 2>&1
此时到node-12上查看集群状态，显示有2个在线节点
把故障节点加入galera 集群
原文地址：http://blog.51cto.com/liuleis/2107146
