一.基本資訊
叢集名：pxczxw
節點一：10.16.24.107
節點二：10.16.24.108
節點三：10.16.24.109

cd /usr/local
ln -s /opt/mysql/Percona-XtraDB-Cluster-5.6.21-rel70.1-25.8.938.Linux.x86_64 mysql

建立目錄如下：
mkdir -p /data/mysql/mysql_3306/{data,logs,tmp}

關閉防火牆和selinux:
/etc/selinux/config中：SELINUX=disabled，並重啟OS
service iptables stop ，並chkconfig iptables off

從其它主機將已準備好的my.cnf檔案copy到/etc/下，並修改3376為3306:
sed -i 's/3376/3306/g' my.cnf
sed -i 's/mysql3306/mysql_3306/g' my.cnf

並將my.cnf檔案scp到另兩臺主機：
scp /etc/my.cnf 10.16.24.108:/etc/
scp /etc/my.cnf 10.16.24.109:/etc/

並調整my.cnf的引數：
10.16.24.108：
server-id       =1083306
10.16.24.109：
server-id       =1093306

建立mysql的使用者及組
groupadd mysql
useradd -g mysql -s /bin/bash -d /usr/local/mysql mysql

建立啟動檔案：
cp /usr/local/mysql/support-files/mysql.server /etc/init.d/mysql

chown -R mysql:mysql /data/mysql/mysql_3306

ln -sf /usr/lib64/libssl.so.10 /usr/lib64/libssl.so.6

ln -sf /usr/lib64/libcrypto.so.10 /usr/lib64/libcrypto.so.6

vi /etc/profile
export PATH=/usr/local/mysql/bin:$PATH

source /etc/profile

四.pxc配置和啟動
在/etc/my.cnf檔案中[mysqld]增加如下：
#percona xtradb cluster
10.16.24.107:
default_storage_engine=InnoDB
innodb_locks_unsafe_for_binlog=1
innodb_autoinc_lock_mode=2
wsrep_provider_options="gcache.size=1999M"
wsrep_cluster_name='pcx_zengxw'
wsrep_cluster_address="gcomm://10.16.24.107,10.16.24.108,10.16.24.109"
wsrep_node_address="10.16.24.107"
wsrep_provider=/usr/local/mysql/lib/libgalera_smm.so
#wsrep_sst_method=rsync
wsrep_sst_method=xtrabackup-v2
wsrep_sst_auth=sst:zengxw1

10.16.24.108:
default_storage_engine=InnoDB
innodb_locks_unsafe_for_binlog=1
innodb_autoinc_lock_mode=2
wsrep_provider_options="gcache.size=1999M"
wsrep_cluster_name='pcx_zengxw'
wsrep_cluster_address="gcomm://10.16.24.107,10.16.24.108,10.16.24.109"
wsrep_node_address="10.16.24.108"
wsrep_provider=/usr/local/mysql/lib/libgalera_smm.so
#wsrep_sst_method=rsync
wsrep_sst_method=xtrabackup-v2
wsrep_sst_auth=sst:zengxw1

10.16.24.109:
default_storage_engine=InnoDB
innodb_locks_unsafe_for_binlog=1
innodb_autoinc_lock_mode=2
wsrep_provider_options="gcache.size=1999M"
wsrep_cluster_name='pcx_zengxw'
wsrep_cluster_address="gcomm://10.16.24.107,10.16.24.108,10.16.24.109"
wsrep_node_address="10.16.24.109"
wsrep_provider=/usr/local/mysql/lib/libgalera_smm.so
#wsrep_sst_method=rsync
wsrep_sst_method=xtrabackup-v2
wsrep_sst_auth=sst:zengxw1

初始化mysql:
cd /usr/local/mysql
./scripts/mysql_install_db

第一個節點啟動(mysql使用者）：
/etc/init.d/mysql bootstrap-pxc

只需在第一個節點上操作：
delete from mysql.user where user!='root' or host!='localhost';
truncate mysql.db;
drop database test;
grant all privileges on *.* to 'zengxw'@'%' identified by 'zengxw1';
grant all privileges on *.* to 'sst'@'localhost' identified by 'zengxw1';
grant reload,lock tables,replication client on *.* to 'sst'@'localhost' identified by 'zengxw1';
flush privileges;

/usr/local/mysql/bin/mysqladmin -u root password 'safe2016'

其它節點啟動(mysql使用者):
/etc/init.d/mysql start
Stale sst_in_progress file in datadir
Starting MySQL (Percona XtraDB Cluster)State transfer in progress, setting sleep higher
...[  OK  ]

關閉：
/etc/init.d/mysql stop

五.pxc叢集檢查

1.通過檢視4567 埠確認叢集是否啟動：
netstat -plantu | grep mysqld

2.檢查叢集是否正常

 Show global status like "wsrep_local_state";
+-------------------+-------+
| Variable_name     | Value |
+-------------------+-------+
| wsrep_local_state | 4     |  ---表示正常監聽
+-------------------+-------+
1 row in set (0.00 sec)

 Show global status like "wsrep_local_state_comment";
 +---------------------------+--------+
| Variable_name             | Value  |
+---------------------------+--------+
| wsrep_local_state_comment | Synced |
+---------------------------+--------+
1 row in set (0.00 sec)

Show global status like "wsrep_cluster_status";
+----------------------+---------+
| Variable_name        | Value   |
+----------------------+---------+
| wsrep_cluster_status | Primary |  ---表示主節點能正常寫入
+----------------------+---------+
1 row in set (0.00 sec)
 Show global status like "wsrep_connected";
+-----------------+-------+
| Variable_name   | Value |
+-----------------+-------+
| wsrep_connected | ON    |
+-----------------+-------+
1 row in set (0.00 sec)

3.監控叢集
檢視MySQL版本: 
SHOW GLOBAL VARIABLES LIKE 'version';
+---------------+----------------------+
| Variable_name | Value                |
+---------------+----------------------+
| version       | 5.6.21-70.1-25.8-log |
+---------------+----------------------+
1 row in set (0.00 sec)

檢視wsrep版本: mysql> SHOW GLOBAL STATUS LIKE 'wsrep_provider_version';
+------------------------+---------------+
| Variable_name          | Value         |
+------------------------+---------------+
| wsrep_provider_version | 3.8(rf6147dd) |
+------------------------+---------------+
1 row in set (0.00 sec)

檢視wsrep有關的所有變數: mysql> SHOW VARIABLES LIKE 'wsrep%' \G

檢視叢集狀態: mysql> show status like 'wsrep%';
監控狀態引數說明:
叢集完整性檢查:
            wsrep_cluster_state_uuid:在叢集所有節點的值應該是相同的,有不同值的節點,說明其沒有連線入叢集.
            wsrep_cluster_conf_id:正常情況下所有節點上該值是一樣的.如果值不同,說明該節點被臨時"分割槽"了.當節點之間網路連線恢復的時候應該會恢復一樣的值.
            wsrep_cluster_size:如果這個值跟預期的節點數一致,則所有的叢集節點已經連線.
            wsrep_cluster_status:叢集組成的狀態.如果不為"Primary",說明出現"分割槽"或是"split-brain"狀況.
節點狀態檢查:
            wsrep_ready: 該值為ON,則說明可以接受SQL負載.如果為Off,則需要檢查wsrep_connected.
            wsrep_connected: 如果該值為Off,且wsrep_ready的值也為Off,則說明該節點沒有連線到叢集.(可能是wsrep_cluster_address或wsrep_cluster_name等配置錯造成的.具體錯誤

需要檢視錯誤日誌)
            wsrep_local_state_comment:如果wsrep_connected為On,但wsrep_ready為OFF,則可以從該項檢視原因.
複製健康檢查:
            wsrep_flow_control_paused:表示複製停止了多長時間.即表明叢集因為Slave延遲而慢的程度.值為0~1,越靠近0越好,值為1表示複製完全停止.可優化wsrep_slave_threads的值

來改善.
            wsrep_cert_deps_distance:有多少事務可以並行應用處理.wsrep_slave_threads設定的值不應該高出該值太多.
            wsrep_flow_control_sent:表示該節點已經停止複製了多少次.
            wsrep_local_recv_queue_avg:表示slave事務佇列的平均長度.slave瓶頸的預兆.
最慢的節點的wsrep_flow_control_sent和wsrep_local_recv_queue_avg這兩個值最高.這兩個值較低的話,相對更好.
檢測慢網路問題:
            wsrep_local_send_queue_avg:網路瓶頸的預兆.如果這個值比較高的話,可能存在網路瓶
衝突或死鎖的數目:
            wsrep_last_committed:最後提交的事務數目
            wsrep_local_cert_failures和wsrep_local_bf_aborts:回滾,檢測到的衝突數目 

六. 高可用測試
1）在節點1上建立demo資料庫，節點2和節點3是自動建立demo
節點1:
(product)[email protected] [(none)]> create database demo;

(product)[email protected] [(none)]> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| demo               |
| mysql              |
| performance_schema |
+--------------------+

節點2:
(product)[email protected] [(none)]> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| demo               |
| mysql              |
| performance_schema |
+--------------------+

節點3:
(product)[email protected] [(none)]> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| demo               |
| mysql              |
| performance_schema |
+--------------------+

2）在關掉節點1後，操作節點2，再啟動節點1，節點2的資料自動與節點1同步
節點1:
-bash-4.1$ /etc/init.d/mysql stop
Shutting down MySQL (Percona XtraDB Cluster)......[  OK  ]

節點2:
(product)[email protected] [(none)]> use demo
(product)[email protected] [demo]> CREATE TABLE `demo1` (
  `ID` int(11) NOT NULL AUTO_INCREMENT COMMENT '主鍵',
  `OID` int(11) NOT NULL,
  PRIMARY KEY (`ID`)
) ENGINE=InnoDB AUTO_INCREMENT=4 DEFAULT CHARSET=utf8

節點1:
(product)[email protected] [(none)]> use demo
Database changed
(product)[email protected] [demo]> show tables;
+----------------+
| Tables_in_demo |
+----------------+
| demo1          |
+----------------+
1 row in set (0.00 sec)

3) 節點3寫入資料，檢查節點一和節點二是否自動同步

節點3:
insert into demo1 select 1,1;
節點1和節點2:
select * from demo1

節點2:
insert into demo1 select 2,2;
節點1和節點2:
select * from demo1

節點1:
insert into demo1 select 3,3;
節點2和節點3:
select * from demo1