mysql MHA 部署

一、环境说明

monitor节点：10.10.2.10

master节点： 10.10.2.72

slave1节点:  10.10.2.73（候选master）

slave2节点:  10.10.2.74

二、配置root免密码登录

1、所有节点执行：

ssh-keygen -t rsa （生成密钥）

2、4台节点公钥的内容统一存到authorized_keys 文件中，分发到4个节点root的/root/.ssh/目录下，权限600；

三、数据库主从配置跳过，下面主要记录主从半同步配置

1、主库

mysql> install plugin rpl_semi_sync_master soname'semisync_master.so';
mysql> set global rpl_semi_sync_master_enabled=1;
mysql> set global rpl_semi_sync_master_timeout=3000;

PS：3s超时，半同步复制是如果从服务端没有开启的话，主服务端第一次会延迟3秒中之后提交，

之后主服务端会降低延迟不再等待从服务端。从服务端开启之后在追赶上主服务端让后在实现半同步。

验证是否打开了semisync功能：

mysql> show global status like 'rpl%';

让mysql在重启时自动加载该功能，在/etc/my.cnf中加入：

rpl_semi_sync_master_enabled=1
rpl_semi_sync_master_timeout=1000

2、从库

mysql> install plugin rpl_semi_sync_slave soname'semisync_slave.so';
mysql> set global rpl_semi_sync_slave_enabled=1;

验证是否打开了semisync功能：

mysql> show global status like 'rpl%';

让mysql在重启时自动加载该功能，在/etc/my.cnf中加入：

rpl_semi_sync_slave_enabled=1

四、设置的复制权限帐号（管理节点、主库、候选主库）

GRANT ALL PRIVILEGES ON *.* TO 'mha_rep'@'10.10.2.10' IDENTIFIED BY '123456';
GRANT ALL PRIVILEGES ON *.* TO 'mha_rep'@'10.10.2.72' IDENTIFIED BY '123456';
GRANT ALL PRIVILEGES ON *.* TO 'mha_rep'@'10.10.2.73' IDENTIFIED BY '123456';
FLUSH PRIVILEGES;

五、MHA安装配置

1、安装

数据库节点：

yum install perl-DBD-MySQL perl-Config-Tiny perl-Log-Dispatch perl-Parallel-ForkManager perl-Config-IniFiles -y
rpm -ivh http://mysql-master-ha.googlecode.com/files/mha4mysql-node-0.53-0.el6.noarch.rpm

管理节点：

yum install perl-DBD-MySQL perl-Config-Tiny perl-Log-Dispatch perl-Parallel-ForkManager perl-Config-IniFiles -y
rpm -ivh http://mysql-master-ha.googlecode.com/files/mha4mysql-node-0.53-0.el6.noarch.rpm
yum -y install perl-Time-HiRes
rpm -ivh http://mysql-master-ha.googlecode.com/files/mha4mysql-manager-0.53-0.el6.noarch.rpm

2、配置（所有节点一样）

mkdir /var/log/masterha/app1 -p #创建work目录
yum -y install mysql #安装命令

撰写配置文件：

[server default]
manager_workdir=/var/log/masterha/app1
manager_log=/var/log/masterha/app1/manager.log
user=mha_rep          //mysql的管理用户名  必要的时候要执行 stop slave ,change master reset slave 等命令。
password=123456       //mysql的管理员密码
ssh_user=root         //使用root账号ssh登录
repl_user=mha_rep     //mysql复制账号
repl_password=123456  //mysql复制账号密码
ping_interval=1       //ping间隔，这里设置一次ping不通即failover
shutdown_script=""    //强制关机脚本，这里设为空
#master_ip_failover_script="/usr/local/bin/master_ip_failover" //主从故障切换脚本
#master_ip_online_change_script=""
#report_script=""
[server1]
hostname=10.10.2.72
candidate_master=1
port=13306
master_binlog_dir=/data/mysql/auth/db/
[server2]
hostname=10.10.2.73
candidate_master=1
port=13306
master_binlog_dir=/data/mysql/auth/db/
[server3]
hostname=10.10.2.74
no_master=1
port=13306

六、测试

1、测试ssh是否正常

[root@db_lvs_master ~]# masterha_check_ssh --conf=/etc/masterha/app1.cnf
Sat Nov 30 15:49:36 2013 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Sat Nov 30 15:49:36 2013 - [info] Reading application default configurations from /etc/masterha/app1.cnf..
Sat Nov 30 15:49:36 2013 - [info] Reading server configurations from /etc/masterha/app1.cnf..
Sat Nov 30 15:49:36 2013 - [info] Starting SSH connection tests..
Sat Nov 30 15:49:37 2013 - [debug]
Sat Nov 30 15:49:36 2013 - [debug]  Connecting via SSH from [email protected](10.10.2.72:22) to [email protected](10.10.2.73:22)..
Sat Nov 30 15:49:36 2013 - [debug]   ok.
Sat Nov 30 15:49:36 2013 - [debug]  Connecting via SSH from [email protected](10.10.2.72:22) to [email protected](10.10.2.74:22)..
Sat Nov 30 15:49:37 2013 - [debug]   ok.
Sat Nov 30 15:49:37 2013 - [debug]
Sat Nov 30 15:49:37 2013 - [debug]  Connecting via SSH from [email protected](10.10.2.73:22) to [email protected](10.10.2.72:22)..
Sat Nov 30 15:49:37 2013 - [debug]   ok.
Sat Nov 30 15:49:37 2013 - [debug]  Connecting via SSH from [email protected](10.10.2.73:22) to [email protected](10.10.2.74:22)..
Sat Nov 30 15:49:37 2013 - [debug]   ok.
Sat Nov 30 15:49:38 2013 - [debug]
Sat Nov 30 15:49:37 2013 - [debug]  Connecting via SSH from [email protected](10.10.2.74:22) to [email protected](10.10.2.72:22)..
Sat Nov 30 15:49:37 2013 - [debug]   ok.
Sat Nov 30 15:49:37 2013 - [debug]  Connecting via SSH from [email protected](10.10.2.74:22) to [email protected](10.10.2.73:22)..
Sat Nov 30 15:49:38 2013 - [debug]   ok.
Sat Nov 30 15:49:38 2013 - [info] All SSH connection tests passed successfully.

2、测试复制是否正常

[root@db_lvs_master ~]# masterha_check_repl --conf=/etc/masterha/app1.cnf
Sat Nov 30 15:50:15 2013 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.
Sat Nov 30 15:50:15 2013 - [info] Reading application default configurations from /etc/masterha/app1.cnf..
Sat Nov 30 15:50:15 2013 - [info] Reading server configurations from /etc/masterha/app1.cnf..
Sat Nov 30 15:50:15 2013 - [info] MHA::MasterMonitor version 0.53.
Sat Nov 30 15:50:15 2013 - [info] Dead Servers:
Sat Nov 30 15:50:15 2013 - [info] Alive Servers:
Sat Nov 30 15:50:15 2013 - [info]   10.10.2.72(10.10.2.72:13306)
Sat Nov 30 15:50:15 2013 - [info]   10.10.2.73(10.10.2.73:13306)
Sat Nov 30 15:50:15 2013 - [info]   10.10.2.74(10.10.2.74:13306)
Sat Nov 30 15:50:15 2013 - [info] Alive Slaves:
Sat Nov 30 15:50:15 2013 - [info]   10.10.2.73(10.10.2.73:13306)  Version=5.5.15-log (oldest major version between slaves) log-bin:enabled
Sat Nov 30 15:50:15 2013 - [info]     Replicating from 10.10.2.72(10.10.2.72:13306)
Sat Nov 30 15:50:15 2013 - [info]     Primary candidate for the new Master (candidate_master is set)
Sat Nov 30 15:50:15 2013 - [info]   10.10.2.74(10.10.2.74:13306)  Version=5.5.15-log (oldest major version between slaves) log-bin:enabled
Sat Nov 30 15:50:15 2013 - [info]     Replicating from 10.10.2.72(10.10.2.72:13306)
Sat Nov 30 15:50:15 2013 - [info]     Not candidate for the new Master (no_master is set)
Sat Nov 30 15:50:15 2013 - [info] Current Alive Master: 10.10.2.72(10.10.2.72:13306)
Sat Nov 30 15:50:15 2013 - [info] Checking slave configurations..
Sat Nov 30 15:50:15 2013 - [info]  read_only=1 is not set on slave 10.10.2.73(10.10.2.73:13306).
Sat Nov 30 15:50:15 2013 - [warning]  relay_log_purge=0 is not set on slave 10.10.2.73(10.10.2.73:13306).
Sat Nov 30 15:50:15 2013 - [info]  read_only=1 is not set on slave 10.10.2.74(10.10.2.74:13306).
Sat Nov 30 15:50:15 2013 - [warning]  relay_log_purge=0 is not set on slave 10.10.2.74(10.10.2.74:13306).
Sat Nov 30 15:50:15 2013 - [info] Checking replication filtering settings..
Sat Nov 30 15:50:15 2013 - [info]  binlog_do_db= , binlog_ignore_db= information_schema,mysql,performance_schema,test
Sat Nov 30 15:50:15 2013 - [info]  Replication filtering check ok.
Sat Nov 30 15:50:15 2013 - [info] Starting SSH connection tests..
Sat Nov 30 15:50:17 2013 - [info] All SSH connection tests passed successfully.
Sat Nov 30 15:50:17 2013 - [info] Checking MHA Node version..
Sat Nov 30 15:50:17 2013 - [info]  Version check ok.
Sat Nov 30 15:50:17 2013 - [info] Checking SSH publickey authentication settings on the current master..
Sat Nov 30 15:50:17 2013 - [info] HealthCheck: SSH to 10.10.2.72 is reachable.
Sat Nov 30 15:50:17 2013 - [info] Master MHA Node version is 0.53.
Sat Nov 30 15:50:17 2013 - [info] Checking recovery script configurations on the current master..
Sat Nov 30 15:50:17 2013 - [info]   Executing command: save_binary_logs --command=test --start_pos=4 --binlog_dir=/data/mysql/auth/db/ --output_file=/var/tmp/save_binary_logs_test --manager_version=0.53 --start_file=mysql-bin.000012
Sat Nov 30 15:50:17 2013 - [info]   Connecting to [email protected](10.10.2.72)..
Creating /var/tmp if not exists..    ok.
Checking output directory is accessible or not..
ok.
Binlog found at /data/mysql/auth/db/, up to mysql-bin.000012
Sat Nov 30 15:50:18 2013 - [info] Master setting check done.
Sat Nov 30 15:50:18 2013 - [info] Checking SSH publickey authentication and checking recovery script configurations on all alive slave servers..
Sat Nov 30 15:50:18 2013 - [info]   Executing command : apply_diff_relay_logs --command=test --slave_user=mha_rep --slave_host=10.10.2.73 --slave_ip=10.10.2.73 --slave_port=13306 --workdir=/var/tmp --target_version=5.5.15-log --manager_version=0.53 --relay_log_info=/data/mysql/auth/db/relay-log.info  --relay_dir=/data/mysql/auth/db/  --slave_pass=xxx
Sat Nov 30 15:50:18 2013 - [info]   Connecting to [email protected](10.10.2.73:22)..
Checking slave recovery environment settings..
Opening /data/mysql/auth/db/relay-log.info ... ok.
Relay log found at /data/mysql/auth/db, up to mysqld-relay-bin.000012
Temporary relay log file is /data/mysql/auth/db/mysqld-relay-bin.000012
Testing mysql connection and privileges.. done.
Testing mysqlbinlog output.. done.
Cleaning up test file(s).. done.
Sat Nov 30 15:50:18 2013 - [info]   Executing command : apply_diff_relay_logs --command=test --slave_user=mha_rep --slave_host=10.10.2.74 --slave_ip=10.10.2.74 --slave_port=13306 --workdir=/var/tmp --target_version=5.5.15-log --manager_version=0.53 --relay_log_info=/data/mysql/auth/db/relay-log.info  --relay_dir=/data/mysql/auth/db/  --slave_pass=xxx
Sat Nov 30 15:50:18 2013 - [info]   Connecting to [email protected](10.10.2.74:22)..
Checking slave recovery environment settings..
Opening /data/mysql/auth/db/relay-log.info ... ok.
Relay log found at /data/mysql/auth/db, up to mysqld-relay-bin.000012
Temporary relay log file is /data/mysql/auth/db/mysqld-relay-bin.000012
Testing mysql connection and privileges.. done.
Testing mysqlbinlog output.. done.
Cleaning up test file(s).. done.
Sat Nov 30 15:50:18 2013 - [info] Slaves settings check done.
Sat Nov 30 15:50:18 2013 - [info]
10.10.2.72 (current master)
+--10.10.2.73
+--10.10.2.74
Sat Nov 30 15:50:18 2013 - [info] Checking replication health on 10.10.2.73..
Sat Nov 30 15:50:18 2013 - [info]  ok.
Sat Nov 30 15:50:18 2013 - [info] Checking replication health on 10.10.2.74..
Sat Nov 30 15:50:18 2013 - [info]  ok.
Sat Nov 30 15:50:18 2013 - [warning] master_ip_failover_script is not defined.
Sat Nov 30 15:50:18 2013 - [warning] shutdown_script is not defined.
Sat Nov 30 15:50:18 2013 - [info] Got exit code 0 (Not master dead).
MySQL Replication Health is OK.

七、启动管理节点进程：

nohup masterha_manager --conf=/etc/masterha/app1.cnf  --ignore_fail_on_start  --ignore_last_failover > /tmp/mha_manager.log  < /dev/null 2>&1 &

查看状态：

[root@db_lvs_master ~]# masterha_check_status --conf=/etc/masterha/app1.cnf
app1 (pid:19135) is running(0:PING_OK), master:10.10.2.72

本文出自 “�潘吭宋�男” 博客，谢绝转载！