vlambda博客
学习文章列表

【干货】Mha-Atlas-MySQL高可用方案实践(一)

点击上方蓝字发现更多精彩

一,mysql-mha环境准备

1.1 实验环境:

主机名 IP地址(NAT) 描述
mysql-db01 eth0:192.168.0.51 系统:CentOS6.5(6.x都可以) 安装:mysql5.6
mysql-db02 eth0:192.168.0.52 系统:CentOS6.5(6.x都可以) 安装:mysql5.6
mysql-db03 eth0:192.168.0.53 系统:CentOS6.5(6.x都可以) 安装:mysql5.6

1.2 软件包

1) mha管理节点安装包:

mha4mysql-manager-0.56-0.el6.noarch.rpmmha4mysql-manager-0.56.tar.gz

2) mha node节点安装包:

mha4mysql-node-0.56-0.el6.noarch.rpmmha4mysql-node-0.56.tar.gz

3) mysql中间件:

Atlas-2.2.1.el6.x86_64.rpm

4) mysql源码安装包

mysql-5.6.17-linux-glibc2.5-x86_64.tar

1.3 主机名映射

[root@mysql-db01 ~]# cat /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4::1 localhost localhost.localdomain localhost6 localhost6.localdomain6192.168.0.51 mysql-db01192.168.0.52 mysql-db02192.168.0.53 mysql-db03

1.4 关闭selinux和iptables

[root@mysql-db01 ~]# vim /etc/sysconfig/selinux [root@mysql-db01 ~]# cat /etc/sysconfig/selinux | grep -v "#"
SELINUX=disabledSELINUXTYPE=targeted
[root@mysql-db01 ~]# setenforce 0[root@mysql-db01 ~]# service iptables stopiptables: Setting chains to policy ACCEPT: filter [ OK ]iptables: Flushing firewall rules: [ OK ]iptables: Unloading modules: [ OK ][root@mysql-db01 ~]# chkconfig iptables off

二,简介

2.1 作者简介

姓名:松信嘉范
MySQL/Linux专家
2001年索尼公司入职
2001年开始使用oracle
2004年开始使用MySQL
2006年9月-2010年8月MySQL从事顾问
2010年-2012年DeNA
2012年至今Facebook

2.2 软件简介

  • MHA(Master High Availability)目前在MySQL高可用方面是一个相对成熟的解决方案,是一套优秀的作为MySQL高可用性环境下故障切换和主从提升的高可用软件。在MySQL故障切换过程中,MHA能做到0~30秒之内自动完成数据库的故障切换操作,并且在进行故障切换过程中,MHA能最大程度上保证数据库的一致性,以达到真正意义上的高可用。

  • MHA由两部分组成:MHA Manager(管理节点)和MHA Node(数据节点)。MHA Manager可以独立部署在一台独立的机器上管理多个Master-Slave集群,也可以部署在一台Slave上。当Master出现故障时,它可以自动将最新数据的Slave提升为新的Master,然后将所有其他的Slave重新指向新的Master。整个故障转移过程对应程序是完全透明的。

2.3 工作流程

  • 从宕机崩溃的master保存二进制日志事件(binlog events);

  • 识别含有最新更新的slave;

  • 应用差异的中继日志(relay log)到其他的slave;

  • 应用从master保存的二进制日志事件(binlog events);

  • 提升一个slave为新的master;

  • 使其他的slave连接新的master进行复制;

2.4 MHA架构图

【干货】Mha-Atlas-MySQL高可用方案实践(一)

2.5 MHA工具介绍

MHA软件由两部分组成,Manager工具包和Node工具包,具体的说明如下:

 
#Manager工具包主要包括以下几个工具:
masterha_check_ssh #检查MHA的SSH配置状况masterha_check_repl #检查MySQL复制状况masterha_check_status #检测当前MHA运行状态masterha_master_monitor #检测master是否宕机masterha_manger #启动MHAmasterha_master_switch #控制故障转移(自动或者手动)masterha_conf_host #添加或删除配置的server信息masterha_secondary_check #试图建立TCP连接从远程服务器masterha_stop #停止MHA
#Node工具包主要包括以下几个工具:
save_binary_logs #保存和复制master的二进制日志apply_diff_relay_logs #识别差异的中继日志事件filter_mysqlbinlog #去除不必要的ROLLBACK事件purge_relay_logs            #清除中继日志

三,mysql环境准备

3.1 环境检查

  • mysql-db01

#系统版本[root@mysql-db01 bin]# cat /etc/redhat-release CentOS release 6.5 (Final)[root@mysql-db01 bin]# uname -r2.6.32-431.el6.x86_64[root@mysql-db01 bin]# hostname -I192.168.0.51 
  • mysql-db02

#系统版本
[root@mysql-db02 ~]# cat /etc/redhat-release CentOS release 6.5 (Final)[root@mysql-db02 ~]# uname -r2.6.32-431.el6.x86_64[root@mysql-db02 ~]# hostname -I192.168.0.52 
  • mysql-db03

#系统版本
[root@mysql-db03 ~]# cat /etc/redhat-release CentOS release 6.5 (Final)[root@mysql-db03 ~]# uname -r2.6.32-431.el6.x86_64[root@mysql-db03 ~]# hostname -I192.168.0.53 

3.2 安装mysql

3.2.1 安装包准备

 
[root@mysql-db01 ~]# lsanaconda-ks.cfg mha4mysql-manager-0.56.tar.gz rpminstall.log mha4mysql-node-0.56.tar.gzinstall.log.syslog mysql-5.6.17-linux-glibc2.5-x86_64.tar.gz[root@mysql-db01 ~]# ll mysql-5.6.17-linux-glibc2.5-x86_64.tar.gz -rw-r--r--. 1 root root 305102088 Sep 3 21:33 mysql-5.6.17-linux-glibc2.5-x86_64.tar.gz

3.2.2 安装(3台都装)

 
[root@mysql-db01 ~]# yum -y install ncurses-devel[root@mysql-db01 ~]# yum -y install libaio[root@mysql-db01 ~]# tar xf mysql-5.6.17-linux-glibc2.5-x86_64.tar.gz -C /usr/local/[root@mysql-db01 ~]# ln -s /usr/local/mysql-5.6.17-linux-glibc2.5-x86_64 /usr/local/mysql[root@mysql-db01 ~]# useradd mysql -s /sbin/nologin -M[root@mysql-db01 ~]# /usr/local/mysql/scripts/mysql_install_db --user=mysql --basedir=/usr/local/mysql --datadir=/usr/local/mysql/data/[root@mysql-db01 ~]# /bin/cp /usr/local/mysql/support-files/my-default.cnf /etc/my.cnf[root@mysql-db01 ~]# /bin/cp /usr/local/mysql/support-files/mysql.server /etc/init.d/mysqld[root@mysql-db01 ~]# ln -s /usr/local/mysql/bin/* /usr/local/bin/[root@mysql-db01 ~]# which mysqladmin/usr/local/bin/mysqladmin

3.2.3 加入开机自启动并启动mysql

 
[root@mysql-db01 ~]# chkconfig mysqld on[root@mysql-db01 ~]# chkconfig mysqld --listmysqld 0:off 1:off 2:on 3:on 4:on 5:on 6:off[root@mysql-db01 ~]# /etc/init.d/mysqld startStarting MySQL. SUCCESS!

3.2.4 配置密码

 
[root@mysql-db01 ~]# mysqladmin -uroot password '123123'

四,配置基于GTID的主从复制

4.1 先决条件

  • 主库和从库都要开启binlog

  • 主库和从库server-id不同

  • 要有主从复制用户

4.2 主库操作(mysql-db01)

4.2.1 修改配置文件

#修改主库配置文件/etc/my.cnf
[root@mysql-db01 mysql]# cat /etc/my.cnf [client]socket = /usr/local/mysql/data/mysql.sock[mysqld]lower_case_table_names = 1default-storage-engine = InnoDBport = 3306datadir = /usr/local/mysql/datacharacter-set-server = utf8socket = /usr/local/mysql/data/mysql.sock
log_bin = mysql-bin #开启binlog日志server_id = 1 #设置server_id
innodb_buffer_pool_size = 200Mslave-parallel-workers = 8thread_cache_size = 600back_log = 600slave_net_timeout = 60max_binlog_size = 512Mkey_buffer_size = 8Mquery_cache_size = 64Mjoin_buffer_size = 2Msort_buffer_size = 2Mquery_cache_type = 1thread_stack = 192K
#重启动MySQL服务[root@mysql-db01 mysql]# /etc/init.d/mysqld restartShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS!

4.2.2 登陆MySQL删除不必要的用户并创建主从复制用户

1)删除不必要的用户

 
 mysql> select user,host from mysql.user;+------+------------+| user | host |+------+------------+| root | 127.0.0.1 || root | ::1 || | localhost || root | localhost || | mysql-db01 || root | mysql-db01 |+------+------------+6 rows in set (0.00 sec)
mysql> drop user root@'127.0.0.1';Query OK, 0 rows affected (0.00 sec)
mysql> drop user root@'::1';Query OK, 0 rows affected (0.00 sec)
mysql> drop user ' '@'localhost';Query OK, 0 rows affected (0.00 sec)
mysql> drop user ' '@'mysql-db01';Query OK, 0 rows affected (0.00 sec)
mysql> select user,host from mysql.user;+------+------------+| user | host |+------+------------+| root | localhost || root | mysql-db01 |+------+------------+2 rows in set (0.00 sec)

2)创建主从复制用户

 
mysql> grant replication slave on *.* to rep@'192.168.0.%' identified by '123123';Query OK, 0 rows affected (0.00 sec)
mysql> select user,host from mysql.user;+------+-------------+| user | host |+------+-------------+| rep | 192.168.0.% || root | localhost || root | mysql-db01 |+------+-------------+3 rows in set (0.00 sec)
mysql> show grants for rep@'192.168.0.%';+--------------------------------------------------------------------------------------------------------------------------+| Grants for rep@192.168.0.% |+--------------------------------------------------------------------------------------------------------------------------+| GRANT REPLICATION SLAVE ON *.* TO 'rep'@'192.168.0.%' IDENTIFIED BY PASSWORD '*E56A114692FE0DE073F9A1DD68A00EEB9703F3F1' |+--------------------------------------------------------------------------------------------------------------------------+1 row in set (0.00 sec)

4.3 从库操作(mysql-db02和mysql-db03)

4.3.1 修改配置文件

#修改mysql-db02配置文件(和mysql-db01配置文件一致)#只需要修改server-id = 5选项[root@mysql-db02 ~]# cat /etc/my.cnf [client]socket = /usr/local/mysql/data/mysql.sock[mysqld]lower_case_table_names = 1default-storage-engine = InnoDBport = 3306datadir = /usr/local/mysql/datacharacter-set-server = utf8socket = /usr/local/mysql/data/mysql.sock
log_bin = mysql-bin #从binlog也要打开server_id = 5 #仅需修改此项
innodb_buffer_pool_size = 200Mslave-parallel-workers = 8thread_cache_size = 600back_log = 600slave_net_timeout = 60max_binlog_size = 512Mkey_buffer_size = 8Mquery_cache_size = 64Mjoin_buffer_size = 2Msort_buffer_size = 2Mquery_cache_type = 1thread_stack = 192K[root@mysql-db02 ~]# /etc/init.d/mysqld restart #重启mysqlShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS!

#修改mysql-db03配置文件(和mysql-db01配置文件一致)#只需要修改server-id = 10选项[root@mysql-db03 ~]# cat /etc/my.cnf[client]socket = /usr/local/mysql/data/mysql.sock[mysqld]lower_case_table_names = 1default-storage-engine = InnoDBport = 3306datadir = /usr/local/mysql/datacharacter-set-server = utf8socket = /usr/local/mysql/data/mysql.sock
log_bin = mysql-bin #从binlog也要打开server_id = 10 #只需修改此项
innodb_buffer_pool_size = 200Mslave-parallel-workers = 8thread_cache_size = 600back_log = 600slave_net_timeout = 60max_binlog_size = 512Mkey_buffer_size = 8Mquery_cache_size = 64Mjoin_buffer_size = 2Msort_buffer_size = 2Mquery_cache_type = 1thread_stack = 192K[root@mysql-db03 ~]# /etc/init.d/mysqld restart #重启mysqlShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS!

特别提示:
在以往如果是基于binlog日志的主从复制,则必须要记住主库的master状态信息。

【干货】Mha-Atlas-MySQL高可用方案实践(一)

但是在MySQL5.6版本里多了一个Gtid的功能,可以自动记录主从复制位置点的信息,并在日志中输出出来。


4.4 开启GTID

#没开启之前先看一下GTID状态mysql> show global variables like '%gtid%';+--------------------------+-------+| Variable_name | Value |+--------------------------+-------+| enforce_gtid_consistency | OFF || gtid_executed | || gtid_mode | OFF || gtid_owned | || gtid_purged | |+--------------------------+-------+5 rows in set (0.00 sec)

编辑mysql配置文件(主库从库都需要修改)

【干货】Mha-Atlas-MySQL高可用方案实践(一)

mysql-db01,mysql-db02,mysql-db03都需要加入上图的上行代码

修改完配置文件以后重启动数据库

[root@mysql-db01 ~]# /etc/init.d/mysqld restartShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS! [root@mysql-db02 ~]# /etc/init.d/mysqld restartShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS![root@mysql-db03 ~]# /etc/init.d/mysqld restartShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS!

再次查看GTID状态

[root@mysql-db01 ~]# mysql -uroot -p123123Warning: Using a password on the command line interface can be insecure.Welcome to the MySQL monitor. Commands end with ; or \g.Your MySQL connection id is 1Server version: 5.6.17-log MySQL Community Server (GPL)
Copyright (c) 2000, 2014, Oracle and/or its affiliates. All rights reserved.
Oracle is a registered trademark of Oracle Corporation and/or itsaffiliates. Other names may be trademarks of their respectiveowners.
Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
mysql> show global variables like '%gtid%';+--------------------------+-------+| Variable_name | Value |+--------------------------+-------+| enforce_gtid_consistency | ON | #执行GTID一致| gtid_executed | | | gtid_mode | ON | #开启GTID模块| gtid_owned | || gtid_purged | |+--------------------------+-------+5 rows in set (0.00 sec)
mysql>

再次提示:
主库从库都必须要开启GTID,否则在做主从复制的时候就会报错.

4.5 配置主从复制(mysql-db02,mysql-db03)

 
mysql> change master to \ -> master_host='192.168.0.51',\ #主库IP -> master_user='rep',\ #主库复制用户 -> master_password='123123',\ #主库复制用密码 -> master_auto_position=1; #GTID位置点(自动追踪需要同步的position)Query OK, 0 rows affected, 2 warnings (0.00 sec)

4.6 开启从库的主从复制功能(mysql-db02,mysql-db03)

mysql> start slave; #开启主从同步功能Query OK, 0 rows affected, 1 warning (0.01 sec)mysql> show slave status\G*************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.0.51 Master_User: rep Master_Port: 3306 Connect_Retry: 60 Master_Log_File: mysql-bin.000003 Read_Master_Log_Pos: 151 Relay_Log_File: mysql-db02-relay-bin.000002 Relay_Log_Pos: 361 Relay_Master_Log_File: mysql-bin.000003 Slave_IO_Running: Yes #此项yes代表成功 Slave_SQL_Running: Yes #此项yes代表成功 Replicate_Do_DB:  Replicate_Ignore_DB:  Replicate_Do_Table:  Replicate_Ignore_Table:  #....以下省略若干行....

两个从库mysql-db02和mysql-db03都执行以上步骤。

4.7 什么是GTID

  • GTID(Global Transaction)全局事务标识符:是一个唯一的标识符,它创建并与源服务器(主)上提交的每个事务相关联。此标识符不仅对其发起的服务器是唯一的,而且在给定复制设置中的所有服务器上都是唯一的。所有交易和所有GTID之间都有1对1的映射。

  • GTID实际上是由UUID+TID组成的。其中UUID是一个MySQL实例的唯一标识。TID代表了该实例上已经提交的事务数量,并且随着事务提交单调递增。

  • 下面是一个GTID的具体形式:

3E11FA47-71CA-11E1-9E33-C80AA9429562:23

4.8 GTID的新特性

(1)支持多线程复制:事实上是针对每个database开启相应的独立线程,即每个库有一个单独的(sql thread)

(2)支持启用GTID,在配置主从复制,传统的方式里,你需要找到binlog和POS点,然后change master to 指向。在mysql5.6里,无须再知道binlog和POS点,只需要知道master的IP/端口/账号密码即可,因为同步复制是自动的,MySQL通过内部机制GTID自动找点同步。

(3)基于Row复制只保存改变的列,大大节省磁盘空间,网络,内存等

(4)支持把Master和Slave的相关信息记录在Table中;原来是记录在文件里,现在则记录在表里,增强可用性

(5)支持延迟复制

4.9 开启方法

 
#mysql配置文件:[mysqld]gtid_mode=ONenforce_gtid_consistency
#查看show global variables like ‘%gtid%’;
4.10 从库设置(mysql-db02,mysql-db03)#登陆从库[root@mysql-db02 ~]# mysql -uroot -p123123
#临时禁用自动删除relay log功能mysql> set global relay_log_purge = 0;Query OK, 0 rows affected (0.00 sec)
#设置只读mysql> set global read_only=1;Query OK, 0 rows affected (0.00 sec)

编辑配置文件/etc/my.cnf

【干货】Mha-Atlas-MySQL高可用方案实践(一)

修改完配置文件,别忘了重启动mysql服务

root@mysql-db02 ~]# /etc/init.d/mysqld restartShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS! root@mysql-db03 ~]# /etc/init.d/mysqld restartShutting down MySQL.. SUCCESS! Starting MySQL. SUCCESS!

五,部署MHA

5.1 环境准备(所有节点mysql-db01,mysql-db02,mysql-db03)

 
#光盘安装依赖包[root@mysql-db01 ~]# yum -y install perl-DBD-MySQL
#安装mha4mysql-node-0.56-0.el6.noarch.rpm[root@mysql-db01 rpm]# rpm -ivh mha4mysql-node-0.56-0.el6.noarch.rpm Preparing... ########################################### [100%] 1:mha4mysql-node ########################################### [100%]
[root@mysql-db01 ~]# mysql -uroot -p123123
mysql> grant all privileges on *.* to mha@'192.168.0.%' identified by '123123';Query OK, 0 rows affected (0.00 sec)
mysql> select user,host from mysql.user where user='mha';+------+-------------+| user | host |+------+-------------+| mha | 192.168.0.% | #主库上创建从库会自动复制+------+-------------+1 row in set (0.00 sec)
#特别提示:3台MySQL都需要安装mha4mysql-node-0.56-0.el6.noarch.rpm

5.2 部署管理节点(mha-manager)

5.2.1 在mysql-db03上部署管理节点

 
#使用阿里云源+epel源wget -O /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-6.repowget -O /etc/yum.repos.d/epel-6.repo http://mirrors.aliyun.com/repo/epel-6.repo
#安装manager依赖包(需要公网源)[root@mysql-db03 ~]# yum -y install perl-Config-Tiny epel-release perl-Log-Dispatch perl-Parallel-ForkManager perl-Time-HiRes
#安装manager包[root@mysql-db03 rpm]# rpm -ivh mha4mysql-manager-0.56-0.el6.noarch.rpm Preparing... ########################################### [100%] 1:mha4mysql-manager ########################################### [100%]

5.2.2 编辑配置文件

 
#创建配置文件目录[root@mysql-db03 ~]# mkdir -p /etc/mha
#创建日志目录[root@mysql-db03 ~]# mkdir -p /var/log/mha/mha1
#创建配置文件(默认没有)[root@mysql-db03 ~]# cd /etc/mha/[root@mysql-db03 mha]# ls[root@mysql-db03 mha]# vim /etc/mha/mha1.cnf[root@mysql-db03 mha]# cat /etc/mha/mha1.cnf[server default]manager_log=/var/log/mha/mha1/manager #manager管理日志存放路径manager_workdir=/var/log/mha/mha1 #manager管理日志的目录路径master_binlog_dir=/usr/local/mysql/data #binlog日志的存放路径user=mha #管理账户password=123123 #管理账户密码ping_interval=2 #存活检查的间隔时间repl_user=rep #主从复制的授权账户repl_password=123123 #主从复制的授权账户密码ssh_user=root #用于ssh连接的账户
[server1]hostname=192.168.0.51 port=3306
[server2]#candidate_master=1 #此条暂时注释掉(后面解释)#check_repl_delay=0 #此条暂时注释掉(后面解释)hostname=192.168.0.52port=3306
[server3]hostname=192.168.0.53port=3306
#**特别提示:**#以上配置文件内容里每行的最后不要留有空格,因此,不能复制的呦

特别说明:

参数:candidate_master=1解释:设置为候选master,如果设置该参数以后,发生主从切换以后会将此从库提升为主库,即使这个主库不是集群中事件最新的slave参数:check_repl_delay=0解释:默认情况下如果一个slave落后master 100M的relay logs 的话,MHA将不会选择该slave作为一个新的master,因为对于这个slave的恢复需要花费很长时间,通过设置check_repl_delay=0,MHA触发切换在选择一个新的master的时候将会忽略复制延时,这个参数对于设置了candidate_master=1的主机非常有用,因为这个候选主在切换的过程中一定是新的master

5.3 配置ssh信任(所有节点mysql-db01,mysql-db02,mysql-db03)

 
#创建密钥对[root@mysql-db03 ~]# ssh-keygen -t dsa -P "" -f ~/.ssh/id_dsa >/dev/null 2>&1
#发送mysql-db03公钥,包括自己[root@mysql-db03 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected][root@mysql-db03 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected][root@mysql-db03 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected]
#发送mysql-db02公钥,包括自己[root@mysql-db02 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected][root@mysql-db02 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected][root@mysql-db02 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected]
#发送mysql-db01公钥,包括自己[root@mysql-db01 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected][root@mysql-db01 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected][root@mysql-db01 ~]# ssh-copy-id -i /root/.ssh/id_dsa.pub [email protected]

5.4 启动测试

5.4.1 ssh检查检测

 
[root@mysql-db03 ~]# masterha_check_ssh --conf=/etc/mha/mha1.cnf #ssh检查命令Tue Sep 5 03:01:38 2017 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping.Tue Sep 5 03:01:38 2017 - [info] Reading application default configuration from /etc/mha/mha1.cnf..Tue Sep 5 03:01:38 2017 - [info] Reading server configuration from /etc/mha/mha1.cnf..Tue Sep 5 03:01:38 2017 - [info] Starting SSH connection tests....中间省略若干行..Tue Sep 5 03:01:40 2017 - [debug] Connecting via SSH from root@192.168.0.53(192.168.0.53:22) to root@192.168.0.52(192.168.0.52:22)..Tue Sep 5 03:01:40 2017 - [debug] ok.Tue Sep  5 03:01:40 2017 - [info] All SSH connection tests passed successfully. #出现这个就表示成功

5.4.2 主从复制检测

(1)错误的主从复制检测

[root@mysql-db03 ~]# masterha_check_repl --conf=/etc/mha/mha1.cnf如果不出意外,同学们的检测结果都会是下面的样子

【干货】Mha-Atlas-MySQL高可用方案实践(一)

因此在mysql-db02和mysql-db03上添加主从复制的用户即可。
grant replication slave on *.* to rep@'192.168.0.%' identified by '123123';
再次检查如下图所示:

【干货】Mha-Atlas-MySQL高可用方案实践(一)

5.5 启动MHA

 
#启动[root@mysql-db03 ~]# nohup masterha_manager --conf=/etc/mha/mha1.cnf --remove_dead_master_conf --ignore_last_failover < /dev/null > /var/log/mha/mha1/manager.log 2>&1 &[root@mysql-db03 ~]# ps -ef | grep perl | grep -v greproot 4961 4690 0 06:33 pts/2 00:00:00 perl /usr/bin/masterha_manager --conf=/etc/mha/mha1.cnf --remove_dead_master_conf --ignore_last_failover
#说明:nohup:启动命令--conf:指定配置文件位置--remove_dead_master_conf:如果有master down了,就去掉配置文件里该master的部分。

5.6 进行mha自动切换master的测试

初始状态:

【干货】Mha-Atlas-MySQL高可用方案实践(一)

(1)登陆mysql-db02(192.168.0.53)查看信息状态

 
#登陆数据库mysql-db02(192.168.0.53)[root@mysql-db03 ~]# mysql -uroot -p123123mysql> show slave status\G*************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.0.51 #这是主库IP地址 Master_User: rep Master_Port: 3306 Connect_Retry: 60 Master_Log_File: mysql-bin.000003 Read_Master_Log_Pos: 1656 Relay_Log_File: mysql-db02-relay-bin.000004 Relay_Log_Pos: 1796 Relay_Master_Log_File: mysql-bin.000003 Slave_IO_Running: Yes Slave_SQL_Running: Yes            ..以下省略若干内容..

(2)停掉mysql-db01(192.168.0.51)上的MySQL服务

 
[root@mysql-db01 ~]# /etc/init.d/mysqld stopShutting down MySQL..... SUCCESS!

(3)查看mysql-db03上的MySQL从库同步状态

 
[root@mysql-db03 ~]# mysql -uroot -p123123 -e 'show slave status\G'Warning: Using a password on the command line interface can be insecure.*************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.0.52 #现在的主库IP Master_User: rep Master_Port: 3306 Connect_Retry: 60 Master_Log_File: mysql-bin.000006 #binlog日志 Read_Master_Log_Pos: 777 #binlog日志位置 Relay_Log_File: mysql-db03-relay-bin.000002 Relay_Log_Pos: 408 Relay_Master_Log_File: mysql-bin.000006 Slave_IO_Running: Yes Slave_SQL_Running: Yes             ..以下省略若干内容..

(4)查看mysql-db02上的MySQL,主库同步状态。

【干货】Mha-Atlas-MySQL高可用方案实践(一)

(5)查看mysql-db03上的mha进程状态

 
[root@mysql-db03 ~]# ps -ef | grep perl | grep -v grep #查询发现mha进程已经没了[root@mysql-db03 ~]#

(6)查看mha配置文件信息

【干货】Mha-Atlas-MySQL高可用方案实践(一)

说明:
当作为主库的mysql-db01上的MySQL宕机以后,mha通过检测发现mysql-db01宕机,那么会将binlog日志最全的从库立刻提升为主库,而其他的从库会指向新的主库进行再次同步。

此处需要进行简单的mha日志记录的讲解:/var/log/mha/mha1/manager

5.7 进行mha的故障还原测试

由于mysql-db01的MySQL服务宕机,因此mha将mysql-db02提升为了主库。因此,我们需要将宕机的mysql-db01的MySQL服务启动,然后作为主库mysql-db02的从库。

初始状态:

【干货】Mha-Atlas-MySQL高可用方案实践(一)

(1)将故障宕机的mysql-db01的MySQL服务启动并授权进行从同步

 
[root@mysql-db01 ~]# /etc/init.d/mysqld startStarting MySQL. SUCCESS! [root@mysql-db01 ~]# mysql -uroot -p123123mysql> CHANGE MASTER TO MASTER_HOST='192.168.0.52', MASTER_PORT=3306, MASTER_AUTO_POSITION=1, MASTER_USER='rep', MASTER_PASSWORD='123123';mysql> start slave;mysql> show slave status\G          #查看从同步状态

(2)将mha配置文件里缺失的部分补全

[root@mysql-db03 ~]# cat /etc/mha/mha1.cnf [server default]manager_log=/var/log/mha/mha1/managermanager_workdir=/var/log/mha/mha1master_binlog_dir=/usr/local/mysql/datapassword=123123ping_interval=2repl_password=123123repl_user=repssh_user=rootuser=mha
[server1]hostname=192.168.0.51port=3306
[server2]hostname=12.168.0.52port=3306
[server3]hostname=192.168.0.53port=3306

(3)启动mha进程

[root@mysql-db03 ~]# nohup masterha_manager --conf=/etc/mha/mha1.cnf --remove_dead_master_conf --ignore_last_failover < /dev/null > /var/log/mha/mha1/manager.log 2>&1 &[root@mysql-db03 ~]# ps -ef | grep perl | grep -v greproot       5226   4690  0 09:42 pts/2    00:00:00 perl /usr/bin/masterha_manager --conf=/etc/mha/mha1.cnf --remove_dead_master_conf --ignore_last_failover

(4)停掉mysql-db02上的MySQL服务

[root@mysql-db02 ~]# /etc/init.d/mysqld stopShutting down MySQL..... SUCCESS!

(5)查看mysql-db03上的主从同步状态:

 
[root@mysql-db03 ~]# mysql -uroot -p123123 -e 'show slave status\G'Warning: Using a password on the command line interface can be insecure.*************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.0.51 #此时的主库IP切换回了mysql-db01 Master_User: rep Master_Port: 3306 Connect_Retry: 60 Master_Log_File: mysql-bin.000005 Read_Master_Log_Pos: 231 Relay_Log_File: mysql-db03-relay-bin.000002 Relay_Log_Pos: 361 Relay_Master_Log_File: mysql-bin.000005 Slave_IO_Running: Yes Slave_SQL_Running: Yes            ..以下省略若干行..

(6)启动mysql-db02上的MySQL服务

 
[root@mysql-db02 ~]# /etc/init.d/mysqld startStarting MySQL. SUCCESS! [root@mysql-db02 ~]# mysql -uroot -p123123mysql> CHANGE MASTER TO MASTER_HOST='192.168.0.51', MASTER_PORT=3306, MASTER_AUTO_POSITION=1, MASTER_USER='rep', MASTER_PASSWORD='123123';mysql> start slave;mysql> show slave status\G          #查看从同步状态

(7)再次补全mha配置文件后,启动mha进程

 
[root@mysql-db03 ~]# cat /etc/mha/mha1.cnf [server default]manager_log=/var/log/mha/mha1/managermanager_workdir=/var/log/mha/mha1master_binlog_dir=/usr/local/mysql/datapassword=123123ping_interval=2repl_password=123123repl_user=repssh_user=rootuser=mha
[server1]hostname=192.168.0.51port=3306
[server2]hostname=192.168.0.52port=3306
[server3]#andidate_master=1#check_repl_delay=0hostname=192.168.0.53port=3306[root@mysql-db03 ~]# nohup masterha_manager --conf=/etc/mha/mha1.cnf --remove_dead_master_conf --ignore_last_failover < /dev/null > /var/log/mha/mha1/manager.log 2>&1 &[root@mysql-db03 ~]# ps -ef | grep perl | grep -v greproot       5226   4690  0 09:42 pts/2    00:00:01 perl /usr/bin/masterha_manager --conf=/etc/mha/mha1.cnf --remove_dead_master_conf --ignore_last_failover

此时的初始状态还原为下图:

【干货】Mha-Atlas-MySQL高可用方案实践(一)

六,MHA参数验证实践(同学们操作)

【干货】Mha-Atlas-MySQL高可用方案实践(一)

mha配置文件内容如下:

附录:源码安装mha的方法

node节点的源码安装方法:

 
[root@mysql-db01 ~]# yum -y install perl-DBD-MySQL perl-Config-Tiny perl-Params-Validate perl-CPAN perl-devel perl-ExtUtils-CBuilder perl-ExtUtils-MakeMaker
[root@mysql-db01 ~]# tar xf mha4mysql-node-0.56.tar.gz -C /usr/src/
[root@mysql-db01 ~]# cd /usr/src/mha4mysql-node-0.56/
[root@mysql-db01 mha4mysql-node-0.56]# perl Makefile.PL
[root@mysql-db01 mha4mysql-node-0.56]# make && make install
manager节点的源码安装方法:[root@mysql-db01 ~]# yum -y install perl-DBD-MySQL perl-Config-Tiny perl-Params-Validate perl-CPAN perl-devel perl-ExtUtils-CBuilder perl-ExtUtils-MakeMaker
[root@mysql-db01 ~]# tar xf mha4mysql-manager-0.56.tar.gz -C /usr/src/
[root@mysql-db01 ~]# cd /usr/src/mha4mysql-manager-0.56/
[root@mysql-db01 mha4mysql-manager-0.56]# perl Makefile.PL
[root@mysql-db01 mha4mysql-manager-0.56]# make && make install

--End--

推荐阅读
















https://www.cnblogs.com/chensiqiqi/p/9162959.html