老张最近两天有些忙,一些老铁一直问,啥时更新博文,我可能做不到天天更新啊,但保证以后一有空就写一些干货知识分享给大家。
我们如果想要做好技术这项工作,一定要做到理论与实践先结合。我一个曾经被数据库虐得体无完肤的过来人给大家一些建议;就是只看书,背理论真的行不通,到时遇到棘手的问题,你还是一样抓瞎。一定要在理论理清的基础上多做实验。
给自己定个目标,3个月做够100-500个实验。然后整理在做实验过程中的各种报错,认真解读分析报错原理,做好笔记。最后再拿起书,重新阅读之前有些可能理解不了的理论知识,我想这次读书的过程,肯定比之前收益会更大。
之前答应过大家,给大家介绍MySQL高可用架构三部曲。今儿先给大家介绍第一步曲;MySQL高可用最主流的架构---MHA
MHA高可用架构是基于主从复制原理而部署的,是最常见,最主流的架构。
MHA简介:
MHA,即MasterHigh Availability Manager and Toolsfor MySQL,是日本的一位MySQL专家采用Perl语言编写的一个脚本管理工具,该工具仅适用于MySQLReplication 环境,目的在于维持Master主库的高可用性。
MHA(Master High Availability)是自动的master故障转移和Slave提升的软件包.它是基于标准的MySQL复制(异步/半同步).
MHA组成部分:
MHA由两部分组成:
MHA Manager(管理节点)
MHA Node(数据节点)
MHA部署解读:
MHA Manager可以单独部署在一台独立机器上管理多个master-slave集群,也可以部署在一台slave上.MHA Manager探测集群的node节点,当发现master出现故障的时候,它可以自动将具有最新数据的slave提升为新的master,然后将所有其它的slave导向新的master上.整个故障转移过程对应用程序是透明的。
MHA node运行在每台MySQL服务器上(master/slave/manager),它通过监控具备解析和清理logs功能的脚本来加快故障转移的。
MHA优缺点介绍:
优点:
1. 故障切换时,可以自行判断哪个从库与主库的数据最接近,就切换到上面,可以减少数据的丢失,保证数据的一致性
2. 支持 binlog server,可提高 binlog 传送效率,进一步减少数据丢失风险。
3. 可以配置 mysql 5.7 的增强半同步,来保证数据的时时同步
缺点:
1. 自动切换的脚本太简单了,而且比较老化,建议后期逐渐完善。
2. 搭建 MHA 架构,需要开启 linux 系统互信协议,所以对于系统安全性来说,是个不小的考验。
原理介绍:
MHA的目的在于维持MySQL Replication中Master库的高可用性,其最大特点是可以修复多个Slave之间的差异日志,最终使所有Slave保持数据一致,然后从中选择一个充当新的Master,并将其它Slave指向它。
当master出现故障时,可以通过对比slave之间I/O thread 读取主库binlog的position号,选取最接近的slave做为备选主库(备胎)。其它的从库可以通过与备选主库对比生成差异的中继日志。在备选主库上应用从原来master保存的binlog,同时将备选主库提升为master。最后在其它slave上应用相应的差异中继日志并开始从新的master开始复制。
MHA工具包功能介绍:
Manager工具:
#masterha_check_ssh : 检查MHA的SSH配置。 #masterha_check_repl : 检查MySQL复制。 #masterha_manager : 启动MHA。 #masterha_check_status : 检测当前MHA运行状态。 #masterha_master_monitor : 监测master是否宕机。 #masterha_master_switch : 控制故障转移(自动或手动)。 #masterha_conf_host : 添加或删除配置的server信息。
2. Node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并应用于其它slave。 #filter_mysqlbinlog : 去除不必要的ROLLBACK事件(MHA已不再使用这个工具)。 #purge_relay_logs : 清除中继日志(不会阻塞SQL线程)。
实战图表展示:
实战开始:
环境介绍:
192.168.56.100 master node 192.168.56.101 slave1 node 192.168.56.102 slave2(manager,node) vip 192.168.56.123
第一步操作:生成ssh无密钥证书
主库(100)执行生成密钥操作:
ssh-keygen -t dsa -P '' -f id_dsa Id_dsa.pub为公钥,id_dsa为私钥,紧接着将公钥文件复制成authorized_keys文件, 这个步骤是必须的,过程如下: cat id_dsa.pub >> authorized_keys
从库(101)执行生成密钥操作:
ssh-keygen -t dsa -P '' -f id_dsa cat id_dsa.pub >> authorized_keys
管理节点(102)生成密钥过程:
ssh-keygen -t dsa -P '' -f id_dsa cat id_dsa.pub >> authorized_keys
主库(100)执行接收密钥的过程:
scp 192.168.56.101:/root/.ssh/id_dsa.pub ./id_dsa.pub.101 scp 192.168.56.102:/root/.ssh/id_dsa.pub ./id_dsa.pub.102
合并密钥: cat id_dsa.pub.101 >> authorized_keys cat id_dsa.pub.102 >> authorized_keys
在主库上传送合成密钥:
scp authorized_keys 192.168.56.101:/root/.ssh/ scp authorized_keys 192.168.56.102:/root/.ssh/
在三台服务器上,编辑/etc/hosts文件,分别加入三台主机hostname
vim /etc/hosts node1 192.168.56.100 node2 192.168.56.101 node3 192.168.56.102
验证主机名登陆,密钥验证:
分别在三台机器上执行:
在192.168.56.100执行 ssh node2 ssh node3 在192.168.56.101执行 ssh node1 ssh node3 在192.168.56.102执行 ssh node1 ssh node2
搭建主从环境(一主两从架构,我这里是mysql 5.7版本)
在所有节点上都要执行
创建主从账号: grant replication slave on *.* to 'repl'@'192.168.56.%' identified by 'repl'; flush privileges;
创建管理账号: grant all privileges on *.* to 'zs'@'192.168.56.%' identified by '123456'; flush privileges;
在Master(100)上面安装数据节点:
首先要先安装mysql依赖的perl环境
yum install perl-DBD-MySQL
解压数据节点的包
tar -zxvf mha4mysql-node-0.57.tar.gz
安装perl-cpan软件包
cd mha4mysql-node-0.57 yum -y install perl-CPAN* perl Makefile.PL make && make install
在从库(101)上面安装数据节点:
同主库一样的安装操作;
在manager管理节点(102)上面安装管理节点:
首先先要安装环境需要的介质包
yum install -y perl-DBD-MySQL* rpm -ivh perl-Params-Validate-0.92-3.el6.x86_64.rpm rpm -ivh perl-Config-Tiny-2.12-1.el6.rfx.noarch.rpm rpm -ivh perl-Log-Dispatch-2.26-1.el6.rf.noarch.rpm rpm -ivh perl-Parallel-ForkManager-0.7.5-2.2.el6.rf.noarch.rpm
下载地址:https://centos.pkgs.org/
再安装数据节点:
tar -zxvf mha4mysql-node-0.57.tar.gz cd mha4mysql-node-0.57 yum -y install perl-CPAN* perl Makefile.PL make && make install
最后安装管理节点:
tar -zxvf mha4mysql-manager-0.57.tar.gz perl Makefile.PL make make install
环境配置,基础安装操作完成。
下面来进行管理节点MHA配置:
创建mha家目录,编辑启动配置文件
mkdir -p /usr/local/mha mkdir -p /etc/mhacd /etc/mha/
编辑MHA配置文件 vim /etc/mha/mha.conf [server default] user=zs password=123456 manager_workdir=/usr/local/mha manager_log=/usr/local/mha/manager.log remote_workdir=/usr/local/mha ssh_user=root repl_user=repl repl_password=repl ping_interval=1 master_ip_failover_script=/usr/local/scripts/master_ip_failover master_ip_online_change_script=/usr/local/scripts/master_ip_online_change
[server1] hostname=192.168.56.100 ssh_port=22 master_binlog_dir=/data/mysql candidate_master=1 port=3306
[server2] hostname=192.168.56.101 ssh_port=22 master_binlog_dir=/data/mysql candidate_master=1 port=3306
[server3] hostname=192.168.56.102 ssh_port=22 master_binlog_dir=/data/mysql no_master=1 port=3306
创建failover,online 脚本的目录
mkdir -p /usr/local/scripts
编辑failover切换脚本:
vim master_ip_failover #!/usr/bin/env perl use strict; use warnings FATAL => 'all'; use Getopt::Long; my ( $command, $ssh_user, $orig_master_host, $orig_master_ip, $orig_master_port, $new_master_host, $new_master_ip, $new_master_port ); my $vip = '192.168.56.123/24'; my $key = '0'; my $ssh_start_vip = "/sbin/ifconfig eth0:$key $vip"; my $ssh_stop_vip = "/sbin/ifconfig eth0:$key down"; GetOptions( 'command=s' => \$command, 'ssh_user=s' => \$ssh_user, 'orig_master_host=s' => \$orig_master_host, 'orig_master_ip=s' => \$orig_master_ip, 'orig_master_port=i' => \$orig_master_port, 'new_master_host=s' => \$new_master_host, 'new_master_ip=s' => \$new_master_ip, 'new_master_port=i' => \$new_master_port, ); exit &main();
sub main { print "\n\nIN SCRIPT TEST====$ssh_stop_vip==$ssh_start_vip===\n\n"; if ( $command eq "stop" || $command eq "stopssh" ) { my $exit_code = 1; eval { print "Disabling the VIP on old master: $orig_master_host \n"; &stop_vip(); $exit_code = 0; }; if ($@) { warn "Got Error: $@\n"; exit $exit_code; } exit $exit_code; } elsif ( $command eq "start" ) { my $exit_code = 10; eval { print "Enabling the VIP - $vip on the new master - $new_master_host \n"; &start_vip(); $exit_code = 0; }; if ($@) { warn $@; exit $exit_code; } exit $exit_code; } elsif ( $command eq "status" ) { print "Checking the Status of the script.. OK \n"; exit 0; } else { &usage(); exit 1; } }
sub start_vip() { `ssh $ssh_user\@$new_master_host \" $ssh_start_vip \"`; } sub stop_vip() { return 0 unless ($ssh_user); `ssh $ssh_user\@$orig_master_host \" $ssh_stop_vip \"`; } sub usage { print "Usage: master_ip_failover --command=start|stop|stopssh|status --orig_master_host=host --orig_master_ip=ip --orig_master_port=port --new_master_host=host --new_master_ip=ip --new_master_port=port\n"; }
编辑online_change的脚本:
cd /usr/local/scripts/
vim master_ip_online_change #!/usr/bin/env perl use strict; use warnings FATAL =>'all'; use Getopt::Long; my $vip = '192.168.1.123/24'; # Virtual IP my $key = "0"; my $ssh_start_vip = "/sbin/ifconfig eth0:$key $vip"; my $ssh_stop_vip = "/sbin/ifconfig eth0:$key down"; my $exit_code = 0; my ( $command, $orig_master_is_new_slave, $orig_master_host, $orig_master_ip, $orig_master_port, $orig_master_user, $orig_master_password, $orig_master_ssh_user, $new_master_host, $new_master_ip, $new_master_port, $new_master_user, $new_master_password, $new_master_ssh_user, ); GetOptions( 'command=s' => \$command, 'orig_master_is_new_slave' => \$orig_master_is_new_slave, 'orig_master_host=s' => \$orig_master_host, 'orig_master_ip=s' => \$orig_master_ip, 'orig_master_port=i' => \$orig_master_port, 'orig_master_user=s' => \$orig_master_user, 'orig_master_password=s' => \$orig_master_password, 'orig_master_ssh_user=s' => \$orig_master_ssh_user, 'new_master_host=s' => \$new_master_host, 'new_master_ip=s' => \$new_master_ip, 'new_master_port=i' => \$new_master_port, 'new_master_user=s' => \$new_master_user, 'new_master_password=s' => \$new_master_password, 'new_master_ssh_user=s' => \$new_master_ssh_user, ); exit &main();
sub main { #print "\n\nIN SCRIPT TEST====$ssh_stop_vip==$ssh_start_vip===\n\n"; if ( $command eq "stop" || $command eq "stopssh" ) { # $orig_master_host, $orig_master_ip, $orig_master_port are passed. # If you manage master ip address at global catalog database, # invalidate orig_master_ip here. my $exit_code = 1; eval { print "\n\n\n***************************************************************\n"; print "Disabling the VIP - $vip on old master: $orig_master_host\n"; print "***************************************************************\n\n\n\n"; &stop_vip(); $exit_code = 0; }; if ($@) { warn "Got Error: $@\n"; exit $exit_code; } exit $exit_code; } elsif ( $command eq "start" ) { # all arguments are passed. # If you manage master ip address at global catalog database, # activate new_master_ip here. # You can also grant write access (create user, set read_only=0, etc) here. my $exit_code = 10; eval { print "\n\n\n***************************************************************\n"; print "Enabling the VIP - $vip on new master: $new_master_host \n"; print "***************************************************************\n\n\n\n"; &start_vip(); $exit_code = 0; }; if ($@) { warn $@; exit $exit_code; } exit $exit_code; }
elsif ( $command eq "status" ) { print "Checking the Status of the script.. OK \n"; `ssh $orig_master_ssh_user\@$orig_master_host \" $ssh_start_vip \"`; exit 0; } else { &usage(); exit 1; } } # A simple system call that enable the VIP on the new master sub start_vip() { `ssh $new_master_ssh_user\@$new_master_host \" $ssh_start_vip \"`; } # A simple system call that disable the VIP on the old_master sub stop_vip() { `ssh $orig_master_ssh_user\@$orig_master_host \" $ssh_stop_vip \"`; } sub usage { print "Usage: master_ip_failover –command=start|stop|stopssh|status –orig_master_host=host –orig_master_ip=ip –orig_master_port=po rt –new_master_host=host –new_master_ip=ip –new_master_port=port\n"; }
创建完两个脚本,记得赋予执行权限
利用mha工具检测ssh
安装需要的环境包:
yum -y install perl-Time-HiRes
执行检测命令;
/usr/local/bin/masterha_check_ssh --conf=/etc/mha/mha.conf
检测结果显示:都为OK,代表ssh检测成功
结果展示:
Sun Jul 23 09:39:09 2017 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping. Sun Jul 23 09:39:09 2017 - [info] Reading application default configuration from /etc/mha/mha.conf.. Sun Jul 23 09:39:09 2017 - [info] Reading server configuration from /etc/mha/mha.conf.. Sun Jul 23 09:39:09 2017 - [info] Starting SSH connection tests.. Sun Jul 23 09:39:11 2017 - [debug] Sun Jul 23 09:39:09 2017 - [debug] Connecting via SSH from root@192.168.56.100(192.168.56.100:22) to root@192.168.56.101(192.168.56.101:22).. Sun Jul 23 09:39:10 2017 - [debug] ok. Sun Jul 23 09:39:10 2017 - [debug] Connecting via SSH from root@192.168.56.100(192.168.56.100:22) to root@192.168.56.102(192.168.56.102:22).. Sun Jul 23 09:39:11 2017 - [debug] ok. Sun Jul 23 09:39:11 2017 - [debug] Sun Jul 23 09:39:10 2017 - [debug] Connecting via SSH from root@192.168.56.101(192.168.56.101:22) to root@192.168.56.100(192.168.56.100:22).. Sun Jul 23 09:39:10 2017 - [debug] ok. Sun Jul 23 09:39:10 2017 - [debug] Connecting via SSH from root@192.168.56.101(192.168.56.101:22) to root@192.168.56.102(192.168.56.102:22).. Sun Jul 23 09:39:11 2017 - [debug] ok. Sun Jul 23 09:39:11 2017 - [debug] Sun Jul 23 09:39:10 2017 - [debug] Connecting via SSH from root@192.168.56.102(192.168.56.102:22) to root@192.168.56.100(192.168.56.100:22).. Warning: Permanently added '192.168.56.102' (RSA) to the list of known hosts. Sun Jul 23 09:39:11 2017 - [debug] ok. Sun Jul 23 09:39:11 2017 - [debug] Connecting via SSH from root@192.168.56.102(192.168.56.102:22) to root@192.168.56.101(192.168.56.101:22).. Sun Jul 23 09:39:11 2017 - [debug] ok. Sun Jul 23 09:39:11 2017 - [info] All SSH connection tests passed successfully. 检测masterha_check_repl --conf=/etc/mha/mha.conf Sun Jul 23 10:14:11 2017 - [warning] Global configuration file /etc/masterha_default.cnf not found. Skipping. Sun Jul 23 10:14:11 2017 - [info] Reading application default configuration from /etc/mha/mha.conf.. Sun Jul 23 10:14:11 2017 - [info] Reading server configuration from /etc/mha/mha.conf.. Sun Jul 23 10:14:11 2017 - [info] MHA::MasterMonitor version 0.57. Sun Jul 23 10:14:11 2017 - [info] Multi-master configuration is detected. Current primary(writable) master is 192.168.56.100(192.168.56.100:3306) Sun Jul 23 10:14:11 2017 - [info] Master configurations are as below: Master 192.168.56.100(192.168.56.100:3306), replicating from 192.168.56.101(192.168.56.101:3306) Master 192.168.56.101(192.168.56.101:3306), replicating from 192.168.56.100(192.168.56.100:3306), read-only Sun Jul 23 10:14:11 2017 - [info] GTID failover mode = 1 Sun Jul 23 10:14:11 2017 - [info] Dead Servers: Sun Jul 23 10:14:11 2017 - [info] Alive Servers: Sun Jul 23 10:14:11 2017 - [info] 192.168.56.100(192.168.56.100:3306) Sun Jul 23 10:14:11 2017 - [info] 192.168.56.101(192.168.56.101:3306) Sun Jul 23 10:14:11 2017 - [info] 192.168.56.102(192.168.56.102:3306) Sun Jul 23 10:14:11 2017 - [info] Alive Slaves: Sun Jul 23 10:14:11 2017 - [info] 192.168.56.101(192.168.56.101:3306) Version=5.7.14-log (oldest major version between slaves) log-bin:enabled Sun Jul 23 10:14:11 2017 - [info] GTID ON Sun Jul 23 10:14:11 2017 - [info] Replicating from 192.168.56.100(192.168.56.100:3306) Sun Jul 23 10:14:11 2017 - [info] Primary candidate for the new Master (candidate_master is set) Sun Jul 23 10:14:11 2017 - [info] 192.168.56.102(192.168.56.102:3306) Version=5.7.14-log (oldest major version between slaves) log-bin:enabled Sun Jul 23 10:14:11 2017 - [info] GTID ON Sun Jul 23 10:14:11 2017 - [info] Replicating from 192.168.56.100(192.168.56.100:3306) Sun Jul 23 10:14:11 2017 - [info] Not candidate for the new Master (no_master is set) Sun Jul 23 10:14:11 2017 - [info] Current Alive Master: 192.168.56.100(192.168.56.100:3306) Sun Jul 23 10:14:11 2017 - [info] Checking slave configurations.. Sun Jul 23 10:14:11 2017 - [info] Checking replication filtering settings.. Sun Jul 23 10:14:11 2017 - [info] binlog_do_db= , binlog_ignore_db= Sun Jul 23 10:14:11 2017 - [info] Replication filtering check ok. Sun Jul 23 10:14:11 2017 - [info] GTID (with auto-pos) is supported. Skipping all SSH and Node package checking. Sun Jul 23 10:14:11 2017 - [info] Checking SSH publickey authentication settings on the current master.. Sun Jul 23 10:14:11 2017 - [info] HealthCheck: SSH to 192.168.56.100 is reachable. Sun Jul 23 10:14:11 2017 - [info] 192.168.56.100(192.168.56.100:3306) (current master) +--192.168.56.101(192.168.56.101:3306) +--192.168.56.102(192.168.56.102:3306) Sun Jul 23 10:14:11 2017 - [info] Checking replication health on 192.168.56.101.. Sun Jul 23 10:14:11 2017 - [info] ok. Sun Jul 23 10:14:11 2017 - [info] Checking replication health on 192.168.56.102.. Sun Jul 23 10:14:11 2017 - [info] ok. Sun Jul 23 10:14:11 2017 - [info] Checking master_ip_failover_script status: Sun Jul 23 10:14:11 2017 - [info] /usr/local/scripts/master_ip_failover --command=status --ssh_user=root --orig_master_host=192.168.56.100 --orig_master_ip=192.168.56.100 --orig_master_port=3306 IN SCRIPT TEST====/sbin/ifconfig eth0:0 down==/sbin/ifconfig eth0:0 192.168.56.123/24=== Checking the Status of the script.. OK Sun Jul 23 10:14:11 2017 - [info] OK. Sun Jul 23 10:14:11 2017 - [warning] shutdown_script is not defined. Sun Jul 23 10:14:11 2017 - [info] Got exit code 0 (Not master dead). MySQL Replication Health is OK.
在主库(100)执行添加vip的过程:(第一次手动添加)
ip addr add 192.168.56.123 dev eth0
在管理节点(102)上,执行mha的启动
nohup masterha_manager --conf=/etc/mha/mha.conf > /tmp/mha_manager.log < /dev/null 2>&1 &
验证启动成功的命令:查看显示状态
masterha_check_status --conf=/etc/mha/mha.conf
模拟主库故障,查看是否自动切换:
在主库(100)上面执行停掉mysql操作。
mysqladmin -uroot -proot123 shutdown 从库192.168.56.101自动获取vip且转换为新的主库,192.168.56.102自动指向新的主库。
切换后,MHA进程会自动停止运行
在管理节点查看:
masterha_check_status --conf=/etc/mha/mha.conf mha is stopped(2:NOT_RUNNING).
恢复操作:
把宕掉的主库192.168.56.100恢复起来;
/usr/local/mysql/bin/mysqld_safe --defaults-file=/etc/my.cnf &
验证结果:重新指向现在的主库192.168.56.101
主从状态一切ok!
MHA搭建演练结束
希望大家每天抽出一点时间,看看书,看看报,看看老张的技术博客哈,我们可能比不了那些富二代,衣来伸手饭来张口,钱可以肆意挥霍!但这样日子很空虚,没有一个自己的奋斗方向!我们还得靠自己的双手努力学习,既然从事了相关技术岗位的工作,那就多做实验,反复练习。经验和技巧都是在反复推敲中获得的!
今年老张打算多出点对老铁们有用的干货技术博客;
大家相互学习,相互请教,争取明年大家的工资翻倍;
企业中MySQL高可用主流架构实战三部曲的第一部曲MHA(完结)
免责声明:本站发布的内容(图片、视频和文字)以原创、转载和分享为主,文章观点不代表本网站立场,如果涉及侵权请联系站长邮箱:is@yisu.com进行举报,并提供相关证据,一经查实,将立刻删除涉嫌侵权内容。