【原】ubuntu下Mysql的HA（corosync+pacemaker+drbd）

一、前提准备：

1、OS：ubuntu 12.04

2、cat /etc/hosts：

 127.0.0.1 localhost
 192.168.153.154 ha1
 192.168.153.155 ha2

3、硬盘：/dev/sdb1

4、vip：192.168.153.200

二、corosync和pacemaker的安装配置：

1、[ha1+ha2]：

 apt-get install -y pacemaker corosync

2、[ha1]：

1 corosync-keygen（时间较长）
2 scp /etc/corosync/authkey root@ha2:/etc/corosync/

3、[ha1+ha2]：取消corosync和pacemaker的开机启动：

 apt-get install -y sysv-rc-conf
 sysv-rc-conf（使用空格把drbd,pacemaker,corosync有X的地方取消掉，q退出）
 sed -i 's/127.0.0.1/192.168.153.0/g' /etc/corosync/corosync.conf
 sed -i 's/ver:       0/ver:       1/g' /etc/corosync/corosync.conf
 mkdir /var/log/cluster
 update-rc.d pacemaker start       . stop    .
 sed -i 's/no/yes/g' /etc/default/corosync
 service corosync start
 corosync-cfgtool -s
 corosync-quorumtool -l
 service pacemaker start
 验证:crm status

 root@ha1:~# crm status
 ============
 Last updated: Tue Feb  ::
 Last change: Tue Feb  ::  via crmd on ha1
 Stack: openais
 Current DC: ha1 - partition with quorum
 Version: 1.1.-9971ebba4494012a93c03b40a2c58ec0eb60f50c
  Nodes configured,  expected votes
  Resources configured.
 ============

 Online: [ ha1 ha2 ]

三、drbd的安装配置：

1、[ha1+ha2]：drbd的安装：

 tar -zxvf drbd8_8.4.4.orig.tar.gz（由于drbd8..2有bug，所以这里选择手动安装8..4版本）
 apt-get install -y debhelper flex xsltproc
 cd drbd8_8.4.4
 ./configure --localstatedir=/var --sysconfdir=/etc --with-km
 make && make install

2、[ha1+ha2]：创建磁盘：

 fdisk /dev/sdb
 n
 p

 ...
 w

3、[ha1+ha2]：创建drbd的mysql资源：

 cat /etc/drbd.d/mysql.res
 resource mysql {
         device /dev/drbd0;
         disk /dev/sdb1;
         meta-disk internal;
         on ha1 {
                 address 192.168.153.154:;
         }
         on ha2 {
                 address 192.168.153.155:;
         }
         syncer {
                 rate 10M;
         }
 }
 echo -e 'drbd # -> added' >> /etc/modules
 drbdadm dump mysql
 drbdadm create-md mysql
 drbdadm up mysql
 验证：drbd-overview
 :mysql/  Connected Secondary/Secondary Inconsistent/Inconsistent C r----- 

4、[ha1]：

 drbdadm -- --overwrite-data-of-peer primary mysql
 cat /proc/drbd：
 root@ha1:~# cat /proc/drbd
 version: 8.4. (api:/proto:-)
 GIT-hash: 599f286440bd633d15d5ff985204aff4bccffadd build by root@ha1, -- ::
   : cs:SyncSource ro:Primary/Secondary ds:UpToDate/Inconsistent C r-----
     ns: nr: dw: dr: al: bm: lo: pe: ua: ap: ep: wo:f oos:
 [>....................] sync'ed:  0.6% (20372/20476)M
 finish: :: speed: , (,) K/sec
 直到同步100%的时候才成功。结果是：
 root@ha1:~# cat /proc/drbd
 version: 8.4. (api:/proto:-)
 GIT-hash: 599f286440bd633d15d5ff985204aff4bccffadd build by root@ha1, -- ::
   : cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r-----
     ns: nr: dw: dr: al: bm: lo: pe: ua: ap: ep: wo:f oos:
 service drbd restart
 mkfs.ext4 /dev/drbd0
 mkdir /mydata
 mount /dev/drbd0 /mydata
 mkdir -p /mydata/testfile
 umount /dev/drbd0
 drbdadm secondary mysql

5、[ha2]：

 service drbd restart
 drbdadm primary mysql
  mkdir /mydata
 mount /dev/drbd0 /mydata
 ll /mydata/ 如果可以看到testfile说明成功。
 umount /dev/drbd0
 drbdadm secondary mysql（如果中间出现错误，可以把ha1和ha2的drbd重启即可。）

6、[ha1]:

 drbdadm primary mysql

7、[ha1+ha2]：mysql的安装配置

 apt-get install -y mysql-server
 sed -i 's/127.0.0.1/0.0.0.0/g' /etc/mysql/my.cnf
 service mysql restart
 mysql -uroot -p
 grant all privileges on *.* to root@"%" identified by "password" with grant option;
 FLUSH PRIVILEGES;
 delete from mysql.user where user="";
 quit;
 service mysql stop

8、[ha1]:

 drbdadm primary mysql
 drbd-overview
 mount /dev/drbd0 /mydata
 cp -R -p /var/lib/mysql /mydata/

9、[ha1+ha2]:

 sed -i 's/2345/!0123456/g' /etc/init/mysql.conf
 sed -i -e '/datadir/cdatadir         = \/mydata\/mysql'  /etc/mysql/my.cnf
 sed -i '/\/var\/lib\/mysql\/ /c\/mydata\/mysql\/ r,' /etc/apparmor.d/usr.sbin.mysqld
 sed -i '/\/var\/lib\/mysql\/\*\* rwk/c\/mydata\/mysql\/** rwk,' /etc/apparmor.d/usr.sbin.mysqld

10、[ha1]:

 service mysql start
 mysql -uroot -p
 CREATE DATABASE test1;
 quit
 service mysql stop
 umount /dev/drbd0
 drbdadm secondary mysql

11、[ha2]:

 drbdadm primary mysql
 mount /dev/drbd0 /mydata/
 service mysql start
 查看是否有数据库test1
 service mysql stop
 umount /dev/drbd0
 drbdadm secondary mysql

12、[ha1]:

 drbdadm primary mysql

13、[ha1+ha2]:

 service drbd stop

14、[ha1]:

 crm configure property stonith-enabled=false
 crm configure property no-quorum-policy=ignore

 crm configure
 primitive mysqldrbd ocf:heartbeat:drbd params drbd_resource=mysql op start timeout= op stop timeout= op monitor role=Master interval= timeout= op monitor role=Slave interval= timeout=
 ms ms_mysqldrbd mysqldrbd meta master-max= master-node-max= clone-max= clone-node-max= notify=true
 commit
 exit
 验证:crm status
 root@ha1:~# crm status
 ============
 Last updated: Tue Feb  ::
 Last change: Tue Feb  ::  via cibadmin on ha1
 Stack: openais
 Current DC: ha1 - partition with quorum
 Version: 1.1.-9971ebba4494012a93c03b40a2c58ec0eb60f50c
  Nodes configured,  expected votes
  Resources configured.
 ============

 Online: [ ha1 ha2 ]

   Master/Slave Set: ms_mysqldrbd [mysqldrbd]
      Masters: [ ha1 ]
       Slaves: [ ha2 ]

 crm configure
 primitive mystore ocf:heartbeat:Filesystem params device=/dev/drbd0 directory=/mydata fstype=ext4 op start timeout= op stop timeout=
 colocation mystore_with_ms_mysqldrbd inf: mystore ms_mysqldrbd:Master
 order mystore_after_ms_mysqldrbd mandatory: ms_mysqldrbd:promote mystore:start
 commit
 exit
 验证:crm status,mount
   Master/Slave Set: ms_mysqldrbd [mysqldrbd]
       Masters: [ ha1 ]
       Slaves: [ ha2 ]
   mystore (ocf::heartbeat:Filesystem): Started ha1

 crm configure
 primitive mysql lsb:mysql
 colocation mysql_with_mystore inf: mysql mystore
 order mysql_after_mystore mandatory: mystore mysql
 commit
 exit
 验证:crm status
 Master/Slave Set: ms_mysqldrbd [mysqldrbd]
       Masters: [ ha1 ]
       Slaves: [ ha2 ]
   mystore (ocf::heartbeat:Filesystem): Started ha1
   mysql (lsb:mysql): Started ha1

 crm configure
 primitive vip ocf:heartbeat:IPaddr params ip=192.168.153.200 nic=eth0 cidr_netmask=255.255.255.0
 colocation vip_with_ms_mysqldrbd inf: ms_mysqldrbd:Master vip
 commit
 exit
 验证crm status
 Master/Slave Set: ms_mysqldrbd [mysqldrbd]
       Masters: [ ha1 ]
       Slaves: [ ha2 ]
   mystore (ocf::heartbeat:Filesystem): Started ha1
   mysql (lsb:mysql): Started ha1
   vip (ocf::heartbeat:IPaddr): Started ha1

四、验证：

1、客户端mysql验证：

 mysql -uroot -p -h 192.168.153.200
 mysql> show databases;
 +--------------------+
 | Database           |
 +--------------------+
 | information_schema |
 | mysql              |
 | performance_schema |
 | test               |
 | test1              |
 +--------------------+

2、模拟故障：

 root@ha1:~# crm
 crm(live)# node
 crm(live)node# standby
 crm(live)node# show
 ha1: normal
 standby: on
 ha2: normal
 crm(live)node# exit
 bye
 此时看一下状态：
 root@ha1:~# crm status
 Node ha1: standby
 Online: [ ha2 ]

  Master/Slave Set: ms_mysqldrbd [mysqldrbd]
      Masters: [ ha2 ]
      Stopped: [ mysqldrbd: ]
  mystore (ocf::heartbeat:Filesystem): Started ha2
  mysql (lsb:mysql): Started ha2
  vip (ocf::heartbeat:IPaddr): Started ha2

此时，mysql已经运行于ha2上了。