innobackupex 备份数据搭建 MySQL Slave

简介：

数据量比较大时，使用 innobackupex 备份数据新增 MySQL Slave 节点。

安装 innobackupex 工具，我这里写过一次：http://www.cnblogs.com/wangxiaoqiangs/p/5961413.html

场景：

  A -> B
    -> C
    -> D -> E

一、增加节点 C

# 由于有从库 B ，所以我们去 B 上面执行备份

shell > innobackupex --user=xx --password=xx --slave-info --safe-slave-backup --no-timestamp alldatabase

# 保存 B 的 Slave 信息，因为新增的节点 C 与 B 同级

shell > scp -r alldatabase mysql-nodeC:/root

# 将备份数据拷贝到新节点 C 上，如果节点 C 在外网，建议先压缩一下

# 登录节点 C

shell > vim /etc/my.cnf  # 准备配置文件

[client]
port           = 3306
socket         = /tmp/mysql.sock

[mysqld]
socket         = /tmp/mysql.sock
datadir        = /data/mysql_data
log-error      = error.log
skip-name-resolve
skip-external-locking
default-storage-engine = MyISAM
key_buffer_size = 3072M
sort_buffer_size = 256M
read_buffer_size = 4M
read_rnd_buffer_size = 8M
myisam_sort_buffer_size = 128M
table_open_cache = 1024
table_definition_cache = 1024
thread_cache_size = 8
back_log = 512
connect_timeout = 20
max_connections = 1500
wait_timeout = 120
interactive_timeout = 120
max_allowed_packet = 1024M

slow_query_log = ON
long_query_time = 2
slow-query-log-file = slow.log

query_cache_type = 0
query_cache_size = 0
query_cache_limit = 0

tmp_table_size = 128M
max_heap_table_size = 128M

innodb_file_per_table = 0
innodb_buffer_pool_size = 1024M

read-only = 1
relay-log = mysql-relay-bin
expire_logs_days = 7
server-id = 1020

# 注意：server-id 必须全局唯一
# 如果新增的是节点 D，那么需要将后4行替换为如下7行

relay-log = mysql-relay-bin
relay_log_purge = ON
log-slave-updates = 1
log-bin = mysql-bin
binlog_format = mixed
expire_logs_days = 7
server-id = 1020

# 因为节点 D 不仅是 Master A 的 Slave，同时也是节点 E 的 Master

shell > innobackupex --apply-log alldatabase  # 在备份目录上应用日志
161111 13:56:04 completed OK!

shell > innobackupex --copy-back alldatabase  # 将备份文件拷贝到 my.cnf 中 datadir 指定目录
161111 14:10:21 completed OK!

shell > chown -R mysql.mysql /data/mysql_data

shell > /etc/init.d/mysql.server start

shell > cat /data/mysql_data/xtrabackup_slave_info  # 记录 master_log_file 和 master_log_pos 的值
CHANGE MASTER TO MASTER_LOG_FILE='mysql-bin.003742', MASTER_LOG_POS=1072788827

shell > mysql -uxx -pxx -A

mysql> stop slave;
mysql> reset slave all;
mysql> change master to master_host='x.x.x.x',master_user='xx',master_password='xx',master_log_file='mysql-bin.003742',master_log_pos=1072788827;
mysql> start slave;
mysql> show slave status\G

             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes
mysql> quit

# master_host='A'
# MySQL Slave 节点 C 搭建完成

二、新增节点 E

# 由于节点 D 后面没有从节点，所以备份要在节点 D 上执行

shell > innobackupex --user=xx --password=xx --no-timestamp alldatabase

# 我们要的是节点 D 的信息，所以不需要保存 Slave 信息( Slave 信息为跟节点 A 同步信息 )

shell > scp -r alldatabase mysql-nodeE:/root

# 将备份数据拷贝到新节点 E 上，如果节点 E 在外网，建议先压缩一下

shell > mysql -uxx -pxx -A  # 跟上面不一样的地方，需要新建立一个授权用户

mysql> grant replication slave on *.* to xx@'%' identified by 'xx';

# 登录节点 E

shell > vim /etc/my.cnf  # 准备配置文件

[client]
port           = 3306
socket         = /tmp/mysql.sock

[mysqld]
socket         = /tmp/mysql.sock
datadir        = /data/mysql_data
log-error      = error.log
skip-name-resolve
skip-external-locking
default-storage-engine = MyISAM
key_buffer_size = 3072M
sort_buffer_size = 256M
read_buffer_size = 4M
read_rnd_buffer_size = 8M
myisam_sort_buffer_size = 128M
table_open_cache = 1024
table_definition_cache = 1024
thread_cache_size = 8
back_log = 512
connect_timeout = 20
max_connections = 1500
wait_timeout = 120
interactive_timeout = 120
max_allowed_packet = 1024M

slow_query_log = ON
long_query_time = 2
slow-query-log-file = slow.log

query_cache_type = 0
query_cache_size = 0
query_cache_limit = 0

tmp_table_size = 128M
max_heap_table_size = 128M

innodb_file_per_table = 0
innodb_buffer_pool_size = 1024M

read-only = 1
relay-log = mysql-relay-bin
expire_logs_days = 7
server-id = 10201

# 同样注意 server-id 全局唯一

shell > chown -R mysql.mysql /data/mysql_data

shell > /etc/init.d/mysql.server start

shell > cat /data/mysql_data/xtrabackup_info

binlog_pos = filename 'mysql-bin.000001', position '112094418'

shell > mysql -uxx -pxx -A

mysql> stop slave;
mysql> reset slave all;
mysql> change master to master_host='x.x.x.x',master_user='rep',master_password='xx',master_log_file='mysql-bin.000001',master_log_pos=112094418;
mysql> start slave;
mysql> show slave status\G

             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes
mysql> quit

# master_host='D'
# master_user='rep' # 这里的用户密码，是节点 D 上的，并不是节点 A 上的
# MySQL Slave 节点 E 搭建完成

# 不采用 innobackupex 时，可以使用如下方法：
# stop slave;
# flush tables with read lock;
# show slave status\G
# 记录 master_log_file，master_log_pos
# 新开一个终端，scp -r /data/mysql_data x.x.x.x:/data/mysql_data，拷贝完成就可以 unlock tables;
# chown -R mysql.mysql /data/mysql_data
# /etc/init.d/mysql.server start
# mysql -uxx -pxx -A
# chanage master to ...
# start slave;

# mysql.sock 如果放在 datadir 下，记得 chmod 755 /data/mysql_data，不然 zabbix_agentd 连接不到数据库
# End