环境：某项目的DB19数据库服务器，无从库也无开过binlog日志，负载有时比较高但基本运行稳定。备份情况：本地每天完整备份保留一周再rsync到异地备份机保留一月，由于数据量的增长加上每天晚上rsync到异地机时流量非常大，所以单独建台虚拟机作为DB19的备份。
备份措施：
1) DB19作为主库开启binlog日志。 2) DB19为从库开设同步帐号并开放复制权限。 3) DB19写脚本在凌晨一点锁表进行一次完整备份，并记录锁表前的binlog文件和position位置点（show master status\G;可查看信息） 4) 将备份的文件导入到从库并设置主从同步. 5) 从库上脚本实施每月1号本地进行一次全备，每天00-20每3小时进行一次增量备份(备份binlog日志)1.DB19 主库设置并开启binlog日志
my.cnf配置文件更改： [mysqld] datadir=/var/lib/mysql socket=/var/lib/mysql/mysql.sock user=mysqlserver-id=1 #id号在LAN中唯一
log-bin=log #开启binlog日志 binlog-ignore-db=mysql #忽略权限库的同步max_connections=1000
long_query_time = 10 log-slow-queries = /var/log/mysql/mysql_slow.log #开启慢查询日志2. DB19为从库开设同步帐号
mysql>grant replication slave on *.* to identified by '123456'; mysql> flush privileges; mysql> show grants for ";3. DB19凌晨一点完整备份脚本并记录锁表前master status信息
#!/bin/bash ################################################ #this scripts is bakdb19 ################################################MYUSER=root
MYPASS="pwd19" MYSOCK=/var/lib/mysql/mysql.sockDATA_PATH=/backup/0314
LOG_FILE=${DATA_PATH}/mysqllogs_0314.log DATA_FILE=${DATA_PATH}/mysql_backup_0314.sql.gzMYSQL_PATH=/usr/bin
MYSQL_CMD="$MYSQL_PATH/mysql -u$MYUSER -p$MYPASS -S $MYSOCK"MYSQL_DUMP="$MYSQL_PATH/mysqldump -u$MYUSER -p$MYPASS -S $MYSOCK -A -B --flush-logs --single-transaction"
$MYSQL_CMD -e "flush tables with read lock;"echo "-----show master status result-----" >>$LOG_FILE
$MYSQL_CMD -e "show master status;" >>$LOG_FILE ${MYSQL_DUMP} | gzip > $DATA_FILE $MYSQL_CMD -e "unlock tables;"脚本加入到crontab中：
0 01 * * * sh /root/dbbak0314.sh>>/root/tasks/backup.log 2>&14. 将备份文件和记录master信息的文件.log复制到从库服务器
备份文件mysql_backup_0314.sql.gz master信息的文件mysqllogs_0314.log 1)从库安装： #yum install mysql mysq-server #my.cnf配置文件： [mysqld] datadir=/data/mysql socket=/data/mysql/mysql.sock user=mysqllog-bin=log
#mysql_install_db #/etc/init.d/mysqld start2)数据导入
#gzip -d mysql_backup_0314.sql.gz #mysql -S /data/mysql/mysql.sock <mysql_backup_0314.sql #more mysqllogs_0314.log -----show master status result----- File Position Binlog_Do_DB Binlog_Ignore_DB log.000003 61139130 mysql3)配置同步
# vi /root/.my.cnf 创建密码文件，省得mysql进去每次输入密码 [client] password=pwd19mysql> CHANGE MASTER TO
-> MASTER_HOST='192.168.7.19', -> MASTER_USER='sqlsync', -> MASTER_PASSWORD='123456', -> MASTER_PORT=3306, -> MASTER_LOG_FILE='log.000003', #指向锁表前记录的binlog文件和位置点，可查看mysqllogs_0314.log -> MASTER_LOG_POS=61139130, -> MASTER_CONNECT_RETRY=60; Query OK, 0 rows affected (0.08 sec)mysql> start slave;
Query OK, 0 rows affected (0.00 sec)mysql> show slave status\G;
*************************** 1. row *************************** Slave_IO_State: Queueing master event to the relay log Master_Host: 192.168.7.19 Master_User: sqlsync Master_Port: 3306 Connect_Retry: 60 Master_Log_File: log.000028 Read_Master_Log_Pos: 4 Relay_Log_File: mysqld-relay-bin.000002 Relay_Log_Pos: 229 Relay_Master_Log_File: log.000003 Slave_IO_Running: Yes Slave_SQL_Running: No Replicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 1062 Last_Error: Error 'Duplicate entry '197963' for key 1' on query. Default database: 'agrocms'. Query: 'INSERT INTO rocms.web_swgkbdxx (web_id,town_code,name,area,bdxx,swgk,bdxxsum,swgksum,day) values ('116597','330522105212','xx村','xx/xx/xx县/xx镇/xx村',0,0,0,17,'2013-03-14')' Skip_Counter: 0 Exec_Master_Log_Pos: 61139130 Relay_Log_Space: 5383075 Until_Condition: None Until_Log_File: Until_Log_Pos: 0 Master_SSL_Allowed: No Master_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: NULL 1 row in set (0.25 sec)ERROR:
No query specifiedmysql> INSERT INTO rocms.web_swgkbdxx (web_id,town_code,name,area,bdxx,swgk,bdxxsum,swgksum,day) values ('116597','330522105212','xx村','xx/xx/xx县/xx镇/xx村',0,0,0,17,'2013-03-14');
ERROR 1062 (23000): Duplicate entry '116597-2013-03-14' for key 2 提示错误重复插入 解决方法： mysql>set global sql_slave_skip_counter=1;不行，因为后面很多重复的insert语句执行不了同样的提示错误 只好将slave-skip-errors=1062加入到my.cnf中 跳过错误代码为1062（重复更新）的所有语句. mysql> show slave status\G; *************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.7.19 Master_User: sqlsync Master_Port: 3306 Connect_Retry: 60 Master_Log_File: log.000029 Read_Master_Log_Pos: 189077906 Relay_Log_File: mysqld-relay-bin.000058 Relay_Log_Pos: 34636888 Relay_Master_Log_File: log.000029 Slave_IO_Running: Yes Slave_SQL_Running: Yes Replicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0 Last_Error: Skip_Counter: 0 Exec_Master_Log_Pos: 189077906 Relay_Log_Space: 34636888 Until_Condition: None Until_Log_File: Until_Log_Pos: 0 Master_SSL_Allowed: No Master_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0 1 row in set (0.00 sec)ERROR:
No query specified已同步OK，可用checksum来检验主从同步数据的一致性，线上机器需锁表，暂不检验了
补充：
Slave_IO_Running: No 极有可能是网络问题或主库设了防火墙阻挡了，或设置的MASTER_LOG_POS位置信息不对.测试OK，最好再将change master to信息在my.cnf配置文件中指定下：
# more /etc/my.cnf [mysqld] datadir=/data/mysql socket=/data/mysql/mysql.sock user=mysqllog-bin=log
expire-logs-days=3 log-slave-updatesserver-id=2
master-host=192.168.7.19 master-user=sqlsync master-password=wer852 master-port=3306 master-connect-retry=60 relay-log=mysqld-relay-bin5. 从库完整+增量备份脚本（暂略）
6. 恢复
1)目录级的备份，恢复时可直接进行目录的恢复frm,MYD,MYI文件
2)mysqldump全备加--tab参数时会生成sql和txt两个文件，sql记录所有包括create table的相关表结构更新语句，txt是数据文件。恢复时mysql先执行sql语句，再mysqlimport执行txt文件，如果表结构无更改，恢复时可直接mysqlimport导入txt文件.
3)mysqdump全备+增量备份，先最近一次的全备恢复，如果有多个binlog可按时间顺序逐个导入.