mysql 主从的复制恢复

Posted 2020-11-25 六月过半

备份恢复

1.从库查询状态

slave status

Master_Log_File: mysql-bin.000120

Slave_IO_Running: Yes

Slave_SQL_Running: No

RELAY_Master_Log_FILE: mysql-bin.000119

Exec_Master_Log_Pos: 87663322

Last_Error: xxxxx

这种状态就是要从mysql-bin.000119 的87663322 行开始恢复，只能一点点比对了。

2.主库查询binlog

show binlog events in ‘mysql-bin.000119‘ from 87663322 limits 100000;
然后看卡在哪个事务上了。

3.从库跳过错误的事务

stop slave;
set GLOBAL SLAVE_SLAVE_SKIP_COUNTER = 1;
start slave;
然后回到步骤1 继续找。如果每个事务都不长，可以调整SLAVE_SLAVE_SKIP_COUNTER的值一次跳过多个事务。

4.slave的过滤配置

官方手册 https://dev.mysql.com/doc/refman/5.7/en/replication-rules.html
按照官方建议，主库的binlogxxx配置有一定作用，在从库设置过滤规则可以让断开复制的几率更小点。
replicate-wild-do-table = xyz.abc_%
类似这种，表名如果有_需要转译。跨库的问题，只能注意了。

5.其它

记录些相关表
mysql.slave_master_info
mysql.slave_relay_log_info
mysql.slave_worker_info
performance_schema.replication_applier_configuration
performance_schema.replication_applier_filters
performance_schema.replication_applier_global_filters
performance_schema.replication_applier_status
performance_schema.replication_applier_status_by_coordinator
performance_schema.replication_applier_status_by_worker
performance_schema.replication_connection_configuration
performance_schema.replication_connection_status
performance_schema.replication_group_member_stats
performance_schema.replication_group_members