概述

今天主要汇总一下MGR常见的一些问题及解决办法，主要思路还是通过报错日志去找对应的解决方法。

ERROR 3077 (HY000): To have multiple channels, repository cannot be of type FILE;

报错：

ERROR 3077 (HY000): To have multiple channels, repository cannot be of type FILE; Please check the repository configuration and convert them to TABLE.

解决：

多源复制加入了一个叫做Channel的概念， 每一个Channel都是一个独立的Slave，都有一个IO_THREAD和SQL_THREAD。原理和普通复制一样。我们只需要对每一个Master执行Change Master 语句，只需要在每个语句最后使用For Channel来进行区分。由于复制的原理没有改变，在没有开启GTID的时候Master的版本可以是MySQL5.5、5.6、5.7。并且从库需要master-info-repository、relay-log-info-repository设置为table，否则会报错

故只需在my.cnf添加这两个参数即可。

ERROR 1123 (HY000): Can't initialize function 'group_replication'; Plugin initialization function failed.

报错截图如下：

解决：

1、查看日志

2、在my.cnf增加参数

master-info-repository=TABLE														#复制元数据存储在系统表而不是文件
relay-log-info-repository=TABLE													#中继日志信息写入到表而不是文件
gtid_mode=on																						#开启全局事务ID
enforce_gtid_consistency=on															#开启全局事务ID强一致性
slave_preserve_commit_order=on													#控制从库并行reply时事务提交的顺序
binlog_checksum=NONE																		#禁用二进制日志时间校验和
transaction_write_set_extraction=XXHASH64								#以便在server收集写集合的同时将其记录到二进制日志。写集合基于每行的主键，并且是行更改后的唯一标识此标识将用于检测冲突。

ERROR 3092 (HY000): The server is not configured properly to be an active member of the group.

添加组成员报错，如下：

解决：

1、查看日志

2、在my.cnf增加参数

LOG-SLAVE-UPDATES

ERROR 3092 (HY000): The server is not configured properly to be an active member of the group.

启动组复制报错：

解决：

1、查看报错日志

[Warning] Plugin group_replication reported: '[GCS] Connection attempt from IP address xxx.116 refused. Address is not in the IP whitelist.
[ERROR] Plugin group_replication reported: '[GCS] Error connecting to the local group communication engine instance.'
[ERROR] Plugin group_replication reported: '[GCS] The member was unable to join the group. Local port: 13306'

2、修改loose-group_replication_ip_whitelist参数，增加IP白名单

从库状态为recovering

查看MGR组成员，发现从库异常

解决：

1、查看报错日志：

[ERROR] Slave I/O for channel 'group_replication_recovery': error connecting to master 'repl@ZCAPZC:3306' - retry-time: 60  retries: 1, Error_code: 2005
 [Note] Slave I/O thread for channel 'group_replication_recovery' killed while connecting to master
[Note] Slave I/O thread exiting for channel 'group_replication_recovery', read up to log 'FIRST', position 4
[ERROR] Plugin group_replication reported: 'For details please check performance_schema.replication_connection_status table and error log messages of Slave I/O for channel group_replication_recovery.'
2020-09-15T10:42:54.468843Z 11 [Note] Plugin group_replication reported: 'Retrying group recovery connection with another donor. Attempt 2/10'

2、这里的问题在于连接不上主库，只需要检查相关的连接即可，排查发现问题在于连接ZCAPZC:3306主库连不上，解析不了ZCAPZC这个主机名，修改/etc/hosts解析即可。

Plugin group_replication reported: 'This member has more executed transactions than those present in the group.

在添加成员时日志报错：

[ERROR] Plugin group_replication reported: 'This member has more executed transactions than those present in the group. Local transactions: 25009bc8-f67a-11ea-8e19-005056bf82fa:1-15 > Group transactions: 25009bc8-f67a-11ea-8e19-005056bf82fa:1-13'

解决:

1、查看报错日志

2、设置参数group_replication_allow_local_disjoint_gtids_join=1以加入成员

3、从库reset master

清空binlog日志

4、set global read_only=0后根据报错日志处理数据不一致问题

5、重新加入组复制

start group_repllication;

以上就是MGR常见的一些报错了，其实根据报错日志对应解决就可以了。

后面会分享更多devops和DBA方面内容，感兴趣的朋友可以关注下！