Mariabackup备份MariaDB副本库时卡在Read redo log up to LSN循环求解决

操作场景与执行命令 #

在主从架构的MariaDB环境中，对副本服务器执行全量备份，使用命令：

mariabackup --backup --target-dir=/backup/ --user=backup --password=backup --host=192.168.xxx.xxx

异常现象 #

命令持续循环输出重复日志内容，无法完成备份：

2023-06-09 13:51:10 0 [Note] InnoDB: Read redo log up to LSN=1625600
[00] 2023-06-09 13:51:10 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:11 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:12 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:13 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:14 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:15 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:16 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:17 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:18 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:19 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:20 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:21 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:22 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:23 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:24 >> log scanned up to (1625437)
2023-06-09 13:51:25 0 [Note] InnoDB: Read redo log up to LSN=1625600
[00] 2023-06-09 13:51:25 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:26 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:27 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:28 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:29 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:30 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:31 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:32 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:33 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:34 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:35 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:36 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:37 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:38 >> log scanned up to (1625437)
[00] 2023-06-09 13:51:39 >> log scanned up to (1625437)

多次尝试均出现相同问题。

服务器进程状态 #

在副本服务器执行show processlist，可见备份连接处于sleep状态，同时存在以下InnoDB后台进程：

1   system user NULL    Daemon  NULL    InnoDB purge worker NULL    0.000
2   system user NULL    Daemon  NULL    InnoDB purge coordinator    NULL    0.000
3   system user NULL    Daemon  NULL    InnoDB purge worker NULL    0.000
4   system user NULL    Daemon  NULL    InnoDB purge worker NULL    0.000
5   system user NULL    Daemon  NULL    InnoDB shutdown handler NULL    0.000

解决思路建议 #

• 检查副本复制状态：执行show slave status\G，确认Slave_IO_Running和Slave_SQL_Running是否为Yes，排查是否存在复制延迟或报错。
• 处理InnoDB shutdown进程：InnoDB shutdown handler进程存在说明数据库可能处于异常关闭流程中，尝试正常重启MariaDB服务，确保进程完全退出后再启动。
• 验证备份权限：确认backup用户拥有RELOAD、LOCK TABLES、REPLICATION CLIENT权限，同时/backup/目录需具备读写权限。
• 修复LSN不一致问题：日志中显示读取的redo log LSN与扫描到的LSN不匹配，可尝试在副本执行FLUSH TABLES WITH READ LOCK后再启动备份，或检查InnoDB日志文件是否损坏。
• 调整备份参数：添加--no-lock参数测试（仅适用于只读副本），或指定--innodb-log-file-size匹配数据库配置，避免日志扫描异常。
• 排查资源瓶颈：确认/backup/目录所在磁盘有足够空间，检查服务器CPU、内存、IO资源是否充足，是否因资源不足导致备份进程停滞。

内容的提问来源于stack exchange，提问作者Stefano P.