在 Centos 6 上从 MySQL 5.6 升级到 MariaDB 10 后，MySQL（又名 MariaDB）服务无法启动

Question

我需要一些帮助来弄清楚为什么在 Centos 6 上从 MySQL 5.6 升级到 MariaDB 10 如此困难。

以下是我正在使用的步骤，随后出现错误

cat /etc/centos-release
CentOS release 6.8 (Final)

mysql -V
mysql  Ver 14.14 Distrib 5.6.35, for Linux (x86_64) using  EditLine wrapper

yum -y update

mysqldump --all-databases -p  > /tmp/backupdatabase.sql 

cp /etc/my.cnf /tmp/my.cnf.bak

service mysqld stop

chkconfig mysqld off

yum remove mysql* mysql-server mysql-devel mysql-libs

edit /etc/yum.repos.d/mariadb.repo to point to MariaDB 10 source

yum clean all

yum install MariaDB-server galera MariaDB-client MariaDB-shared MariaDB-backup MariaDB-common

到目前为止一切看起来都很好：

service mysql start

service mysql start
Starting MariaDB.190801 13:39:12 mysqld_safe Logging to '/var/lib/mysql/designtv.designtv.net.err'.
190801 13:39:12 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
...../etc/init.d/mysql: line 260: kill: (3056) - No such process

错误日志

190801 14:00:25 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
190801 14:00:25 [ERROR] Can't read from messagefile '/usr/share/mysql/english/errmsg.sys'
/usr/libexec/mysqld: Unknown error 1146
190801 14:00:25 [ERROR] Can't open the mysql.plugin table. Please run mysql_upgrade to create it.
190801 14:00:25  InnoDB: Initializing buffer pool, size = 8.0M
190801 14:00:25  InnoDB: Completed initialization of buffer pool
190801 14:00:25  InnoDB: Started; log sequence number 0 44233
190801 14:00:25 [ERROR] Aborting

190801 14:00:25  InnoDB: Starting shutdown...
190801 14:00:30  InnoDB: Shutdown completed; log sequence number 0 44233
190801 14:00:30 [Note]
190801 14:00:30 mysqld_safe mysqld from pid file /var/lib/mysql/designtv.designtv.net.pid ended

“/var/lib/mysql/designtv.designtv.net.err” 是全世界可读的

ls -la /usr/share/mysql/english/errmsg.sys
-rw-r--r-- 1 root root 61764 Jan 29  2019 /usr/share/mysql/english/errmsg.sys

“[错误] 无法打开 mysql.plugin 表。请运行 mysql_upgrade 来创建它。”

在服务器运行之前，我无法运行 mysql_ugrade。

使用备份我完全删除了 /var/lib/mysql 中现有的数据库文件后进行了彻底的重新安装。

现在我收到此错误：

190801 16:50:40  InnoDB: Initializing buffer pool, size = 8.0M
190801 16:50:40  InnoDB: Completed initialization of buffer pool
InnoDB: Error: log file ./ib_logfile0 is of different size 0 50331648 bytes
InnoDB: than specified in the .cnf file 0 5242880 bytes!
190801 16:50:40 [ERROR] Plugin 'InnoDB' init function returned error.
190801 16:50:40 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
190801 16:50:40 [ERROR] Aborting

190801 16:50:40 [Note]
190801 16:50:40 mysqld_safe mysqld from pid file /var/lib/mysql/designtv.designtv.net.pid ended
190801 16:53:35 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
190801 16:53:35 [ERROR] Can't read from messagefile '/usr/share/mysql/english/errmsg.sys'
190801 16:53:35  InnoDB: Initializing buffer pool, size = 8.0M
190801 16:53:35  InnoDB: Completed initialization of buffer pool
InnoDB: Error: log file ./ib_logfile0 is of different size 0 50331648 bytes
InnoDB: than specified in the .cnf file 0 5242880 bytes!
190801 16:53:35 [ERROR] Plugin 'InnoDB' init function returned error.
190801 16:53:35 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
190801 16:53:35 [ERROR] Aborting

Answer 1

您认为删除数据库文件后就进行了全新安装，但事实并非如此。

否则，第二个错误日志将不会表明 ib_logfile0 的大小与配置文件指定的大小不同。

至少你忘记删除 ib_logfile0 和 /etc/mysql 目录。

我建议你：

service mysql stop
killall -9 mysqld
mv /etc/mysql /etc/mysql.old
mv /var/lib/mysql /var/lib/mysql.old

此时，使用包管理器删除已安装的 mysql。

重新安装 mysql，它会要求您输入初始化详细信息，例如 root 密码等。

重新配置 mysql 以满足您的需要并重新启动它。

如果您有单独的备份，请逐个恢复数据库。避免恢复 mysql db 并手动重新创建所有用户。这对您来说会更容易。

这应该会带来更新且可正常运行的 mysql。