我重新启动我的VPS盒子（手动/硬重新启动），从此，无论如何，MySQL无法启动。
我做了一个tail /var/log/syslog ，我得到这个：
Feb 20 11:49:33 kyrgyznews mysqld[11461]: ) ;InnoDB: End of page dump 575 Feb 20 11:49:33 kyrgyznews mysqld[11461]: 110220 11:49:33 InnoDB: Page checksum 1045788239, prior-to-4.0.14-form checksum 236985105 576 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: stored checksum 1178062585, prior-to-4.0.14-form stored checksum 236985105 577 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: Page lsn 0 10651, low 4 bytes of lsn at page end 10651 578 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: Page number (if stored to page already) 3, 579 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: space id (if created with >= MySQL-4.1.1 and stored already) 0 580 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: Database page corruption on disk or a failed 581 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: file read of page 3. 582 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: You may have to recover from a backup. 583 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: It is also possible that your operating 584 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: system has corrupted its own file cache 585 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: and rebooting your computer removes the 586 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: error. 587 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: If the corrupt page is an index page 588 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: you can also try to fix the corruption 589 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: by dumping, dropping, and reimporting 590 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: the corrupt table. You can use CHECK 591 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: TABLE to scan your table for corruption. 592 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: See also InnoDB: http://dev.mysql.com/doc/refman/5.0/en/forcing-recovery.html 593 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: about forcing recovery. 594 Feb 20 11:49:33 kyrgyznews mysqld[11461]: InnoDB: Ending processing because of a corrupt database page. 595 Feb 20 11:49:33 kyrgyznews mysqld_safe[11469]: ended 596 Feb 20 11:49:47 kyrgyznews /etc/init.d/mysql[12228]: 0 processes alive and '/usr/bin/mysqladmin --defaults-file=/etc/mysql/debian.cnf ping' resulted in 597 Feb 20 11:49:47 kyrgyznews /etc/init.d/mysql[12228]: ^G/usr/bin/mysqladmin: connect to server at 'localhost' failed 598 Feb 20 11:49:47 kyrgyznews /etc/init.d/mysql[12228]: error: 'Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)' 599 Feb 20 11:49:47 kyrgyznews /etc/init.d/mysql[12228]: Check that mysqld is running and that the socket: '/var/run/mysqld/mysqld.sock' exists! 600 Feb 20 11:49:47 kyrgyznews /etc/init.d/mysql[12228]: 601 Feb 20 11:49:56 kyrgyznews mysqld_safe[13437]: started 602 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: The log sequence number in ibdata files does not match 603 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: the log sequence number in the ib_logfiles! 604 Feb 20 11:49:56 kyrgyznews mysqld[13440]: 110220 11:49:56 InnoDB: Database was not shut down normally! 605 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: Starting crash recovery. 606 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: Reading tablespace information from the .ibd files... 607 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: Restoring possible half-written data pages from the doublewrite 608 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: buffer... 609 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: Database page corruption on disk or a failed 610 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: file read of page 3. 611 Feb 20 11:49:56 kyrgyznews mysqld[13440]: InnoDB: You may have to recover from a backup.
我已经看过它引用的页面， http://dev.mysql.com/doc/refman/5.0/en/forcing-innodb-recovery.html ，但在搞乱任何设置之前，我想知道经验丰富的DBA会build议做什么？ 强迫复苏是否有害？
PS – 我没有任何更新到MySQL。 版本是MySQL Ver 14.12 Distrib 5.0.51a，用于使用readline 5.2的debian-linux-gnu（i486）。
根据该日志中的内容，我build议您运行一些磁盘检查，因为它有可能导致数据库损坏的磁盘错误。 如果磁盘检出，如果无法修复数据库，则可能需要从备份中恢复。
确保文件/var/run/mysqld/mysqld.sock存在于你的服务器中。 如果不存在，您可以创build它。 这个tuts将帮助你http://anthologyoi.com/computers/cant-connect-to-local-mysql-server-through-socketerror.html