Re: InnoDB recovery

[Marko Mäkelä]

Rudolf,

You could proceed as follows:

1) dump the internal InnoDB data dictionary using innodb_table_monitor:
http://dev.mysql.com/doc/refman/5.1/en/innodb-troubleshooting.html
http://dev.mysql.com/doc/refman/5.1/en/innodb-monitor.html

2) ensure that the frm files correspond to the InnoDB data dictionary

3) for each table, do SELECT * FROM table

InnoDB will probably crash for some tables in step 3) because of missing or unwritten pages. Just restart and proceed to next table.

The InnoDB redo log is essential for correct operation. There can be large numbers of dirty pages in the InnoDB buffer pool (RAM) that have not been written to disk. In crash recovery (when mysqld is restarted), these dirty pages will be reconstructed from the redo log (ib_logfiles).

Best regards,

Marko Mäkelä
Innobase Oy/Oracle Corp.