corrupt database.

[Dillan]

hi all, i want to develop an application that automate a recovery from a corrupt database. to simulate a corrupt database, i opened .ibd files in a hex editor, and modified the contents with garbage but mysql would not even start.
on modifying the middle of the file, mysql started, and it operated like normal. but on performed a CHECK TABLE, it crashed one time, and also, indicated a corrupt table. what steps would you take to recover data from it? -- This message may have been cut off and the rest will only be shown to members. To become a member, click here --

[sierra]

Try to restore from backup and replay the binary log or you lose data. Things happen, backup and binary log are the minimum required tools to survive them.

Or use replication i.e. a master and two slave setup. Lets call them A, B and C:

A fails

failover to B

copy data from C to A to recover A without interfering with B

When not involved in failover scenarios, the B and C servers would probably be used for serving queries, to improve performance and get an increased return on the investment.

But replication is not a substitute for backup and binary log. You still need those to recover from DROP TABLE. -- This message may have been cut off and the rest will only be shown to members. To become a member, click here --