So I wanted to test MySQL recovery from Innodb corruption.
file-per-table is enabled.
I have a few test db's on the MySQL instance.
I shutdown mysql. Manually edited/added random characters to an .idb file within a DB.
... I wanted to test credibility of innodbchecksum, mysqlcheck, etc.
MySQL started but shut down right away with errors. Only innodb-force-recovery=6 seemed to keep MySQL up.
I couldn't dump the corrupted DB at all. I couldn't select the table. if I tried MySQL crashed again. I also couldn't drop the table or database even. error.log mentioned something about not being able to remove the folder.
I manually did an rm on the entire db folder. edit: I understand it may be impossible to get the data back after wrecking the .idb file, I am only interested in bringing MySQL back to normal status.
MySQL seemed to have started but with errors everywhere in the log saying:
121122 17:34:01 InnoDB: Error: page 7 log sequence number 1 2040399532 InnoDB: is in the future! Current system log sequence number 0 8204. InnoDB: Your database may be corrupt or you may have copied the InnoDB InnoDB: tablespace but not the InnoDB log files...
So my question is, how can this MySQL instance be recovered? If an .idb file is corrupted due to disk error or such, is the entire MySQL instance hosed? is there any recovery from this?
Edit: there must be a way of forcing drop a corrupted DB or even table without breaking the tablespace.