I have a large database (180 GB), running under Windows, with a runaway rollback that starts running as soon as the database is started up.

I have started the database with innodb_force_recovery=3 to prevent the rollback from running while I try to fix the problem. As I understand it from other questions on StackExchange, the correct thing to do is to drop the table causing the rollback and restore it from a backup. All that would be no problem. However, I am unable to identify the table triggering the rollback. SHOW ENGINE INNODB STATUS is no help -- it just says that there is a rollback. I am running mysqlcheck over all the tables, but everything is so far showing up "OK".

Is there any way to determine what table(s) are causing the rollback?

  • What does SHOW BINARY LOGS; send back? Commented Sep 23, 2015 at 9:01
  • What monster query was running when it crashed? If it does not get rolled back, you will have trash in the table(s).
    – Rick James
    Commented Oct 25, 2015 at 6:32


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.