In some databases, it’s possible to disable the redo log for individual queries. Sure, you have to be careful and you have to be absolutely sure of what you are doing, but it can be especially helpful if you know exactly what you’re doing.
In a particular case, I need to delete tens of millions of rows from a table. The only way I’ve figured out how to monitor this process is to do “show engine innodb status”, find the particular transaction in question, and then follow the growing “undo log entries” value.
I see after 10 minutes that I have 4 million undo log entries - I have a ways to go yet.
But this leads me to think. If I can actually disable undo log entries for this operation - true I wouldn’t be able to track progress anymore - but then wouldn’t the operation run that much faster?