Our tables have gotten large enough that it takes hours to do an alter table. So, this could be a very good thing. It doesn’t relieve any of the issues dealing with the table being locked during this time, but at least we’ll know how much longer it will be locked.
Makes me wonder if (since our servers are in the cloud), we could apply any new indexes on a slave, let the slave catch up, then promote the slave to master. That quite possibly could work with little down time. Hmmmm.
Comments are closed.