We host a Rails application on AWS that uses mysql in the amazon-rds database. We have a giant table in the database that we want to migrate, but the migration takes several days, due to millions of rows. In particular, we move VARCHAR to the TEXT column
How can I migrate this large production database if users are not idle?
One of the ideas I heard is to set up a copy of the database and migrate there and switch it to the main database when it is done. However, I am not sure how this takes into account user data entered during the migration.
Update: This may be relevant: amazon-rds suggests reading replicas and Multi-AZ deployments that seem to be done for this type of thing. However, for the first time, having done this, we will welcome guidance on any method, whether this or that.
source share