I hear this suggested in talks / on the web frequently enough to wonder if any company has ever successfully migrated from mySQL to PostgresQL, or from database technology foo to database technology bar, without coming close to killing the company or blocking the development team from doing more profitable work? Or is this something that many people suggest but have never actually done?<p>Just move to a "real" database and all your problems will be solved, right? I don't think so - you will just replace one set of problems with another. And when you're dealing with a schema that is multiples of TB, hundreds of tables, stored procs, triggers, poorly normalised, contains a mass of undocumented domain knowledge, goes back over a decade, then it's not going to happen.<p>If it does happen then it's probably over a period of years, or effectively a rewrite. Databases are like mortgages, they're technical debt that take decades to pay off.<p>So, have you ever done the migration? How did it go?