Based on my experience: MySQL is significantly easier to set up, and it's much easier to admin. PG still seems to do the thing where it assumes your system user is your database user unless you do a bunch of things, and the weird blurred lines between the operating system and the database can get
really confusing at the start.
On top of that, MySQL is well-documented and lots of people use it for lots of different things, which makes the most common (and most of the uncommon) problems eminently google-able.
Anecdotally, I've avoided Postgres also in part because it had the same problem that Perl had when I got into Linux: the community was pretty toxic and unwelcoming. I asked about replication once, and I got a bunch of answers that fell into a few categories: 'Why on earth do you need replication?', 'You don't need replication, Postgres is fast enough', 'set up a script to rsync your data directories over and over again', 'mysql replication is unreliable so you shouldn't use it so it doesn't matter'.
That sort of attitude drove me (and, I'd wager, a lot of other people) away from Postgres and towards the vastly more welcoming MySQL community, and I'd wager that a lot of people just kind of stuck there because why switch?
Postgres has added a lot of pretty great features lately, and it seems good? But I'm left wondering why I should use Postgres for my RDBMS, my JSON data store, and my KV store, when history has shown that one piece of software that expands to do everything rarely does any of it well.
(Postgres has since gained replication abilities)
