The error I gave is a similar one to the one I used to get with “major” upgrades that happened when Ubuntu decided it was time to upgrade.
It happens and I seriously never claimed that it was an ultra common problem, merely that upgrades in Postgres are more intentional and not painful except for a little extra work between major versions. The standard upgrade path within major versions; 9.x or 10.x or 11.x or 12.x is working just the same as MySQL, except I have much more experience of MySQL completely fumbling their “automatic unattended” upgrade or even the mysql_upgrade command.
Mostly because in the real world outside of engineering cultures databasen are massively abused, ISAM tables that are constantly updated, InnoDB ibdata1 in the terabytes, poor configs, replicas that have skipped a few queries, column changes inside a transaction that failed but actually modified data, it happens. Usually I am called in to clean the mess.
Major difference here is that Postgres doesn’t leave a mess, so I never have the kind of issues that I am describing in this thread with it, and you don’t because I am guessing that you’re there when they’re installed, someone with knowledge was actively maintaining. or you have a lot of people to help with shortcomings.
I get it though. you’ve got your sunk cost knowledge of MySQL and you’ve been on large support teams for it. Maybe you’re afraid I’m suggesting that this knowledge goes out the window. and it has gotten better, but I wouldn’t give my kids watered down led infused soft drinks just because I had suffered through led poisoning. I remember coming to blows with you in other threads over the years because you think MySQL can be saved or is totally fine, but honestly, just, no.