On 2020-11-24 13:32, Magnus Hagander wrote:
I think it boils down to that today the output from initdb is entirely geared towards people running initdb directly and starting their server manually, and very few people outside the actual PostgreSQL developers ever do that. But there are still a lot of people who run initdb through their wrapper manually (for redhat you have to do that, for debian you only have to do it if you're creating a secondary cluster but that still a pretty common operation).
Perhaps it's worth asking whom the advice applies to then. You suggest it's mostly developers. I for one am still grumpy that in 9.5 we removed the variant of the hint that suggested "postgres -D ..." instead of pg_ctl. I used to copy and paste that a lot. The argument back then was that the hint should target end users, not developers. I doubt that under the current circumstances, running pg_ctl start from the console is really appropriate advice for a majority of users. (For one thing, systemd will kill it when you log out.) I don't know what better advice would be, though. Maybe we need to add some kind of text adventure game into initdb.
-- Peter Eisentraut 2ndQuadrant, an EDB company https://www.2ndquadrant.com/
