Re: pg_upgrade from 9.4 to 10.4

Bruce Momjian <bruce(at)momjian(dot)us> writes:
> The comment at the top of src/port/system.c explains why we need those
> quotes. Spaces was not the issue.

So, while starting to prepare the release notes, I looked at this patch
again and I'm still wondering why it's designed this way. AFAICS,
the current state of affairs is:

1. previous server was shut down nicely: all good.
2. previous server was shut down with "-m immediate": we complain and die.
3. previous server was shut down with "kill -9": we clean up and press on.

I am not really sure why case 2 deserves a wrist slap and making the user
clean it up manually when we're willing to clean up automatically in
case 3. If we're going to treat them differently, that's backwards.

Right now is probably not a good time to fix this, but it seems like
something that could be improved. I'd be kind of inclined to remove
the pidfile checking business altogether in favor of inspecting the
state in pg_control; or at least do them both in the same place with
the same recovery attempt if we don't like what we see.

regards, tom lane