> > I personally find the behavior of pg_ctl connecting to the
> > postmaster _after_ it's started up to verify that it has indeed
> > started up, to be a rather broken behavior and think that it
> > shouldn't be the default.
>
> ?? It isn't.
>
> I agree that we could do with a better mechanism, though.
Hrm, I stand corrected. The startup script passes -w. Hrm, idea.
Any chance the PID filename could be hard linked to the format:
postmaster-%d.pid where %d is the start time of the postmaster?
Cleanup would have to be altered to remove both files, and the
postmaster would have to be updated to retain the time at which it
started. Anyway, my thought being that the new and upcoming protocol
could include an uptime query that'd include the time that the server
was started and compare it with the pid file. Odds are slim that
they'd be the same and it'd allow for verification that the server had
started without authentication using hard to replicate, insecure data.
-sc
--
Sean Chittenden