Re: [HACKERS] Re: Clarifying "server starting" messaging in pg_ctlstart without --wait

* Michael Paquier (michael.paquier@gmail.com) wrote:
> On Thu, Jan 19, 2017 at 6:20 AM, Stephen Frost <sfrost@snowman.net> wrote:
> > On Wed, Jan 18, 2017 at 16:15 Robert Haas <robertmhaas@gmail.com> wrote:
> >> On Wed, Jan 18, 2017 at 3:59 PM, Stephen Frost <sfrost@snowman.net> wrote:
> >> > For non-cold standby configurations, pg_ctl is going to return just as
> >> > soon as the database has finished crash recovery, which in most cases
> >> > will probably be on the order of a few seconds.
> >>
> >> /me is poleaxed.
> >>
> >> Yeah, that was a confused sentence- most of the time it's going to return
> >> on the order of a few seconds because we're doing regular startup without
> >> having to do any crash recovery.
> >>
> >>
> >> For actual crash recovery cases, it'll take between a few seconds and
> >> checkpoint_timeout, as I described up-thread, based on the amount of
> >> outstanding WAL.
>
> Recovering up to the minumum recovery point could take minutes!

Right, in an actual crash recovery case, it'll take longer, up to
checkpoint_timeout.

I am certainly hopeful that, most of the time, we're actually starting
up from a cleanly shut down system.

Thanks!

Stephen