Re: pg_rewind vs superuser

On Thu, Apr 11, 2019 at 10:33:13AM +0200, Magnus Hagander wrote:
> And I think it would make sense to wait by default, but we could then also
> have a commandline parameter that says "don't wait, instead error out in
> case the checkpoint isn't done".
> 
> Or something like that?

Yes, that would be the idea.  You still need to cover the case where
both instances are on the same timeline, in which case you could wait
for a checkpoint forever, so we'd need to change the current behavior
a bit by making sure that we always throw an error if both nodes are
still on the same timeline after the timeout (see TAP test
005_same_timeline.pl).  I am not sure that you need a separate option
to control the case where you don't want to wait though.  Perhaps we
could have a separate switch, but a user could also just set
--timeout=0 to match that behavior.
--
Michael