On 2017-03-23 06:55:53 +0800, Craig Ringer wrote:
> On 22 March 2017 at 21:06, Andres Freund <andres@anarazel.de> wrote:
> > Hi,
> >
> > On 2017-03-21 09:05:26 +0800, Craig Ringer wrote:
> >> > 0002 should be doable as a whole this release, I have severe doubts that
> >> > 0003 as a whole has a chance for 10 - the code is in quite a raw shape,
> >> > there's a significant number of open ends. I'd suggest breaking of bits
> >> > that are independently useful, and work on getting those committed.
> >>
> >> That would be my preference too.
> >
> >
> >> The parts I think are important for Pg10 are:
> >
> >> * Ability to create logical slots on replicas
> >
> > Doesn't this also imply recovery conflicts on DROP DATABASE? Besides,
> > allowing to drop all slots using a database upon DROP DATABASE, is a
> > useful thing on its own.
>
> Definitely beneficial, otherwise recovery will stop until you drop
> slots, which isn't ideal.
s/isn't ideal/not acceptable/ ;)
> >> * Ability to advance (via feedback or via SQL function) - no need to
> >> actually decode and call output plugins at al
> >
> > That pretty much requires decoding, otherwise you really don't know how
> > much WAL you have to retain.
>
> Yes, and to update restart_lsn and catalog_xmin correctly.
> I was thinking that by disallowing snapshot use and output plugin
> invocation we'd avoid the need to support cancellation on recovery
> conflicts, etc, simplifying things considerably.
That seems like it'd end up being pretty hacky - the likelihood that
we'd run into snapbuild error cross-checks seems very high.
Greetings,
Andres Freund
