Tom Lane wrote:
> Lamar Owen <lamar.owen@wgcr.org> writes:
> > That is a great side-effect. Now, if there were a way to initdb just
> > template0, leaving everything else in place, then rebuilding template1
> I'm missing something --- I don't see how this affects pg_upgrade one
> way or the other, except of course that it should be prepared to cope
> with user data in template1 (not sure if it does or not now).
Maybe I spoke too soon....
> pg_upgrade won't be usable for the 7.1 transition anyway, because of WAL
> changes (page header format is changing). I dunno whether it will be
> usable at all under WAL --- Vadim will have to comment on that.
Of course, the upgrade from 7.0 to 7.1 involves a physical on disk
format change (implying pg_upgrade's uselessness in doing its job
there). You know, our version numbers aren't at all consistent WRT disk
format. ISTM that 6.5 should have been 7.0 due to its format change,
and 7.1 should be 8.0.
Changing from 6.5 to 7.0 is less of a format change than 6.4 to 6.5 or
7.0 to 7.1.
But, what's in a version number.... :-) They don't _have_ to be
consistent, really.
--
Lamar Owen
WGCR Internet Radio
1 Peter 4:11