On Thu, Jul 26, 2012 at 3:01 PM, Daniel Farina <daniel@heroku.com> wrote:
> For example: suppose pg_upgrade emitted full-page-write records in the
> format of the new postgres version on an unoccupied timeline. One can
> use PG.next tools to report on the first txid
and by txid I meant WAL position, which mucks it up a bit because
that's not an option on old Postgres-es. But the overall point is the
same: I think pg_upgrade would be much more safe and usable in the
standby setting if it emitted data that depend upon a posterior
WAL-position at the point the cluster diverges onto a new timeline
that is the new version of Postgres, and it would be nice if that data
incremented the WAL position rather than being applied out-of-band.
And, it seems that emitting full page image records in the format of
the new version is a way to accomplish that.
--
fdr
