On Sat, Jun 30, 2007 at 02:16:11PM +0000, Christopher Browne wrote:
> 
> 1.  It isn't true of the origin node.  On the origin node, the schema
> isn't actually broken (because we aren't hiding triggers/rules).  You
> *can*, *always*, run pg_dump successfully against the origin.

But there's no such thing as an origin node, purely; it's only true
of the origin node for a table; and pg_dump won't proceed if there is
even one table that is a replica, because that table has caused some
catalogue corruption.  You simply cannot pg_dump from a database that
has even _one_ replica in it, unless you are using an older pg_dump,
in which case you're missing a number of sanity checks that could
leave you with a broken dump.

> In Slony-I 2.0, on PG 8.3, this is all alleviated.  

Or will be, maybe, assuming the code arrives and works as advertised.
None of it exists yet.  (And I'd caution people against being too
keen to rush 2.0 into production. The .0 releases of Slony have so
far been plagued with really nasty bugs.  That's not surprising in
code as young as this, but it does nobody any good to pretend
otherwise.)

A

-- 
Andrew Sullivan  | ajs at crankycanuck.ca
Everything that happens in the world happens at some place.
		--Jane Jacobs