On Fri, Dec 7, 2012 at 09:16:56AM +0000, Simon Riggs wrote:
> On 7 December 2012 04:02, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
> > Andrew Dunstan wrote:
> >>
> >> On 12/06/2012 09:23 PM, Bruce Momjian wrote:
> >
> >> >As soon as pg_dump stopped dumping the CREATE INDEX, pg_upgrade would
> >> >stop creating creating it in the new cluster, and not transfer the index
> >> >files.
> >>
> >> So we'll lose the index definition and leave some files behind? This
> >> sounds a bit messy to say the least.
> >
> > I find it hard to get excited about this being a real problem. If the
> > index has been kept invalid, how come the definition is so valuable?
>
> Agreed.
>
> I don't see the problem... just say "rebuild any invalid indexes
> before you run pg_upgrade".
>
> I don't think pg_upgrade should take the responsibility of fixing
> everything broken before you run an upgrade. Making that rod for our
> backs looks pretty bad here and could get worse if other situations
> come to light.
>
> Maybe it should have a mode where it detects that it would fail if you
> attempted the upgrade...
That's what pg_upgrade --check does, but see my email about in-process
concurrent index builds also causing a failure.
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +