Re: pg_upgrade problem with invalid indexes

On Tue, Dec 11, 2012 at 09:19:34PM +0100, Andres Freund wrote:
> > > > Only 9.2 :(. Before that there was no DROP INDEX CONCURRENTLY and in 9.3
> > > > there's an actual indislive field and indisready is always set to false
> > > > there if indislive is false.
> > > >
> > > > But I see no problem using !indisvalid || !indisready as the condition
> > > > in all (supported) versions.
> > >
> > > OK, updated patch attached.
> >
> > Patch applied back to 9.0.
> >
> > Now that it is applied, I need to publicize this.  How do I do that?
> > Josh mentioned my blog.
> 
> Not really sure why this needs to be so much more prominent than other
> fixes?

Because it creates an post-upgrade system that might return wong results
with no warning.

> > What would cause these invalid indexes?  Just CREATE INDEX CONCURRENTLY
> > failures?  What types of failures would users have if these invalid
> > indexes had been upgraded by pg_upgrade?
> 
> Afaics it could be anything from wrong responses to queries to crashes
> although the former seems to be more likely.
> 
> > Can they test their indexes in any way?  I assume they can't run
> > queries on the old cluster to check.
> 
> I don't see how.

That's what I thought.  This is not good.  :-(

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com
 + It's impossible for everything to be true. +