On Thu, Dec 6, 2012 at 09:41:00PM -0500, Andrew Dunstan wrote:
>
> On 12/06/2012 09:23 PM, Bruce Momjian wrote:
> >On Thu, Dec 6, 2012 at 09:10:21PM -0500, Tom Lane wrote:
> >>Bruce Momjian <bruce@momjian.us> writes:
> >>>On Thu, Dec 6, 2012 at 07:53:57PM -0500, Tom Lane wrote:
> >>>>Because CREATE INDEX CONCURRENTLY can't drop the index if it's already
> >>>>failed. It's not because we want to do that, it's an implementation
> >>>>restriction of the horrid kluge that is CREATE/DROP INDEX CONCURRENTLY.
> >>>Well, what is the logic that pg_dump dumps it then, even in
> >>>non-binary-upgrade mode?
> >>Actually, I was thinking about proposing exactly that. Ideally the
> >>system should totally ignore an invalid index (we just fixed some bugs
> >>in that line already). So it would be perfectly consistent for pg_dump
> >>to ignore it too, with or without --binary-upgrade.
> >>
> >>One possible spanner in the works for pg_upgrade is that this would mean
> >>there can be relation files in the database directories that it should
> >>ignore (not transfer over). Dunno if that takes any logic changes.
> >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.
You will lose the index definition, but the new cluster will not have
the invalid index files from the old cluster.
> Making the user fix it seems much more sensible to me. Otherwise I
> suspect we'll find users who get strangely surprised when they can
> no longer find any trace of an expected index in their upgraded
> database.
Well, the indexes weren't being used.
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +