On Wed, Apr 26, 2006 at 06:42:53PM -0400, Tom Lane wrote:
> "Jim C. Nasby" <jnasby@pervasive.com> writes:
> > Try running a first index build by itself and then running them in
> > parallel.
>
> Yeah, this is probably the best workaround for now. I think we should
> look at making it fully concurrent-safe per upthread comments, but that
> won't be happening in existing release branches.
>
> Also, the only case where it's a problem is if the first two index
> builds finish at almost exactly the same time. It might be possible to
> overlap the first two index builds with reasonable safety so long as you
> choose indexes with very different sorting costs (eg, integer vs text
> columns, different numbers of columns, etc).
What about not updating if the tuplecount is within X percent? Would
that be safe enough to back-port? I've been trying to think of a reason
why disabling the current behavior of CREATE INDEX forcing reltuples to
be 100% accurate but I can't think of one...
--
Jim C. Nasby, Sr. Engineering Consultant jnasby@pervasive.com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461
