Tom Lane wrote:
> I don't immediately see any good way to fix this. I think the "triple
> parity" rule as it stands is hopelessly broken, but I don't know what
> to replace it with, even granting that we don't need to maintain on-disk
> compatibility. (We'd have to tell people to reindex SPGist indexes
> anyway, because of the risk that they already contain circular links;
> so switching to a new layout rule doesn't seem to add any more pain.)
> Or we could try to modify the insertion algorithm so it doesn't lock
> two levels of the tree at once, but that seems pretty risky.
Is this the chance to add a metapage?
--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services