Kevin Grittner <kgrittn(at)ymail(dot)com> wrote:
> I am investigating.
Information from the syscache entry for the table was being
referenced after the relation was closed. We were generally
getting away with it until the new locking caused a check for
invalidation messages at a new spot. Moving the heap_close() call
down a few lines fixed it by keeping the syscache entry valid until
we were through with it.
--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company