On Thu, Jan 2, 2014 at 3:19 PM, Andres Freund <andres@2ndquadrant.com> wrote:
>> I was wondering if we could somehow arrange to not
>> release the subtransaction's AccessShareLock on the table, as long as it
>> was protecting toasted references someplace.
>
> Sounds fairly ugly...
I think the only principled fixes are to either retain the lock or
forcibly detoast before releasing it. The main problem I see with
retaining the lock is that you'd need a way of finding out the
relation OIDs of all toast pointers you might later decide to expand.
I don't have an amazingly good idea about how to figure that out.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company