On Thu, Mar 13, 2014 at 12:11 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Andres Freund <andres@2ndquadrant.com> writes:
>> On 2014-03-13 11:26:10 -0400, Robert Haas wrote:
>>> I have however had the thought before that it would be nice to allow
>>> for callbacks of invalidation functions of some kind even on catalogs
>>> that don't have catcaches.
>
>> Unfortunately the format catcache invalidations have is pretty tightly
>> tied to the hash function catcaches use internally. And we need
>> something that can be included in the WAL, otherwise it won't work on HS
>> nodes.
>
> Note that the existence of a cache doesn't mean it's necessarily
> populated. In this example, a catcache on pg_seclabels could be used just
> fine, as long as it wasn't used to load labels for large objects. Even if
> it were never used at all, it would still provide a usable conduit for
> invalidation events.
That'd need some commenting, but it seems like a possibly workable
approach that wouldn't require changing much.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
