Re: Inadequate thought about buffer locking during hot standby replay

On Tue, Nov 13, 2012 at 10:32 AM, Merlin Moncure <mmoncure(at)gmail(dot)com> wrote:
> On Tue, Nov 13, 2012 at 9:03 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> Ok. It won't help all that much on 9.0, though.
>>
>> Well, it won't help GIST much, but the actually-reported-from-the-field
>> case is in btree, and it does fix that.
>>
>> It occurs to me that if we're sufficiently scared of this case, we could
>> probably hack the planner (in 9.0 only) to refuse to use GIST indexes
>> in hot-standby queries. That cure might be worse than the disease though.
>
> if anything, it should be documented. if you do this kind of thing
> people will stop installing bugfix releases.

Agreed. I think doing that in a back-branch release would be
extremely user-hostile.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company