Here is Pavel's patch rebased to master branch, added the dropdb
--force option, a test case & documentation.
I'm willing to work on it if needed. What are possible bad things that
could happen here? Is the documentation clear enough?
Thanks.
On Tue, Dec 18, 2018 at 4:34 PM Marti Raudsepp <marti@juffo.org> wrote:
>
> Hi
>
> > út 18. 12. 2018 v 16:11 odesílatel Filip Rembiałkowski <filip.rembialkowski@gmail.com> napsal:
> >> Please share opinions if this makes sense at all, and has any chance
> >> going upstream.
>
> Clearly since Pavel has another implementation of the same concept,
> there is some interest in this feature. :)
>
> On Tue, Dec 18, 2018 at 5:20 PM Pavel Stehule <pavel.stehule@gmail.com> wrote:
> > Still one my customer use a patch that implement FORCE on SQL level. It is necessary under higher load when is not
easyto synchronize clients.
>
> I think Filip's approach of setting pg_database.datallowconn='false'
> is pretty clever to avoid the synchronization problem. But it's also a
> good idea to expose this functionality via DROP DATABASE in SQL, like
> Pavel's patch, not just the 'dropdb' binary.
>
> If this is to be accepted into PostgreSQL core, I think the two
> approaches should be combined on the server side.
>
> Regards,
> Marti Raudsepp