Re: Create subscription with `create_slot=false` and incorrect slot name

On 01/06/17 04:44, Peter Eisentraut wrote:
> On 5/31/17 09:40, Robert Haas wrote:
>> On Tue, May 30, 2017 at 3:01 PM, Peter Eisentraut
>> <peter(dot)eisentraut(at)2ndquadrant(dot)com> wrote:
>>> On 5/25/17 17:26, Peter Eisentraut wrote:
>>>> Another way to fix this particular issue is to not verify the
>>>> replication slot name before doing the drop. After all, if the name is
>>>> not valid, then you can also just report that it doesn't exist.
>>>
>>> Here is a possible patch along these lines.
>>
>> I don't see how this solves the problem. Don't you still end up with
>> an error message telling you that you can't drop the subscription, and
>> no guidance as to how to fix it?
>
> Well, the idea was to make the error message less cryptic.
>
> But I notice that there is really little documentation about this. So
> how about the attached documentation patch as well?
>
> As mentioned earlier, if we want to do HINT messages, that will be a bit
> more involved and probably PG11 material.
>

I think the combination of those patches is probably good enough
solution for PG10 (I never understood the need for name validation in
ReplicationSlotAcquire() anyway).

--
Petr Jelinek http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services