Re: pg_background (and more parallelism infrastructure patches)

On 09/09/14 18:49, Robert Haas wrote:
> On Tue, Sep 9, 2014 at 12:33 PM, Petr Jelinek <petr@2ndquadrant.com> wrote:
>> I am also not sure that I like the pq_redirect_to_shm_mq being directly
>> exposed for use in bgworker, what I would like is to have elog interface to
>> which you tell that you want errors sent to shm_mq handle and that one would
>> then do the necessary calls (It's basically same principle but for the sake
>> of cleanliness/correctness I think it should be elog and not pq from the
>> point of bgworker).
>
> I think that's completely wrong.  As the patch series demonstrates,
> it's not limited to propagating ErrorResponse and NoticeResponse.  It
> can also propagate NotifyResponse and RowDescription and DataRow and
> anything else that comes along.  We are not just propagating errors;
> we are propagating all protocol messages of whatever type.  So tying
> it to elog specifically is not right.
>

Oh in that case, I think what Andres proposed is actually quite good. I 
know the hook works fine it just seems like using somewhat hackish 
solution to save 20 lines of code.
The reason why I am not proposing anything better is that my solution 
when I did prototyping in this area was to just check if my pq_dsm_mq 
handle is set or not and call the appropriate function from the wrapper 
based on that which does not seem like big improvement over the hook 
approach... Anything better/more flexible that I could come up with is 
basically same idea what Andres already wrote.

>> Please don't forget to add this to next commitfest.
>
> OK, done.  But it would be awfully nice if we could actually make some
> progress on hammering out a design everyone can live with here.
> Letting it sit for another 5 weeks is not going to help anything, and
> it will hold up getting any more stuff after this done in time for
> 9.5.
>

Yeah I said that just as formality, I wrote the response now and not in 
5 weeks exactly for this reason, I am all for discussing this now and I 
think it's important to hammer this infrastructure out sooner rather 
than later.

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