Re: WIP: Upper planner pathification

On Mon, Mar 7, 2016 at 11:52 AM, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
>
> On Sun, Mar 6, 2016 at 9:59 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> > Perhaps it was intentional when written, but if Robert's advice is
correct
> > that the new upper-planner path nodes should copy up parallel_degree
from
> > their children, then it cannot be the case that parallel_degree>0 in a
> > node above the scan level implies that that node type has any special
> > behavior for parallelism.
> >

Right.

>
> > I continue to bemoan the lack of documentation about what these fields
> > mean.
> >

Point taken and if Robert doesn't feel otherwise, I can try to write a
patch to explain the newly added fields.

> > As far as I can find, the sum total of the documentation about
> > this field is
> >
> > int parallel_degree; /* desired parallel degree; 0 = not
parallel */
>
> While it doesn't particularly relate to parallel joins, I've expressed
> a general concern about the max_parallel_degree GUC that I think is
> worth considering again:
>
>
http://www.postgresql.org/message-id/CAM3SWZRs1mTvrKkAsY1XBShGZXkD6-HNxX3gq7x-p-dz0ZtkMg@mail.gmail.com
>
> In summary, I think it's surprising that a max_parallel_degree of 1
> doesn't disable parallel workers entirely.
>

I have responded on the thread where you have raised that point with my
thoughts, discussing it here on a separate point can dilute the purpose of
this thread.

With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com