Re: parallel vacuum comments

and

On Fri, Dec 3, 2021 at 6:56 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>
> On Fri, Dec 3, 2021 at 2:33 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
> >
> > On Thu, Dec 2, 2021 at 6:01 PM Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com> wrote:
> > >
> > > I've attached updated patches.
> > >
> >
> > I have a few comments on v4-0001.
> >
>
> The new test proposed by v4-0003 is increasing the vacuum_parallel.sql
> timing by more than 10 times. It appears to be taking the highest time
> among all the tests in make check. I think it is because of a large
> amount of data being processed by the test.

Right.

> I think it is good to use
> it for testing of patches during development but won't be a good idea
> to commit. Can we reduce its timings?

On reflection, we already have test cases for:

* a parallel vacuum does bulkdeletion and cleanup
* a parallel vacuum does only cleanup

and the case that the new tests add is that a vacuum does bulkdeletion
twice and cleanup. Given the increasing the regression test time, the
thrid patch might not worth to be added.

Regards,

--
Masahiko Sawada
EDB: https://www.enterprisedb.com/