Re: parallel vacuum comments - Mailing list pgsql-hackers

From Masahiko Sawada
Subject Re: parallel vacuum comments
Date
Msg-id CAD21AoCV68wkXUo4JmhdydTSXuOz-UGPuJR6gturCTn7UwPPqA@mail.gmail.com
Whole thread Raw
In response to Re: parallel vacuum comments  (Masahiko Sawada <sawada.mshk@gmail.com>)
Responses Re: parallel vacuum comments  (Amit Kapila <amit.kapila16@gmail.com>)
RE: parallel vacuum comments  ("houzj.fnst@fujitsu.com" <houzj.fnst@fujitsu.com>)
List pgsql-hackers
On Tue, Nov 30, 2021 at 4:45 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
>
> On Tue, Nov 30, 2021 at 3:00 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
> >
> > On Tue, Nov 30, 2021 at 11:03 AM houzj.fnst@fujitsu.com
> > <houzj.fnst@fujitsu.com> wrote:
> > >
> > > On Mon, Nov 29, 2021 11:38 AM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> > > >
> > >
> > > 2)
> > > +               /* Reinitialize the parallel context to relaunch parallel workers */
> > > +               if (!pvs->first_time)
> > >
> > > It seems the ParallelVacuumState::first_time was not initialized before ?
> > >
> >
> > Yeah, I also notice this while looking at the patch.
>
> Thank you for the comments, Amit and Hou.
>
> >
> > One more thing it seems the patch has removed even the existing error
> > callback from parallel_vacuum_main. I suggested that we can enhance or
> > add a new one if required in a separate patch but let's keep the
> > current one as it is.
>
> Understood.
>
> >
> > Can we think of splitting the patch in the following manner: (a) the
> > patch to get rid of bitmap to represent whether particular index
> > supports parallel vacuum and rename of functions (b) any other stuff
> > to improve the current implementation, (c) move the parallel vacuum
> > related code to a separate file?
>
> Okay, I'll split the patch and submit them.
>

I've attached updated patches.

The first patch is the main patch for refactoring parallel vacuum
code; removes bitmap-related code and renames function names for
consistency. The second patch moves these parallel-related codes to
vacuumparallel.c as well as common functions that are used by both
lazyvacuum.c and vacuumparallel.c to vacuum.c. The third patch adds
regression tests for parallel vacuum on different kinds of indexes
with multiple index scans. Please review them.

Regards,

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

Attachment

pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: pg_dump versus ancient server versions
Next
From: Daniel Gustafsson
Date:
Subject: Re: pgbench logging broken by time logic changes