Horton

Re: [PoC] Improve dead tuple storage for lazy vacuum

Masahiko Sawada <sawada.mshk@gmail.com>

View thread View on postgr.es
From: Masahiko Sawada <sawada.mshk@gmail.com>
To: John Naylor <johncnaylorls@gmail.com>
Cc: PostgreSQL-development <pgsql-hackers@postgresql.org>, Nathan Bossart <nathandbossart@gmail.com>, Andres Freund <andres@anarazel.de>, Matthias van de Meent <boekewurm+postgres@gmail.com>, Yura Sokolov <y.sokolov@postgrespro.ru>
Date: 2023-12-20T11:35:38Z
Lists: pgsql-hackers

Attachments

On Tue, Dec 19, 2023 at 4:37 PM John Naylor <johncnaylorls@gmail.com> wrote:
>
> On Tue, Dec 19, 2023 at 12:37 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> >
> > On Mon, Dec 18, 2023 at 3:41 PM John Naylor <johncnaylorls@gmail.com> wrote:
> > > Let's do it in just one place. In TidStoreCreate(), do
> > >
> > > /* clamp max_bytes to at least the size of the empty tree with
> > > allocated blocks, so it doesn't immediately appear full */
> > > ts->control->max_bytes = Max(max_bytes, {rt, shared_rt}_memory_usage);
> > >
> > > Then we can get rid of all the worry about 1MB/2MB, 64kB, 70kB -- all that.
> >
> > But doesn't it mean that even if we create a shared tidstore with
> > small memory, say 64kB, it actually uses 1MB?
>
> This sounds like an argument for controlling the minimum DSA segment
> size. (I'm not really in favor of that, but open to others' opinion)
>
> I wasn't talking about that above -- I was saying we should have only
> one place where we clamp max_bytes so that the tree doesn't
> immediately appear full.

Thank you for your clarification. Understood.

I've updated the new patch set that incorporated comments I got so
far. 0007, 0008, and 0012 patches are updates from the v45 patch set.
In addition to the review comments, I made some changes in tidstore to
make it independent from heap. Specifically, it uses MaxOffsetNumber
instead of MaxHeapTuplesPerPage. Now we don't need to include
htup_details.h. It enlarged MaxBlocktableEntrySize but it's still 272
bytes.

BTW regarding the previous comment I got before:

> - RT_PTR_ALLOC *slot;
> + RT_PTR_ALLOC *slot = NULL;
>
> We have a macro for invalid pointer because of DSA.

I think that since *slot is a pointer to a RT_PTR_ALLOC it's okay to set NULL.

As for the initial and maximum DSA segment sizes, I've sent a summary
on that thread:

https://www.postgresql.org/message-id/CAD21AoCVMw6DSmgZY9h%2BxfzKtzJeqWiwxaUD2T-FztVcV-XibQ%40mail.gmail.com

I'm going to update RT_DUMP() and RT_DUMP_NODE() codes for the next step.


Regards,

--
Masahiko Sawada
Amazon Web Services: https://aws.amazon.com

Commits

  1. radixtree: Fix SIGSEGV at update of embeddable value to non-embeddable.

  2. Get rid of anonymous struct

  3. Teach radix tree to embed values at runtime

  4. Teach TID store to skip bitmap for small numbers of offsets

  5. Use bump context for TID bitmaps stored by vacuum

  6. Fix alignment of stack variable

  7. Use TidStore for dead tuple TIDs storage during lazy vacuum.

  8. Rethink create and attach APIs of shared TidStore.

  9. Fix inconsistent function prototypes with function definitions.

  10. Fix a calculation in TidStoreCreate().

  11. Fix potential integer handling issue in radixtree.h.

  12. Add TIDStore, to store sets of TIDs (ItemPointerData) efficiently.

  13. Fix link error for test_radixtree module on Windows

  14. Blind attempt to fix ODR violations

  15. Fix incorrect format specifier for int64

  16. Fix redefinition of typedefs

  17. Add template for adaptive radix tree

  18. Fix signedness error in 9f225e992 for gcc

  19. Introduce helper SIMD functions for small byte arrays

  20. Optimize vacuuming of relations with no indexes.

  21. Add bound check before bsearch() for performance

  22. Allocate consecutive blocks during parallel seqscans