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: 2023-12-08T08:05:53Z
Lists: pgsql-hackers 
On Fri, Dec 8, 2023 at 3:45 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
>
> On Fri, Dec 8, 2023 at 1:37 PM John Naylor <johncnaylorls@gmail.com> wrote:
> >
> > On Fri, Dec 8, 2023 at 8:57 AM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> >
> > > It's still unclear to me why the value doesn't need to contain the size.
> > >
> > > If I understand you correctly, in RT_GET(), the tree allocs a new
> > > memory and updates the slot where the value is embedded with the
> > > pointer to the allocated memory, and returns the pointer to the
> > > caller. Since the returned value, newly allocated memory, is still
> > > empty, the callner needs to copy the contents of the old value to the
> > > new value and do whatever else it needs to.
> > >
> > > If the value is already a single-leave value and RT_GET() is called
> > > with a larger size, the slot is always replaced with the newly
> > > allocated area and the caller needs to copy the contents? If the tree
> > > does realloc the value with a new size, how does the tree know the new
> > > value is larger than the existing value? It seems like the caller
> > > needs to provide a function to calculate the size of the value based
> > > on the length.
> >
> > Right. My brief description mentioned one thing without details: The
> > caller would need to control whether to re-alloc. RT_GET would pass
> > the size. If nothing is found, the tree would allocate. If there is a
> > value already, just return it. That means both the address of the
> > slot, and the local pointer to the value (with embedded, would be the
> > same address).

BTW Given that the actual value size can be calculated only by the
caller, how does the tree know if the value is embedded or not? It's
probably related to how to store combined pointer/value slots. If leaf
nodes have a bitmap array that indicates the corresponding slot is an
embedded value or a pointer to a value, it would be easy. But since
the bitmap array is needed only in the leaf nodes, internal nodes and
leaf nodes will no longer be identical structure, which is not a bad
thing to me, though.

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