On Tue, May 18, 2021 at 06:28:49PM +0900, Masahiko Sawada wrote:
> On Mon, May 10, 2021 at 3:03 PM Bruce Momjian <br...@momjian.us> wrote:
> >
> > I have committed the first draft of the PG 14 release notes. You can
> > see the most current build of them here:
> >
> > https://momjian.us/pgsql_docs/release-14.html
> >
>
> I think we need to mention in the release note that
> vacuum_cleanup_index_scale_factor GUC parameter has been removed and
> vacuum_cleanup_index_scale_factor storage parameter has been
> deprecated (please refer to commit 9f3665fb and effdd3f3b63).
Looking at the full commit message:
commit 9f3665fbfc
Author: Peter Geoghegan <p...@bowt.ie>
Date: Wed Mar 10 16:27:01 2021 -0800
Don't consider newly inserted tuples in nbtree VACUUM.
Remove the entire idea of "stale stats" within nbtree VACUUM (stop
caring about stats involving the number of inserted tuples). Also
remove the vacuum_cleanup_index_scale_factor GUC/param on the master
branch (though just disable them on postgres 13).
The vacuum_cleanup_index_scale_factor/stats interface made the
nbtree AM
partially responsible for deciding when pg_class.reltuples stats
needed
to be updated. This seems contrary to the spirit of the index AM
API,
though -- it is not actually necessary for an index AM's bulk
delete and
cleanup callbacks to provide accurate stats when it happens to be
inconvenient. The core code owns that. (Index AMs have the
authority
to perform or not perform certain kinds of deferred cleanup based on
their own considerations, such as page deletion and recycling, but
that
has little to do with pg_class.reltuples/num_index_tuples.)
This issue was fairly harmless until the introduction of the
autovacuum_vacuum_insert_threshold feature by commit b07642db,
which had
an undesirable interaction with the
vacuum_cleanup_index_scale_factor
mechanism: it made insert-driven autovacuums perform full index
scans,
even though there is no real benefit to doing so. This has been
tied to
a regression with an append-only insert benchmark [1].
Also have remaining cases that perform a full scan of an index
during a
cleanup-only nbtree VACUUM indicate that the final tuple count is
only
an estimate. This prevents vacuumlazy.c from setting the index's
pg_class.reltuples in those cases (it will now only update pg_class
when
vacuumlazy.c had TIDs for nbtree to bulk delete). This arguably
fixes
an oversight in deduplication-related bugfix commit 48e12913.
[1]
https://smalldatum.blogspot.com/2021/01/insert-benchmark-postgres-is-still.html
Author: Peter Geoghegan <p...@bowt.ie>
Reviewed-By: Masahiko Sawada <sawada.m...@gmail.com>
Discussion:
https://postgr.es/m/cad21aoa4whthn5uu6+wscz7+j_rcejmcuh94qcoupub42sh...@mail.gmail.com
--> Backpatch: 13-, where autovacuum_vacuum_insert_threshold was added.
This was backpatched into PG 13.3, which was released last week:
<listitem>
<!--
Author: Peter Geoghegan <p...@bowt.ie>
Branch: master [9f3665fbf] 2021-03-10 16:27:01 -0800
Branch: REL_13_STABLE [9663d1244] 2021-03-10 16:26:58 -0800
Author: Peter Geoghegan <p...@bowt.ie>
Branch: master [5f8727f5a] 2021-03-10 17:07:57 -0800
Branch: REL_13_STABLE [1fc5a5738] 2021-03-10 17:07:55 -0800
-->
<para>
Disable the <varname>vacuum_cleanup_index_scale_factor</varname>
parameter and storage option (Peter Geoghegan)
</para>
<para>
The notion of tracking <quote>stale</quote> index statistics
proved
to interact badly with
the <varname>autovacuum_vacuum_insert_threshold</varname>
parameter,
resulting in unnecessary full-index scans and consequent
degradation
of autovacuum performance. The latter mechanism seems superior,
so
remove the stale-statistics logic. The control parameter for
that,
<varname>vacuum_cleanup_index_scale_factor</varname>, will be
removed entirely in v14. In v13, it remains present to avoid
breaking existing configuration files, but it no longer does
anything.
</para>
</listitem>
Therefore, it didn't show up in my src/tools/git_changelog output, and I
did not include it.
--
Bruce Momjian <br...@momjian.us> https://momjian.us
EDB https://enterprisedb.com
If only the physical world exists, free will is an illusion.
