MIME-Version: 1.0
References: <CA+TgmoY3m3RimCHifQnxU3YDVV7oYRbRUX4GZ4FuYkhnhe0giA@mail.gmail.com>
 <20220203215048.rmwjzixdzaetedxq@alap3.anarazel.de> <CAFBsxsHp3brGqWu5vU_6d5wfOU5i40c8u3Hg=3TtN4qZwmAecw@mail.gmail.com>
 <CAH2-WzkEStBgnk0buJYAKw5jpRYUE+-tuLYn5U+vDyhXNfu+DQ@mail.gmail.com>
 <CAFBsxsFiMPxQ-dHZ8tOgktn=+ffeJT3+GinZ4zdOGbmAnCYadA@mail.gmail.com>
 <CAH2-WzkpZcat1YmEVAKpcMBaWRXUd2vGVKHtrBjzBGDL=j6mbA@mail.gmail.com>
 <CAD21AoCAKf+yq8pHLpmyxxS=A5zNFUvxTeergLOhobjYaJbF+Q@mail.gmail.com>
 <CAH2-Wznb+W7HdrcT7vLe5hvxA131ay2EwdY+r3z8Uehd_Jr5uw@mail.gmail.com>
 <CA+Tgmoadjx+r8_gGbbnNifL6vEyjZntiQRPzyixrUihvtZ5jdQ@mail.gmail.com>
 <CAH2-Wz=5RuakSPA9euKmUCa3WX5HB2GFeLd7xKXyH5AwbpCGtg@mail.gmail.com>
 <20220216181846.4syh3mnq5qzjrdiz@alap3.anarazel.de> <CAH2-Wz=gkPvzf+H6nXtcxT2STGvh7-hmys4X-oGsihddZQte1Q@mail.gmail.com>
In-Reply-To: <CAH2-Wz=gkPvzf+H6nXtcxT2STGvh7-hmys4X-oGsihddZQte1Q@mail.gmail.com>
From: Peter Geoghegan <pg@bowt.ie>
Date: Wed, 16 Feb 2022 11:18:19 -0800
Message-ID: <CAH2-Wzkb2ixRbxNsUZ2PA5-uqSh4+Xfq-NmcEKD=JNcncyDPTA@mail.gmail.com>
Subject: Re: do only critical work during single-user vacuum?
To: Andres Freund <andres@anarazel.de>
Cc: Robert Haas <robertmhaas@gmail.com>, Masahiko Sawada <sawada.mshk@gmail.com>, 
	John Naylor <john.naylor@enterprisedb.com>, 
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Archived-At: <https://www.postgresql.org/message-id/CAH2-Wzkb2ixRbxNsUZ2PA5-uqSh4%2BXfq-NmcEKD%3DJNcncyDPTA%40mail.gmail.com>
Precedence: bulk

On Wed, Feb 16, 2022 at 10:27 AM Peter Geoghegan <pg@bowt.ie> wrote:
> True, but the most recent version where that's actually possible is
> PostgreSQL 8.0, which was released in early 2005.

It just occurred to me that the main historic reason for the single
user mode advice was the lack of virtual XIDs. The commit that added
the xidStopLimit behavior (commit 60b2444cc3) came a couple of years
before the introduction of virtual transaction IDs (in commit
295e63983d). AFAICT, the advice about single-user mode was added at a
time where exceeding xidStopLimit caused the system to grind to a halt
completely -- even trivial SELECTs would have failed once Postgres 8.1
crossed the xidStopLimit limit.

It seems as if the advice about single user mode persisted for no
great reason at all. Technically there were some remaining reasons to
keep it around (like the truncation thing), but overall these
secondary reasons could have been addressed much sooner if somebody
had thought about it.

-- 
Peter Geoghegan