Re: MultiXactId concept underdocumented

From: Andres Freund <andres(at)2ndquadrant(dot)com>
To: Peter Eisentraut <peter_e(at)gmx(dot)net>
Cc: pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: MultiXactId concept underdocumented
Date: 2013-08-08 02:57:24
Message-ID: 20130808025724.GB6350@alap2.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On 2013-08-07 21:25:06 -0400, Peter Eisentraut wrote:
> The new MultiXactId concept appears in a number of user-facing error
> messages, including in the scary context of transaction ID wraparound,
> so it seems kind of important, but it doesn't appear to be documented in
> any user-facing places. Should this be rectified? Maybe some of the
> material from README.tuplock should be moved to the user manual. Maybe
> a less jargony term could be found as well.

Multixacts aren't really new, they've been there since 8.1 or so. It's
just that their use has been a bit expanded and that their internal
format changed a bit.

WRT user facing messages: I think it'd suffice to explain the
anti-wraparound part of the docs that the same holds true for
multixacts. Doing a more detailed explanation seems to require a good
amount of low level explanations and thus seems unlikely to be actually
read.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2013-08-08 03:05:12 Re: 9.4 regression
Previous Message Peter Eisentraut 2013-08-08 01:55:14 timeline signedness