Re: Move bki file pre-processing from initdb to bootstrap - Mailing list pgsql-hackers

From Andres Freund
Subject Re: Move bki file pre-processing from initdb to bootstrap
Date
Msg-id 20230918231333.odexzr4r4elrx67k@awork3.anarazel.de
Whole thread Raw
In response to Re: Move bki file pre-processing from initdb to bootstrap  (Peter Eisentraut <peter@eisentraut.org>)
List pgsql-hackers
Hi,

On 2023-09-01 14:59:57 +0200, Peter Eisentraut wrote:
> On 01.09.23 14:37, Tom Lane wrote:
> > Krishnakumar R <kksrcv001@gmail.com> writes:
> > > This patch moves the pre-processing for tokens in the bki file from
> > > initdb to bootstrap. With these changes the bki file will only be
> > > opened once in bootstrap and parsing will be done by the bootstrap
> > > parser.
> > 
> > You haven't provided any iota of evidence why this would be an
> > improvement.
> 
> I had played with similar ideas in the past, because it would shave some
> time of initdb, which would accumulate noticeably over a full test run.
> 
> But now with the initdb caching mechanism, I wonder whether this is still
> needed.

I think it's still relevant - it's not just our own test infrastructure that
runs a lot of initdbs, it's also lots of projects using postgres.


The main reason I'd like to move this infrastructure to the backend is that I
really would like to get rid of single user mode. It adds complications all
over, it's barely tested, pointlessly hard to use. I wrote a rough prototype
of that a while back:
https://postgr.es/m/20220220214439.bhc35hhbaub6dush%40alap3.anarazel.de

Greetings,

Andres Freund



pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: XLog size reductions: smaller XLRec block header for PG17
Next
From: Michael Paquier
Date:
Subject: Re: dikkop seems unhappy because of openssl stuff (FreeBSD 14-BETA1)