Horton

Re: use CREATE DATABASE STRATEGY = FILE_COPY in pg_upgrade

Nathan Bossart <nathandbossart@gmail.com>

View thread View on postgr.es
From: Nathan Bossart <nathandbossart@gmail.com>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Matthias van de Meent <boekewurm+postgres@gmail.com>, Dilip Kumar <dilipbalaut@gmail.com>, Ranier Vilela <ranier.vf@gmail.com>, pgsql-hackers@postgresql.org
Date: 2024-06-19T13:37:17Z
Lists: pgsql-hackers 
On Wed, Jun 19, 2024 at 09:17:00AM -0400, Robert Haas wrote:
> OK, I have a (probably) stupid question. The comment says:
> 
> + * In binary upgrade mode, we can skip this checkpoint because neither of
> + * these problems applies: we don't ever replay the WAL generated during
> + * pg_upgrade, and we don't concurrently modify template0 (not to mention
> + * that trying to take a backup during pg_upgrade is pointless).
> 
> But what happens if the system crashes during pg_upgrade? Does this
> patch make things worse than they are today? And should we care?

My understanding is that you basically have to restart the upgrade from
scratch if that happens.  I suppose there could be a problem if you try to
use the half-upgraded cluster after a crash, but I imagine you have a good
chance of encountering other problems if you do that, too.  So I don't
think we care...

-- 
nathan

Commits

  1. Use CREATE DATABASE ... STRATEGY = FILE_COPY in pg_upgrade.

  2. Use TRUNCATE to preserve relfilenode for pg_largeobject + index.

  3. initdb: When running CREATE DATABASE, use STRATEGY = WAL_COPY.