Horton

Re: allow_in_place_tablespaces vs. pg_basebackup

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: Thomas Munro <thomas.munro@gmail.com>, Kyotaro Horiguchi <horikyota.ntt@gmail.com>, pgsql-hackers@postgresql.org
Date: 2023-03-22T17:14:09Z
Lists: pgsql-hackers

Attachments

On Tue, Mar 21, 2023 at 7:59 PM Michael Paquier <michael@paquier.xyz> wrote:
> On Mon, Mar 20, 2023 at 07:56:42AM -0400, Robert Haas wrote:
> > Anyone want to comment on this?
>
> I have not checked the patch in details, but perhaps this needs at
> least one test?

Sure. I was sort of hoping to get feedback on the overall plan first,
but it's easy enough to add a test so I've done that in the attached
version. I've put it in a separate file because 010_pg_basebackup.pl
is pushing a thousand lines and it's got a ton of unrelated tests in
there already. The test included here fails on master, but passes with
the patch.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

  1. Fix pg_basebackup with in-place tablespaces some more.

  2. Fix pg_basebackup with in-place tablespaces.

  3. Allow "in place" tablespaces.

  4. Avoid including tablespaces inside PGDATA twice in base backups

  5. Tablespaces. Alternate database locations are dead, long live tablespaces.