Horton

Re: pg15b2: large objects lost on upgrade

Justin Pryzby <pryzby@telsasoft.com>

View thread
From: Justin Pryzby <pryzby@telsasoft.com>
To: Robert Haas <robertmhaas@gmail.com>
Cc: pgsql-hackers@postgresql.org, Bruce Momjian <bruce@momjian.us>, Shruthi Gowda <gowdashru@gmail.com>
Date: 2022-07-07T17:10:19Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Revert recent changes to 002_pg_upgrade.pl.

  2. Revise test case added in 43746996399541ecb5c7b188725a5f097c15ceae.

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

  4. Preserve relfilenode of pg_largeobject and its index across pg_upgrade.

  5. Have VACUUM warn on relfrozenxid "in the future".

  6. Set relfrozenxid to oldest extant XID seen by VACUUM.

  7. pg_upgrade: Preserve relfilenodes and tablespace OIDs.

  8. Perform a lot more sanity checks when freezing tuples.

Attachments

On Wed, Jul 06, 2022 at 08:25:04AM -0400, Robert Haas wrote:
> On Wed, Jul 6, 2022 at 7:56 AM Justin Pryzby <pryzby@telsasoft.com> wrote:
> > I'm looking into it, but it'd help to hear suggestions about where to put it.
> > My current ideas aren't very good.
> 
> In main() there is a comment that begins "Most failures happen in
> create_new_objects(), which has just completed at this point." I am
> thinking you might want to insert a new function call just before that
> comment, like remove_orphaned_files() or tidy_up_new_cluster().
> 
> Another option could be to do something at the beginning of
> transfer_all_new_tablespaces().

That seems like the better option, since it has access to the custer's
filenodes.

I checked upgrades from 9.2, upgrades with/out vacuum full, and upgrades with a
DB tablespace.

Maybe it's a good idea to check that the file is empty before unlinking...

-- 
Justin