Horton

Re: A recent message added to pg_upgade

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Peter Smith <smithpb2250@gmail.com>, Kyotaro Horiguchi <horikyota.ntt@gmail.com>, bharath.rupireddyforpostgres@gmail.com, alvherre@alvh.no-ip.org, pgsql-hackers@lists.postgresql.org
Date: 2023-11-11T00:16:22Z
Lists: pgsql-hackers 
On Fri, Nov 10, 2023 at 03:27:25PM +0530, Amit Kapila wrote:
> I don't think this comment is correct because there won't be any apply
> activity on the new cluster as after restoration subscriptions should
> be disabled. On the old cluster, I think one problem is that the
> origins may move forward after we copy them which can cause data
> inconsistency issues. The other is that we may not prefer to generate
> additional data and WAL during the upgrade. Also, I am not completely
> sure about using the word 'corruption' in this context.

What is your suggestion here?  Would it be better to just aim for
simplicity and just say that we don't want it to run because "it can
lead to some inconsistent behaviors"?
--
Michael

Commits

  1. Fix the handling of two GUCs during upgrade.

  2. Rework how logirep launchers are stopped during pg_upgrade

  3. Prohibit max_slot_wal_keep_size to value other than -1 during upgrade.

  4. Prevent startup of logical replication launcher during pg_upgrade

  5. Migrate logical slots to the new node during an upgrade.