Re: Issues with ON CONFLICT UPDATE and REINDEX CONCURRENTLY

From: Mihail Nikalayeu <mihailnikalayeu(at)gmail(dot)com>
To: Álvaro Herrera <alvherre(at)kurilemu(dot)de>
Cc: Alexander Lakhin <exclusion(at)gmail(dot)com>, Michael Paquier <michael(at)paquier(dot)xyz>, Noah Misch <noah(at)leadboat(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Issues with ON CONFLICT UPDATE and REINDEX CONCURRENTLY
Date: 2025-12-08 11:34:00
Message-ID: CADzfLwXDS7QfJ6=vbWBQ2w-pATwR+3x60iRPXfj7eGWd21FemQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hello, Álvaro!

On Mon, Dec 8, 2025 at 10:58 AM Álvaro Herrera <alvherre(at)kurilemu(dot)de> wrote:
> Rereading this -- did you mean to propose that a possible fix was to
> remove the "invalid arbiter index list" error? I had understood
> something different.

Yes, it was the initial idea.

> Your idea downthread of changing the way that check works (so that we
> don't throw an error in this case, but we continue to double-check that
> the arbiter list is sensible) sounds reasonable to me. Do you want to
> propose a specific check for it?

I think the next logic is correct:
* for each IS indisvalid arbiter in the parent table we should have AT
LEAST ONE compatible indisvalid pair in the partition (we may have
multiple or a few more ready-only)
* for each NOT indisvalid arbiter in parent - nothing is expected
from partition

I'll try to create a patch with such later.

Best regards,
Mikhail.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Dilip Kumar 2025-12-08 11:45:37 Re: Proposal: Conflict log history table for Logical Replication
Previous Message Amit Kapila 2025-12-08 11:29:53 Re: POC: enable logical decoding when wal_level = 'replica' without a server restart