Content-Type: multipart/alternative;
 boundary="------------GozTQQNiZu3ZrrHjAz8hGhK5"
Message-ID: <17622f79-117a-4a44-aa8e-0374e53faaf0@gmail.com>
Date: Sat, 6 Dec 2025 08:00:01 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Issues with ON CONFLICT UPDATE and REINDEX CONCURRENTLY
To: =?UTF-8?Q?=C3=81lvaro_Herrera?= <alvherre@kurilemu.de>,
 Mihail Nikalayeu <mihailnikalayeu@gmail.com>
Cc: Michael Paquier <michael@paquier.xyz>, Noah Misch <noah@leadboat.com>,
 PostgreSQL Hackers <pgsql-hackers@postgresql.org>
References: <202512021255.r3ouzsfyxg5r@alvherre.pgsql>
Content-Language: en-US
From: Alexander Lakhin <exclusion@gmail.com>
In-Reply-To: <202512021255.r3ouzsfyxg5r@alvherre.pgsql>
Archived-At: 
 <https://www.postgresql.org/message-id/17622f79-117a-4a44-aa8e-0374e53faaf0%40gmail.com>
Precedence: bulk

This is a multi-part message in MIME format.
--------------GozTQQNiZu3ZrrHjAz8hGhK5
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

Hello Álvaro and Mihail,

02.12.2025 15:07, Álvaro Herrera wrote:
> Thanks, pushed this one after some more editorialization.

I've discovered that despite removing FIXME (in 90eae926a), the error
"invalid arbiter index list" can still be triggered with:
CREATE TABLE pt (a int PRIMARY KEY) PARTITION BY RANGE (a);
CREATE TABLE p1 PARTITION OF pt FOR VALUES FROM (1) to (2) PARTITION BY RANGE (a);
CREATE TABLE p1_1 PARTITION OF p1 FOR VALUES FROM (1) TO (2);
CREATE UNIQUE INDEX ON ONLY p1 (a);
INSERT INTO p1 VALUES (1) ON CONFLICT (a) DO NOTHING;
ERROR:  XX000: invalid arbiter index list
LOCATION:  ExecInitPartitionInfo, execPartition.c:863

The first commit it produced on with this script is bc32a12e0.

Best regards,
Alexander
--------------GozTQQNiZu3ZrrHjAz8hGhK5
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 8bit

<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <div class="moz-cite-prefix">Hello Álvaro and Mihail,<br>
      <br>
      02.12.2025 15:07, Álvaro Herrera wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:202512021255.r3ouzsfyxg5r@alvherre.pgsql">
      <pre wrap="" class="moz-quote-pre">Thanks, pushed this one after some more editorialization.
</pre>
    </blockquote>
    <br>
    I've discovered that despite removing FIXME (in 90eae926a), the
    error<br>
    "invalid arbiter index list" can still be triggered with:<br>
    CREATE TABLE pt (a int PRIMARY KEY) PARTITION BY RANGE (a);<br>
    CREATE TABLE p1 PARTITION OF pt FOR VALUES FROM (1) to (2) PARTITION
    BY RANGE (a);<br>
    CREATE TABLE p1_1 PARTITION OF p1 FOR VALUES FROM (1) TO (2);<br>
    CREATE UNIQUE INDEX ON ONLY p1 (a);<br>
    INSERT INTO p1 VALUES (1) ON CONFLICT (a) DO NOTHING;<br>
    ERROR:  XX000: invalid arbiter index list<br>
    LOCATION:  ExecInitPartitionInfo, execPartition.c:863<br>
    <br>
    The first commit it produced on with this script is bc32a12e0.<br>
    <br>
    Best regards,<br>
    Alexander<br>
  </body>
</html>

--------------GozTQQNiZu3ZrrHjAz8hGhK5--