Re: BUG #18644: ALTER PUBLICATION ... SET (publish_via_partition_root) wrong/undocumented behavior.

From: Maxim Boguk <maxim(dot)boguk(at)gmail(dot)com>
To: Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>
Cc: pgsql-bugs(at)lists(dot)postgresql(dot)org
Subject: Re: BUG #18644: ALTER PUBLICATION ... SET (publish_via_partition_root) wrong/undocumented behavior.
Date: 2024-10-04 14:07:26
Message-ID: CAK-MWwRd_XXXcCdgY6u9W4tHD+Jz3SiMUbN1WYZrCnOoVzYsLg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

> Can you tell us what additional information you want to document other
> than what is already documented in CREATE PUBLICATION [1] related to
> this parameter?

At least put a warning that alter publication set
(publish_via_partition_root) will unrecoverable break any existing
logical replication subscribers if there at least one partitioned
table with data in replication set,
so alter publication set (publish_via_partition_root) should be run
only before any subscription activation.
Ideally - make this use case actually working (but looking into the
code - it would be very complicated I feel).

>>It would be useful if you can create a small test case to show the exact problem and what is your usecase for the same?
Usecase - after initial load of logical replica I decided that on the
replica I better split future data into weekly partitions due huge
size (instead of monthly partitions on the master/publisher)
exactly the case for "alter publication set (publish_via_partition_root)".

My main issues with this case - there is no way to fix this problem if
it happened less than reloading whole logical replication from blank.

--
Maxim Boguk
Senior Postgresql DBA

Phone UA: +380 99 143 0000
Phone AU: +61 45 218 5678

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Cameron Vogt 2024-10-04 18:26:03 Re: PostgreSQL 17 Segmentation Fault
Previous Message Даниил 2024-10-04 14:00:54 Re: BUG #18646: The problem with the installer