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

From: PG Bug reporting form <noreply(at)postgresql(dot)org>
To: pgsql-bugs(at)lists(dot)postgresql(dot)org
Cc: maxim(dot)boguk(at)gmail(dot)com
Subject: BUG #18644: ALTER PUBLICATION ... SET (publish_via_partition_root) wrong/undocumented behavior.
Date: 2024-10-01 13:50:17
Message-ID: 18644-6866bbd22178ee16@postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

The following bug has been logged on the website:

Bug reference: 18644
Logged by: Maxim Boguk
Email address: maxim(dot)boguk(at)gmail(dot)com
PostgreSQL version: 15.8
Operating system: Ubuntu
Description:

Hi,

When ALTER PUBLICATION ... SET (publish_via_partition_root) executed on the
existing logical replication with data
(following ALTER SUBSCRIPTION ... REFRESH PUBLICATION), the database start
copy whole partitioned table from start (thus breaking existing logical
replication).
What's worse - I didn't found any way get out of such situation less than
redo all multi-TB logical replication from blank db.

Also ALTER SUBSCRIPTION ... REFRESH PUBLICATION (copy_data=false) - cannot
be used as workaround because it lead to loose any changes in partitioned
table between run ALTER PUBLICATION and ALTER SUBSCRIPTION.

Afterthought this behavior not surprising at all but I think better to
document it somewhere (or even better disable ALTER PUBLICATION ... SET
(publish_via_partition_root) for any publication with existing subscriptions
because it will break things for sure).

After I look into pub/sub code - I feel it will be very complicated task to
make it work correctly.

--
Maxim

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Tom Lane 2024-10-01 16:43:46 Re: BUG #18643: EXPLAIN estimated rows mismatch
Previous Message Takeshi Ideriha 2024-10-01 13:37:47 Re: BUG #18641: Logical decoding of two-phase commit fails with TOASTed default values