Re: Allowing TRUNCATE of FK target when session_replication_role=replica - Mailing list pgsql-hackers

From Euler Taveira
Subject Re: Allowing TRUNCATE of FK target when session_replication_role=replica
Date
Msg-id bbcb9163-0b3e-4d44-a990-40881d96adcb@app.fastmail.com
Whole thread Raw
In response to Re: Allowing TRUNCATE of FK target when session_replication_role=replica  (Hannu Krosing <hannuk@google.com>)
List pgsql-hackers


On Tue, Oct 31, 2023, at 3:21 PM, Hannu Krosing wrote:
One thing though re:
> The former is true but the latter is not. Logical replication requires
> wal_level = logical. That's also true for skipping FSM.

wal_level=logical is only needed *at provider* side, at least when
running pglogical.

It is not a requirement for the subscriber. However, it increases the
complexity for a real scenario (in which you set up backup and sometimes
additional physical replicas) because key GUCs require a restart.


--
Euler Taveira

pgsql-hackers by date:

Previous
From: Nathan Bossart
Date:
Subject: Re: CRC32C Parallel Computation Optimization on ARM
Next
From: Matthias van de Meent
Date:
Subject: btree: implement dynamic prefix truncation (was: Improving btree performance through specializing by key shape, take 2)