From: | Peter Smith <smithpb2250(at)gmail(dot)com> |
---|---|
To: | PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
Subject: | PG Docs for ALTER SUBSCRIPTION REFRESH PUBLICATION - copy_data option |
Date: | 2021-06-25 03:49:48 |
Message-ID: | CAHut+PtXzBV5UVU+6YDEwOcEkQdajVAyK4A6L5EMJvL9grkFsw@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi.
(For brevity, in this mail I refer to "ALTER SUBSCRIPTION sub REFRESH
PUBLICATION" as "ASRP")
--
The PG Docs for ASRP WITH (copy_data = true), says "(Previously
subscribed tables are not copied.)" [1].
I thought this rule meant that only tables which got added by ALTER
PUBLICATION pubname ADD TABLE ... [2] after the CREATE SUBSCRIPTION
would be affected by the copy_data.
But I recently learned that when there are partitions in the
publication, then toggling the value of the PUBLICATION option
"publish_via_partition_root" [3] can also *implicitly* change the list
published tables, and therefore that too might cause any ASRP to make
use of the copy_data value for those implicitly added
partitions/tables.
It seems a bit too subtle.
I was wondering if this should be made more obvious by a note added to
the PG Docs for the ASRP [1]. e.g. "Previously subscribed tables are
not copied. Note: Tables may also be newly subscribed by changes to
the publish_via_partition_root option [link]"
Or perhaps, the "publish_via_partition_root option" Docs [3] should
say something. e.g. "Note: Changing this option can affect the ASRP
copy_data [link].
Thoughts?
-----
[1] https://www.postgresql.org/docs/devel/sql-altersubscription.html
[2] https://www.postgresql.org/docs/devel/sql-alterpublication.html
[3] https://www.postgresql.org/docs/devel/sql-createpublication.html
Kind Regards,
Peter Smith.
Fujitsu Australia
From | Date | Subject | |
---|---|---|---|
Next Message | Amit Kapila | 2021-06-25 03:53:44 | Re: subscription/t/010_truncate.pl failure on desmoxytes in REL_13_STABLE |
Previous Message | Amit Kapila | 2021-06-25 03:20:03 | Re: Decoding speculative insert with toast leaks memory |