Re: Question on any plans to use the Create Server/Create blink_ Mapping to provide Logical Replication Subscriptions the user/password in an encrypted manner - Mailing list pgsql-hackers

From Ashutosh Bapat
Subject Re: Question on any plans to use the Create Server/Create blink_ Mapping to provide Logical Replication Subscriptions the user/password in an encrypted manner
Date
Msg-id CAExHW5v3LgDHSNYOHaJMbDy1ytYv8jrCpwE=iCdcUFTt3X5nKw@mail.gmail.com
Whole thread Raw
List pgsql-hackers
On Tue, Jul 15, 2025 at 7:28 PM Vitale, Anthony, Sony Music
<anthony.vitale@sonymusic.com> wrote:
>
> Hello All
>
>
>
> Postgresql dblinks and dblink_fdw allow for the use of Server and user mapping to be able to store the user/password
ofa connection and save it in an encrypted manner. 
>
>
>
> Logical replication subscription syntax regarding connection info allows for the user/password to be supplied within
thesubscription ddl. 
>
>
>
> And the Subscription connection info is visible via the pg_subscription.subconninfo column, which can contain
plain-textpasswords, is intentionally restricted. Only the pg_read_all_settings role, superusers, and the owner of the
subscriptioncan SELECT from this column. 
>
>
>
> In a dblink the connection info can be provided via the same connection parameters as allowed by the logical
subscriptionsyntax, however it is allowed to use a Created ServerName with a user mapping in the connections. 
>
>
>
> I am not familiar on what it would take to allow logical subscriptions to use User Server/Mapping logic as the dblink
extensionallows but if it where possible then this would assure that only the role creating the User server/mapping can
setthe connection user/password and then it can be totally hidden from prying eyes. 
>
>
>
> I was wondering if this is within any plan to implement in future releases.
>

Please check if the proposal at [1] suits your needs.

[1] https://www.postgresql.org/message-id/149ff9264db27cdf724b65709fbbaee4bf316835.camel%40j-davis.com

--
Best Wishes,
Ashutosh Bapat



pgsql-hackers by date:

Previous
From: Malthe
Date:
Subject: Composite types for updatable views
Next
From: Rintaro Ikeda
Date:
Subject: Re: Suggestion to add --continue-client-on-abort option to pgbench