User mapping security
Hackorum builds and tests every patch posted to the lists, not only commitfest submissions. This is Hackorum's own CI rather than the PostgreSQL project's, and it is still under testing - please report anything that looks wrong.
You can run a PostgreSQL built from this patch straight from Docker, with no checkout and no build:
docker run --rm -p 5432:5432 ghcr.io/hackorum-dev/postgres-patch:t76507psql -h localhost -U postgresBuilt from patchset v2 (message #2), August 17, 2026 at 10:12 AM.
The following documentation comment has been logged on the website:
Page: https://www.postgresql.org/docs/10/postgres-fdw.html
Description:
Regarding the documentation pages
https://www.postgresql.org/docs/10/postgres-fdw.html?origin_team=T02HEPYKQ
and https://www.postgresql.org/docs/10/sql-createusermapping.html
I suppose it should be warned on the pages that foreign credentials with be
stored as simple text and will be available for viewing in pg_user_mappings.
On Tue, Jul 16, 2019 at 02:01:00AM +0000, PG Doc comments form wrote:
The following documentation comment has been logged on the website:
Page: https://www.postgresql.org/docs/10/postgres-fdw.html
Description:Regarding the documentation pages
https://www.postgresql.org/docs/10/postgres-fdw.html?origin_team=T02HEPYKQ
and https://www.postgresql.org/docs/10/sql-createusermapping.htmlI suppose it should be warned on the pages that foreign credentials with be
stored as simple text and will be available for viewing in pg_user_mappings.
I know this is four years old, but the attached patch documents it. I
don't think postgresql-fdw needs it since it relies on user mapping and
discourages passwords in the connection string.
--
Bruce Momjian <bruce@momjian.us> https://momjian.us
EDB https://enterprisedb.com
Only you can decide what is important to you.
Attachments:
usermap.difftext/x-diff; charset=us-asciiDownload+5-1
Greetings,
* Bruce Momjian (bruce@momjian.us) wrote:
On Tue, Jul 16, 2019 at 02:01:00AM +0000, PG Doc comments form wrote:
The following documentation comment has been logged on the website:
Page: https://www.postgresql.org/docs/10/postgres-fdw.html
Description:Regarding the documentation pages
https://www.postgresql.org/docs/10/postgres-fdw.html?origin_team=T02HEPYKQ
and https://www.postgresql.org/docs/10/sql-createusermapping.htmlI suppose it should be warned on the pages that foreign credentials with be
stored as simple text and will be available for viewing in pg_user_mappings.I know this is four years old, but the attached patch documents it. I
don't think postgresql-fdw needs it since it relies on user mapping and
discourages passwords in the connection string.
A bit on the fence about it ... but I do wonder if we should encourage
use of gssapi and credential delegation now that we support that and
point out that storing passwords isn't required if you're using gssapi.
Thanks,
Stephen
Show quoted text
diff --git a/doc/src/sgml/ref/create_user_mapping.sgml b/doc/src/sgml/ref/create_user_mapping.sgml index 55debd5401..e93bfe48f6 100644 --- a/doc/src/sgml/ref/create_user_mapping.sgml +++ b/doc/src/sgml/ref/create_user_mapping.sgml @@ -92,7 +92,11 @@ CREATE USER MAPPING [ IF NOT EXISTS ] FOR { <replaceable class="parameter">user_ This clause specifies the options of the user mapping. The options typically define the actual user name and password of the mapping. Option names must be unique. The allowed option - names and values are specific to the server's foreign-data wrapper. + names and values are specific to the server's foreign-data + wrapper. Option values, including passwords, are visible in the + <link + linkend="catalog-pg-user-mapping"><structname>pg_user_mapping</structname></link> + system view. </para> </listitem> </varlistentry>
Bruce Momjian <bruce@momjian.us> writes:
On Tue, Jul 16, 2019 at 02:01:00AM +0000, PG Doc comments form wrote:
I suppose it should be warned on the pages that foreign credentials with be
stored as simple text and will be available for viewing in pg_user_mappings.
I know this is four years old, but the attached patch documents it. I
don't think postgresql-fdw needs it since it relies on user mapping and
discourages passwords in the connection string.
This is far too alarmist. It ignores the privilege restrictions that
are built into the pg_user_mappings view. Random users can't see
umoptions.
regards, tom lane
On Thu, Nov 9, 2023 at 05:03:33PM -0500, Tom Lane wrote:
Bruce Momjian <bruce@momjian.us> writes:
On Tue, Jul 16, 2019 at 02:01:00AM +0000, PG Doc comments form wrote:
I suppose it should be warned on the pages that foreign credentials with be
stored as simple text and will be available for viewing in pg_user_mappings.I know this is four years old, but the attached patch documents it. I
don't think postgresql-fdw needs it since it relies on user mapping and
discourages passwords in the connection string.This is far too alarmist. It ignores the privilege restrictions that
are built into the pg_user_mappings view. Random users can't see
umoptions.
True. I wasn't sure how much of an issue it was to expose passwords at
the SQL level, but I will drop the idea.
--
Bruce Momjian <bruce@momjian.us> https://momjian.us
EDB https://enterprisedb.com
Only you can decide what is important to you.