Re: Feature request: permissions change history for auditing

From: Andrew Dunstan <andrew(at)dunslane(dot)net>
To: Thom Brown <thombrown(at)gmail(dot)com>
Cc: Glyn Astill <glynastill(at)yahoo(dot)co(dot)uk>, pgsql-hackers(at)postgresql(dot)org
Subject: Re: Feature request: permissions change history for auditing
Date: 2009-11-30 14:00:01
Message-ID: 4B13CFE1.2060602@dunslane.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Thom Brown wrote:
> 2009/11/30 Glyn Astill <glynastill(at)yahoo(dot)co(dot)uk
> <mailto:glynastill(at)yahoo(dot)co(dot)uk>>
>
> --- On Mon, 30/11/09, Thom Brown <thombrown(at)gmail(dot)com
> <mailto:thombrown(at)gmail(dot)com>> wrote:
>
> > As far as I am aware, there is no way to tell when a
> > user/role was granted permissions or had permissions
> > revoked, or who made these changes. I'm wondering if
> > it would be useful for security auditing to maintain a
> > history of permissions changes only accessible to
> > superusers?
>
> I'd have thought you could keep track of this in the logs by
> setting log_statement >= ddl ?
>
> I'm pretty sure this is a feature that's not wanted, but the
> ability to add triggers to these sorts of events would surely make
> more sense than a specific auditing capability.
>
>
> I concede your suggestion of the ddl log output. I guess that could
> then be filtered to obtain the necessary information.
>
>

This could probably be defeated by making the permissions changes in a
stored function. Or even a DO block, I suspect, unless you had
log_statement = all set.

I do agree with Glyn, though, that making provision for auditing one
particular event is not desirable.

cheers

andrew

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2009-11-30 14:03:21 Re: Patch: Remove gcc dependency in definition of inline functions
Previous Message Thom Brown 2009-11-30 13:46:00 Re: Feature request: permissions change history for auditing