Re: Why are absolute paths considered a security risk?

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Hadley Willan <hadley(dot)willan(at)deeperdesign(dot)co(dot)nz>
Cc: Postgresql General <pgsql-general(at)postgresql(dot)org>
Subject: Re: Why are absolute paths considered a security risk?
Date: 2003-02-26 00:31:21
Message-ID: 25797.1046219481@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Hadley Willan <hadley(dot)willan(at)deeperdesign(dot)co(dot)nz> writes:
> The documentation (7.2.1) mentions that allowing absolute paths when
> creating a db is a security risk and is off by default.
> However, it seems fairly hard to exploit, and I was wondering if anybody
> has any examples of how much of a risk this is?
> Reason I ask is we're considering turning them on in our server and want
> to consider these risks.

The difficulty is that someone who is allowed to create databases (but
isn't necessarily a superuser) will be able to cause the backend to
scribble in any directory that the postgres user has write access to.
The potential damage is somewhat limited since "base/DBOID" gets tacked
onto the user-specified string, and the user has little if any control
over the DBOID part. Still, it's a risk.

regards, tom lane

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Andrew Bartley 2003-02-26 00:59:07 WITHOUT OIDS
Previous Message Cristian Custodio 2003-02-26 00:24:49 OID or lo