Re: Dumping roles improvements?

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: Andrew Dunstan <andrew(at)dunslane(dot)net>
Cc: Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Dumping roles improvements?
Date: 2011-10-11 16:47:39
Message-ID: 201110111647.p9BGld012062@momjian.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Andrew Dunstan wrote:
>
>
> On 10/11/2011 12:40 PM, Bruce Momjian wrote:
> > Josh Berkus wrote:
> >> It occurs to me that we could really use two things to make it easier to
> >> move copies of database stuff around:
> >>
> >> pg_dump -r, which would include a CREATE ROLE for all roles needed to
> >> restore the database (probably without passwords), and
> >>
> >> pg_dumpall -r --no-passwords which would dump the roles but without
> >> CREATE PASSWORD statements. This would be useful for cloning databases
> >> for use in Dev, Test and Staging, where you don't what to copy the md5s
> >> of passwords for possible cracking.
> > What would this do that pg_dumpall --globals-only doesn't?
> >
>
> As stated, it would not export the passwords.

What is the logic for not dumping passwords but the CREATE ROLE
statement? I don't see how anyone would recognize that behavior as
logical. If you want to add a --no-passwords option to pg_dumpall, that
seems more logical to me.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Simon Riggs 2011-10-11 17:11:26 Re: Overhead cost of Serializable Snapshot Isolation
Previous Message Kevin Grittner 2011-10-11 16:46:05 Re: Overhead cost of Serializable Snapshot Isolation