Re: Dumping roles improvements? - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Dumping roles improvements?
Date
Msg-id 8922.1318394624@sss.pgh.pa.us
Whole thread Raw
In response to Re: Dumping roles improvements?  (Josh Berkus <josh@agliodbs.com>)
Responses Re: Dumping roles improvements?  (Josh Berkus <josh@agliodbs.com>)
Re: Dumping roles improvements?  (Andrew Dunstan <andrew@dunslane.net>)
List pgsql-hackers
Josh Berkus <josh@agliodbs.com> writes:
> The reason I want to have the dependant roles created as part of a
> database dump is so that we can ship around dump files as a single file,
> and restore them with a single command.  This is considerably simpler
> than the current requirements, which are:

> 1. pg_dumpall the roles
> 2. pg_dump the database
> 3. tar both files
> 4. ship file
> 5. untar both files
> 6. psql the role file
> 7. pg_restore the database file

I don't find this terribly convincing.  I can see the rationales for two
endpoint cases: (1) restore these objects into exactly the same
ownership/permissions environment that existed before, and (2) restore
these objects with the absolute minimum of ownership/permissions
assumptions.  The latter case seems to me to be covered already by
--no-owner --no-privileges.  Cases in between those endpoints seem
pretty special-purpose, and I don't want to buy into the assumption that
we should fix them by creating a plethora of --do-it-joshs-way switches.
Can we invent something comparable to the --list/--use-list mechanism,
that can handle a range of use cases with a bit more manual effort?
        regards, tom lane


pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Index only scan paving the way for "auto" clustered tables?
Next
From: Heikki Linnakangas
Date:
Subject: Re: Overhead cost of Serializable Snapshot Isolation