Re: User Quota Implementation

Hi Josh, hi jonah,

On Mon, Jun 13, 2005 at 12:36:12PM -0700, Josh Berkus wrote:
>
> > Don't get me wrong, I think we need tablespace maximums. What I'm
> > looking at is a user/group-based quota which would allow a superuser to
> > grant say, 2G of space to a user or group. Any object that user owned
> > would be included in the space allocation.
> >
> > So, if the user owns three tablespaces, they can still only have a
> > maximum of 2G total. This is where I think it would be wise to allow
> > the tablespace owner and/or superuser to set the maximum size of a
> > tablespace.
>
> Yeah, the problem is that with the upcoming "group ownership" I see
> user-based quotas as being rather difficult to implement unambiguously.
> Even more so when we get "local users" in the future. So I'd only want
> to do it if there was a real-world use case that tablespace quotas
> wouldn't satisfy.

Well, I think in one way jona is right, that I mixed up two things.
Indeed a max size for a tablespace is something different, than a quota.
In my opinion, it makes only sense to use quotas for ressource-owners on
ressources, i.e. tablespaces. To as an example I think about some
tablespace whith a MAXSIZE of 2 GB (that it won't grow until the disk is
full) and a QUOTA of 500 MB for user A on that certain tablespace. In
general (of cause this is only my experience in using quotas in dbms)
you will create different tablespaces for different object kinds/types
i.e. one for indexes, one for dimensions and at least one for the fact
data in a dwh. So to allow users to store their comparable tables in the
appropriate tablespace you'd set up a quota for them.

Regards,
Yann