On Tue, Sep 13, 2005 at 11:56:03AM +0100 or thereabouts, William Hay wrote:
> > On Mon, Sep 12, 2005 at 09:56:21AM +0100 or thereabouts, William Hay wrote:
> > > You wrote:
> > > > Sites should only appear on this map only if they are running
> > > > LCG2_6_0 which added latitude and long to the schema.
> > > >
> > > BTW is there a reason why these figures were added to the schema rather
> > > than just requesting sites to add a LOC record (RFC1876) to the DNS
> > > for an approprate host?
> >
> > There are not actually any hosts in the schema, only endpoints. Plus if
> > I want to find the physical location of some service end points I would
> > have to join across both DNS and Glue Schemas.
> >
> Like you have to do anyway with network location which is probably
> more useful. Also the coordinates in the schema aren't associated with
> the endpoints AFAICT but with the fictional entity called a site which
> doesn't necessarily have a single physical location.
The endpoinits objects contain a foreign key of the sitename. Joining
on tree of course a weird thing to do.
> Hosts usually can
> be said to have such a location. UCL-CCC will probably be acquiring
> some new nodes which will be on the other side of UCL-HEP from where
> the current nodes are so the site location will be a bit dispersed.
>
> William Hay, UCL-CCC System Administrator, Information Systems, EISD, UCL
--
Steve Traylen
[log in to unmask]
http://www.gridpp.ac.uk/
|