On Tue, 26 Oct 2004, Steve Traylen wrote:
> On Tue, Oct 26, 2004 at 10:56:58PM +0100 or thereabouts, John Gordon wrote:
> > The HEPiX profiles contain exits for groups as well as the user. The idea
> > is that an experiment person maintains the group profile, setting the
> > group parts of the environment. This works on the UI where the unix group
> > means something although it doesn't translate to the WN. The sysadmin
> > isn't really involved.
> This is like the case today where Steve used --config-vo option to point
> to config file. This being maintained by a bunch of atlas users is
> about the best option today. You could even generate it out of the
> info system.
One implication of the thin UI model is surely that the configuration
*will* end up being maintained by individual end-users on their laptops,
rather than the the local Gridmeister; so today's case isn't atypical.
The hard bit is getting the data in the first place or hearing about
updates: should this be done by experiments(VOs) directly? Atlas could
easily post an atlas/edg_wl_ui.conf on their website for each UI to pull
down in a cron job overnight, for example.
Incidentally, what's the difference between the LBAddresses in the
VO/edg_wl_ui.conf - which can be totally different between VOs - and
LoggingDestination in edg_wl_ui_cmd_var.conf, which is the same for
everyone?
Henry
--
Dr. Henry Nebrensky [log in to unmask]
http://www.brunel.ac.uk/~eesrjjn
"The opossum is a very sophisticated animal.
It doesn't even get up until 5 or 6 p.m."
|