LHC Computer Grid - Rollout
> [mailto:[log in to unmask]] On Behalf Of Dimitris Zilaskos
said:
> Yes, that was my initial reaction too when I first saw it,
> but thinking
> of it later, I thought why not? It happens for all kinds of grid
> services already, why this should be an exception?
What do you mean by it happening for many services? I would hope that
it's unusual for the same service to be published multiple times - it
may cause problems for exactly the sort of reason discussed here, i.e.
clients don't expect it and may not handle it properly. We do have
multiple CE headnodes all feeding the same farm but publishing the
cluster/subcluster information multiple times, but that also cause
problems and Steve Traylen has been working on improving that (should be
released soon I think).
Stephen
--
Scanned by iCritical.
|