Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of Steve Traylen said:
> But it doesn't actually matter. So long as the SC that is linked with
> the cluster is
> as representative as the single one you used to publish. The
> unlinked ones will
> never be found by the WMS only the linked one.
Dug appeared to be talking about having both linked:
> >> Correct me if I am wrong but you can publish an
> inhomogeneous cluster
> > as two separate sub-clusters from the same CE.
although perhaps by CE he meant head node? (I assumed that Santanu meant
the GlueCE, i.e. having multiple queues, but it can be hard to tell!) Of
course it's true that a subcluster not attached to any (glue) CE will
never be used by the WMS - but actually in general I'd think that's the
sort of thing we'd like to flag as an error in gstat ...
Stephen
--
Scanned by iCritical.
|