 |
 |
More to the point, the Glue 1.3 standard (say) explicitly notes that the Cluster element can have one or more SubClusters. "SubCluster.UniqueID 1..* A cluster can be decomposed in one or more subclusters:" (There is the question over "UniqueIDs", of course, but since the Clusters all do refer to the same SubClusters, this a weakness in Glue 1.3. In fact, I'm pretty sure this is one of the things that got changed in Glue 2.0?) Sam On 13 March 2015 at 10:26, Andrew Lahiff <[log in to unmask]> wrote: > > That's what I've never understood about glue - it seems to be designed to NOT be able to describe correctly 99% of grid sites, where there is a single cluster made up of distinct groups of worker nodes with similar properties. > > Regards, > Andrew. > > ________________________________________ > From: Testbed Support for GridPP member institutes [[log in to unmask]] on behalf of Stephen Burke [[log in to unmask]] > Sent: Friday, March 13, 2015 10:10 AM > To: [log in to unmask] > Subject: Re: Publishing Issues (Rebus vs. Gstat) > > Testbed Support for GridPP member institutes [mailto:TB- >> [log in to unmask]] On Behalf Of Gareth Roy said: >> As far as I can tell the LDAP we publish is correct > > Err, no. Your GlueCluster svr009.gla.scotgrid.ac.uk is messed up; it has four dependent SubClusters called Viglen, SuperMicro, Interlagos and SandyBridge. That's invalid, you're only supposed to have one SubCluster per Cluster, and also those names are supposed to be Unique IDs and aren't constructed to be unique. The other CEs look OK, at least in that respect, so I'm not sure what happened there. > > Stephen