This has not really got as far as tier2s yet but according to "Installed capacity document you have" ./installed-hepspec06-wlcg-cpu.pl -v -s UKI-SOUTHGRID-CAM-HEP GlueSiteName: UKI-SOUTHGRID-CAM-HEP GlueSiteUniqueId: UKI-SOUTHGRID-CAM-HEP ClusterUniqueID: serv03.hep.phy.cam.ac.uk GlueSubCluster: serv03.hep.phy.cam.ac.uk ClusterUniqueID: serv07.hep.phy.cam.ac.uk GlueSubCluster: serv07.hep.phy.cam.ac.uk si00=2700 log=152 hs2006=102;alice=0.100000;atlas=0.100000;cms=0;lhcb=0.100000;other=0.700000 i.e 102 HepSpec2006 units. * 4 to get KSI00 units. You should ignores the shares for now, they only make sense really once you publish a share and you are waiting on yaim for that. http://www.sysadmin.hep.ac.uk/svn/grid-monitoring/trunk/installed-capacity/cpu/ for the script. If those numbers are correct then chances over the next few weeks I hope we will fix all the monitoring to conform as it were. Steve On Tue, Apr 7, 2009 at 7:24 PM, Burke, S (Stephen) <[log in to unmask]> wrote: > LHC Computer Grid - Rollout >> [mailto:[log in to unmask]] On Behalf Of Santanu Das said: >> That make does sense but does it explain why "CE Total CPU" >> is 152 and 184 at the same time > > Your serv07 subcluster publishes 152 physical CPUs so it takes that > value. serv03 publishes 0 so it looks at the CE which has TotalCPUs=184. > How many CPUs do you actually have? > >> or the "CE Free CPU" is 87 and 89? > > Again, your serv03 CEs are currently showing 75 CPUs free and serv07 has > 79 (all the CEs for each head node have the same values). If that > doesn't correspond to something reasonable you need to check your info > provider. > >> Any idea how GlueSubClusterPhysicalCPUs values are actually >> being returned? > > What do you mean by "returned"? > > Stephen > -- > Scanned by iCritical. > -- Steve Traylen