RAL Tier1 Computing Element

WARNING: information not up-to-date

Service Details

The RAL Tier1 runs several computing elements or CEs. These are lcgce02.gridpp.rl.ac.uk, lcgce03.gridpp.rl.ac.uk, lcgce04.gridpp.rl.ac.uk and lcgce05.gridpp.rl.ac.uk.

These feed jobs to a remote Torque batch server currently hosted on csflnx353.rl.ac.uk

SubClusters

There are currently four GlueClusters and GlueSubClusters published by all CEs. These SubClusters are identical in every way other than their values for memory. An additional Cluster and SubCluster is published on lcgce03.gridpp.rl.ac.uk and lcgce05.gridpp.rl.ac.uk to provide a 3000MB queue for Atlas.

(Note that this example only shows lcgce02 for clarity, in practice this query would show all CEs)

 $ ldapsearch -x -H ldap://site-bdii.gridpp.rl.ac.uk:2170 \
       -b 'Mds-vo-name=RAL-LCG2,o=grid' '(objectClass=GlueSubCluster)' \
        GlueSubClusterName GlueHostMainMemoryRAMSize

 # lcgce02.gridpp.rl.ac.uk, lcgce02.gridpp.rl.ac.uk, RAL-LCG2, grid
 dn: GlueSubClusterUniqueID=lcgce02.gridpp.rl.ac.uk,GlueClusterUniqueID=lcgce02
  .gridpp.rl.ac.uk,mds-vo-name=RAL-LCG2,o=grid
 GlueHostMainMemoryRAMSize: 512
 GlueSubClusterName: lcgce02.gridpp.rl.ac.uk

 # 700-lcgce02.gridpp.rl.ac.uk, 700-lcgce02.gridpp.rl.ac.uk, RAL-LCG2, grid
 dn: GlueSubClusterUniqueID=700-lcgce02.gridpp.rl.ac.uk,GlueClusterUniqueID=700
  -lcgce02.gridpp.rl.ac.uk,mds-vo-name=RAL-LCG2,o=grid
 GlueHostMainMemoryRAMSize: 700
 GlueSubClusterName: 700-lcgce02.gridpp.rl.ac.uk

 # 1000-lcgce02.gridpp.rl.ac.uk, 1000-lcgce02.gridpp.rl.ac.uk, RAL-LCG2, grid
 dn: GlueSubClusterUniqueID=1000-lcgce02.gridpp.rl.ac.uk,GlueClusterUniqueID=10
  00-lcgce02.gridpp.rl.ac.uk,mds-vo-name=RAL-LCG2,o=grid
 GlueHostMainMemoryRAMSize: 1000
 GlueSubClusterName: 1000-lcgce02.gridpp.rl.ac.uk

 # 2000-lcgce02.gridpp.rl.ac.uk, 2000-lcgce02.gridpp.rl.ac.uk, RAL-LCG2, grid
 dn: GlueSubClusterUniqueID=2000-lcgce02.gridpp.rl.ac.uk,GlueClusterUniqueID=20
  00-lcgce02.gridpp.rl.ac.uk,mds-vo-name=RAL-LCG2,o=grid
 GlueHostMainMemoryRAMSize: 2000
 GlueSubClusterName: 2000-lcgce02.gridpp.rl.ac.uk

We have subclusters for 512, 700, 1024 and 2048 megabyte resources named lcgce02.gridpp.rl.ac.uk, 700-lcgce02.gridpp.rl.ac.uk, 1000-lcgce02.gridpp.rl.ac.uk and 2000-lcgce02.gridpp.rl.ac.uk respectively.

A job submitted via an RB with JDL containing a requirement of

  Requirements = (other.GlueHostMainMemoryRAMSize>600) ;

would only match the last three sub clusters and consequently be submitted to the batch system requesting 700, 1000 or 2000 megabytes of memory.