Introduction
The University of Manchester's GridPP Tier2 resource is provided by the University's High Energy Physics Group. The cluster is primarily used by LHC experiments and data analysis via the EGEE/LCG grid.
System Configuration
The cluster has 1800 cores [900 Dell PowerEdge nodes with 2 x 2.8 GHz Xeon processors, 2x250 GB hdd, dual 1 Gb Ethernet connections] and over 400 terabytes of storage space. Each node has 4 GB of main memory shared between processors. The cluster runs the Scientific Linux 4.4 operating system.
Users' home directories are local to each worker node where their jobs run. There is up to 50GB of temporary space available in the /home area on the worker nodes and a further 20GB in /scratch. Our cluster does not support parallel jobs. Jobs cannot be submitted to the site with a maximum duration longer than 72 hours. More information about the hardware configuration can be found here.
Middleware, compilers and other software
The gLite grid middleware version 3.1 is installed to allow users to manage jobs and transfer files across all NGS resources with a single sign-on. Based on Globus Toolkit 2, gLite affords the usual NGS facilities for job management and file transfer. The usual suite of GCC compilers, JDK etc are provided by the base Scientific Linux 4 operating system (based on RedHat Enterprise Linux 4).
The cluster comprises two sub-clusters. Jobs can be submitted directly to both computing element (CE) head nodes, or via external resource brokers. The CE hosts are ce01.tier2.hep.manchester.ac.uk and ce02.tier2.hep.manchester.ac.uk and the contact string to submit to the ce01 cluster is
ce01.tier2.hep.manchester.ac.uk:2119/jobmanager-lcgpbs
Contact
Any queries regarding the NGS at The University of Manchester should be submitted by email to the NGS support desk: support@grid-support.ac.uk
More Information
