The new IPPP Grid cluster is now available for use and has been in full operation since the beginning of January 2009. The new cluster now provides 1 Million SpecInt2k – a factor of 12 increase on the old cluster. This is a significant improvement in computational power to assist particle physics research through the pheno VO and other grid users.
The new cluster consists of 3 new front end machines and 84 new worker nodes. Using twin-servers, two machines can be packed in a 1U server, providing huge CPU power in a small area. A total of 672 job slots are available to provide the 1 M SpecInt2k – with each worker node consisting of:
Dual processor, quad core providing 8 cores per machine.
Low-power Xeon L5430 for greater power efficiency and lower running costs.
16GB RAM per machine, providing 2GB per core.
Dual bonded gigabit ethernet
0.5TB Hard Disk
Installed with the Scientific Linux 4.7 OS
A new high spec UI is available to allow the preparation, submission and retrieval of jobs and data. The cluster also provides 3 disk servers providing a total of 30TB of usable grid storage.
The management and functionality of the cluster has also improved dramatically. Switchable PDUs and a private IPMI network has been setup for improved service management. The old front end machines have been replaced with new shiny models, and each front-end service is now running as a virtual machine. This allows for great flexibility and helps lower costs with the following services all running as Virtual Machines:
Two CEs (Computing Elements) for Grid job submission and control.
Torque server for the PBS batching system to control job scheduling.
BDII for information handling to report cluster usage.
SE (Storage Element) for storage head node functionality.
Install host for kickstarting and cfengine fabric management.
Mon box for publishing accounting information.
Monitoring box running ganglia and nagios.
