The HPC team is busy testing a prototype SLURM cluster. This will replace both the hex and hpc clusters, although it is likely that the hex infrastructure will be incorporated into the new cluster. The time frame for this is a few months away as ICTS is currently building its new data center, see below. The HPC team made the decision to move away from Torque PBS as it has a very old code base and the scheduler, Maui, is no longer being maintained. SLURM is free, is very widely used in the professional HPC community and has an extremely good support community.
New Data Center
ICTS is building a new data centre at 7 On Main which will host eResearch kit such as the new HPC cluster. The data centre will have a standby emergency power generator in-line, dual UPS distribution feeds, hot and cold aisles partitioning and environmental management.
The FHGFS scratch area is now available to HPC users. This is a very high speed, temporary work area. Access to this volatile storage is on request only. User directories have quotas set. Additionally old files are automatically deleted, the higher the quota the lower the retention time. This is to ensure that the working space does not fill up, as users tend to be fairly relaxed about disk space utilization. The current default setting is 100GB with an 80 day expiration limit.