We've been working to improve the performance and reliability of our data centers in the Sliger and Dirac buildings. This effort includes a major overhaul of our network and a consolidation of resources into the Sliger Data Center.
We are pleased to announce that are upgrading our parallel storage systems for the HPC and Spear. We have procured a new GPFS filesystem, and over the next several months, we will migrate all user and research data to this new system from Panasas and Lustre. After the consolidation, we will retire the older systems.
We are pleased to announce a new service offering at the RCC: low-cost archival storage. This is a general-purpose storage platform that replaces the NoleStor service and provides long term, durable storage for research data.
The RCC Systems Team is working to improve the reliability of our systems and services. In that regard, we have scheduled several maintenance windows over the next two months for various systems.
An issue on our Lustre storage system has caused some data to be irrerovacbly lost. Approximately 12TB (~6.5% of total storage) was affected.
Two events may affect the HPC and other RCC systems in the coming weeks: Chiller Maintenance and Hurricane Irma.
FSU is at increased risk of power outage during August. If an outage occurs, RCC staff will turn off some systems in order to prevent damage.
We've recently purchased and implemented 64 new Lenovo nodes, adding 1,652 new Intel cores to our HPC cluster. This brings the theroetical peak performance up to 264.6 TeraFlops.
We have upgraded MATLAB to the latest version, 2017a. The new version is now available on HPC and Spear.
As part of the maintenance upgrade, we have upgraded many software packages on the HPC to newer versions. Here is the complete list.