We are happy to announce that two new storage systems went into operation!
All HSUper users have now access to an additional file system designed specifically for backups. This new storage system, which can be found using the $BACKUP environment variable, built on top of ZFS, provides excellent data resilience and automatic snapshots of stored files. A key feature of this backup storage configuration is its ability to retain deleted data for a period of 30 days, from once per night created snapshots. For the past two years, monthly snapshots are kept. This provides an additional layer of protection for research data. Note that HSUper users are responsible for backup of their own data. The backup storage must be used manually, no automatic backup service is provided for other file systems. The backup storage has a netto capacity of 1300 TB on the hard disk drives. Thanks to automatic compression, we expect to be able to store up to 2000 TB of data.
Users of the Interactive Scientific Computing Cloud (ISCC) can now enjoy improved performance, thanks to our new flash-based dedicated storage system. This upgrade is designed for running a very large number of virtual machines (VMs) in parallel. It significantly accelerates virtualization management, making tasks such as cloning a VM remarkably faster and more efficient. Cloning a VM now takes only seconds, instead of several hours or even days. The hardware is based on fast solid-state drives and provides a usable raw capacity of ca. 380 TB, plus compression and deduplication. Our benchmarks measure an 11.5x times increased R/W-bandwidth for file operations inside VMs, compared to the previous storage system.
