Case Study: CI TASK achieves 1.7PB+ scalable, high-performance Lustre storage and improved supercomputer availability with Hitachi Vantara

Polish Supercomputer Center CI TASK Stores More Than 1.7PB of Data on Hitachi Storage With the Lustre File System

CI TASK, one of Poland’s largest supercomputer centers, faced rapid data growth and needed a high‑performance, scalable storage solution that would preserve usable capacity and performance even during disk failures. To support the Galera Plus cluster and a single global Lustre namespace, CI TASK selected Hitachi Vantara and deployed Hitachi Unified Storage VM systems and a HUS 150 alongside a Lustre file system front end.

Hitachi Vantara implemented three HUS VM arrays and one HUS 150 with dedicated Lustre metadata and data servers, delivering a SAN‑based Lustre architecture that sustained about 15 GB/s write performance even during disk rebuilds. The deployment now stores slightly over 1.7 PB on 564 × 3 TB SAS‑2 disks, improved availability for more than 10,000 processing cores, reduced rebuild impact on performance, and provides unrestricted licensing and virtualization for easy, petabyte‑scale growth.

CI TASK

Michal Bialoskorski

Administrator, Supercomputer Department