Pica provides project storage and is comprised by 7 Hitachi HNAS cluster nodes and 3 HUS150 controllers providing roughly 5.5 PB of storage.
Lupus provides storage for Irma and is a Dell IEEL (Intel Enterprise Edition for Lustre) storage system. The system includes 9 Dell PowerEdge R630 servers configured as follows:
- 1 management server
- 2 meta data servers working as a one high availability cluster. The cluster is connected to one Dell PowerVault MD3420 storage array.
- 6 object storage servers connected to 6 Dell PowerVault MD3460 storage arrays it the way ensuring redundancy for all the elements of the system.
Lupus is running Lustre file system version 2.5. It supports data striping over multiple nodes and disk targets and can give a summary read/write speed up to 25 Gbit per second. Total capacity is roughly 1.1 PB.
Castor is a custom built storage system running GlusterFS dedicated to Bianca. The system consists of 54 Huawei 5288 V3 servers, each server is equipped with 36 x 3TB SATA disks working as one logical volume (with redundancy) and providing 109TB raw disk space per one server.
This gives about 5,7 PB raw disk space in total.
Each storage server is connected to network with 2 x 40 Gb/s Ethernet links working as one aggregated link 80 Gb/s.
Rackham's storage system is a DDN (DataDirect Networks) EXAScaler filesystem based on the ES14KX platform. Crex uses 420 6TB NL-SAS drives and 24 300GB SAS drives for metadata storage. The total volume is 2 PB, with 1 PB reserved for SciLifeLab, 0.5 PB reserved for SNIC projects, and 0.5 PB for UPPMAX use. The filesystem is Lustre, a highly scalable filesystem common in HPC.