(858) 613-1818Quote list 0

4U 36 bay HDD storage server

Capacity tier storage server with 36 hot swap 3.5 inch bays for up to 1 PB raw on HDD, NVMe cache, dual EPYC 9005 or Xeon 6 and 100 GbE or 200 Gb/s ports.

Illustration of a 4U GPU server
Drives
36 x 3.5 inch SAS or SATA HDD, 20 TB to 30 TB each, up to 1 PB raw
Cache
4 x U.2 NVMe for metadata and write cache
CPU
2 x AMD EPYC 9005 or 2 x Intel Xeon 6
Memory
16 or 24 x DDR5 RDIMM, 256 GB to 1 TB
Networking
2 x 100 GbE or 2 x 200 Gb/s, 2 x 25 GbE onboard
Power
Redundant PSUs, about 1.3 kW per node planning figure

Overview

AI projects accumulate data that is read rarely but must be kept: raw corpora, archived checkpoints, logs, rendered frames and backups. The 4U 36 bay server stores it at the lowest cost per terabyte, with up to a petabyte of HDD behind a dual socket host, NVMe cache drives for metadata and writes, and 100 GbE or 200 Gb/s networking.

It is deployed as the capacity tier under a flash layer, as an object store with erasure coding across several nodes, or as a backup target. The host is powerful enough to run software defined storage with compression and deduplication, and PCIe slots allow expansion shelves to be attached when a single chassis is not enough.

Configuration includes 36 enterprise HDDs of 20 TB to 30 TB, four U.2 NVMe cache drives, mirrored boot, 256 GB to 1 TB of memory and redundant power. Plan for about 1.3 kW per node and confirm rack rails and floor loading for the populated weight.

Use cases

  • Capacity tier under flash storage
  • Object storage and archives
  • Backup targets
  • Media and render archives
  • Data lake storage

Specifications

Storage
Drive bays36 x 3.5 inch hot swap, 24 front and 12 rear, SAS or SATA
CapacityUp to about 1 PB raw with 30 TB drives
Cache4 x U.2 NVMe, 3.84 TB to 15.36 TB
Boot2 x M.2 NVMe, mirrored
ControllerSAS HBA for software defined storage or hardware RAID
Compute
CPUs2 x AMD EPYC 9005 series or 2 x Intel Xeon 6
System memory16 or 24 x DDR5 RDIMM, 256 GB to 1 TB
ExpansionPCIe Gen5 slots for additional HBAs and JBOD expansion
Networking
Storage network2 x 100 GbE or 2 x 200 Gb/s ConnectX-7
Frontend2 x 25 GbE onboard
ManagementDedicated 1 GbE BMC port, TPM 2.0
Physical
Form factor4U rackmount, air cooled, rails rated for the loaded weight
Power suppliesRedundant 1,600 W class PSUs, 200 to 240 V input
Planning powerAbout 1.3 kW per node fully populated

Figures describe the system class. Exact configuration, weights and power draw are confirmed on the quotation.

What's included

  • Rails, power cords and cable management arm
  • Burn in and acceptance test report
  • Operating system imaging and firmware baseline on request
  • 3 year warranty, options to 5 years with advance replacement
  • Drive health report with every drive serial

Compliance

This product is not a data center GPU system and does not require NVIDIA end user registration. Standard export documentation applies. Read about export compliance.

Compare it with the next class up or down

The nearest classes we quote are the 2U 24 bay all NVMe storage server and the 2U JBOD or flash expansion shelf. If neither is quite right, tell us the workload and we specify the system between them.

Talk to an engineerRequest a quotation