Hardware

This is a hardware specification designed for vertical scalability. It is optimised for those who plan to built an initial LogMan.io cluster with the lower possible cost yet with the possibility to add more hardware gradually as the cluster grows. This specification is also fully compatible with the horizontal scalability strategy, which means adding one or more new server node to the cluster.

Specification

  • Chasis: 2U
  • CPU: 1x AMD EPYC 16 Cores
  • CPU Slots: 1 occupied, 2 available
  • RAM: 256GB DDR4 3200
  • Data SSD: 2x 2TB SSD NVMe, PCIe 3.0+, no RAID
  • Data HDD: 3x 18TB SATA 2/3+ or SAS 1/2/3+, 6+ Gb/s, 7200 rpm, hot-swap, HBA (no RAID)
  • OS HDD: 2x 256MB SSD SATA 2/3+, HBA (no RAID)
  • NIC: 2x 10Gbps Ethernet
  • HBA controller card (or IT mode) for Data HDDs
  • Power supply: Redundant 920W
  • IMPI (Remote Platform Management)

Remark: 3U and 4U variants are also available with 16 respective 24 drive bays.

Remark: RAID is implemented in software (OS).

Vertical scalability options

  • Add one more CPU (2 CPUs in total)
  • Add RAM up to 512GB
  • Add up to 9 additional Data HDDs, total 12 drive bays, max. 198 TB space using 12x 18 TB HDDs in RAID5

We recommend to build a cluster of three servers. It is also possible to start with one or two servers but such setup is not providing a state-of-art availability.

Recommended geo-cluster setup is six servers but variants with four, five or more are possible as well.

It is recommended to preffer even numbers of servers in the cluster. In case of two geographically separated locations, smaller “arbiter” site should be build. The arbiter can be implemented as a virtual server with much smaller specifications.


Last update: Sep 2022