IT-Beratung HALBE, which operates its own data center, wanted to realign its strategy by replacing the conventional storage and virtualization environment based on DSS V7 and VMware ESXi with a new infrastructure. Taking into account a general cost minimization, the goal was to create a future-proof starting position so that competitive services in the IT and hosting area can be still offered to the customers.
The requirements for a suitable storage solution included:- The basis of the planning was to provide low latencies and high IOPS.
- High Availability and load balancing were also relevant.
- An important part of the project was the license costs reduction by switching the hypervisor to Microsoft Hyper-V.
The Chosen Solution
To achieve the goals, IT-Beratung HALBE asked for help from the Boston Deutschland – as they had done in the past on several occasions when similar projects were carried out. Boston Deutschland proposed a customized all-flash High Availability storage solution based on the Open-E JovianDSS system with two nodes and a shared JBOD as a storage extension, with a new virtualization host as a package.
Positive Side Effects
The space required in the server rack has been significantly reduced. In the past, an entire server rack was filled with hardware for this purpose, but now it’s been decreased to only one third. Comparable solutions could only be realized by significantly exceeding the budget of more than 20,000 EUR. The handling of failures of any kind has been improved considerably with Open-E JovianDSS compared to DSS V7.
The main reason we have chosen the redundant fail-proof storage solution based on Open-E JovianDSS presented by Boston Server & Storage Solutions was the perfect harmonization of stability, performance, scalability, and flexibility (which is more and more important these days). It allows us to react perfectly to the market needs. Another noteworthy aspect is the amount of electricity saved, about 4 kWh per day, which, thanks to the modern hardware, also benefits the environment and has a positive impact on our electricity costs. Especially the transparent advice in the area of hardware design possibilities on the part of Boston Germany, along with a fair price, convinced me - à la bonne heure. With this solution, we have created the optimal set of their applications in the home office for our customers.
Hardware details
| Configuration per cluster node | |
|---|---|
| Server: | 1HE Supermicro SYS-1029P-WTRT |
| Chassis: | SuperChassis 116AC2 R706WB2 |
| Motherboard: | X11DDW-NT |
| Power supply: | Redundant 740W - 80 PLUS Platinum Level Certified |
| Processor: | 2x Intel® Xeon® Silver 4112 2.60 – 3,00GHz |
| RAM: | 192 GB DDR4 2933 ECC (12x16GB) Samsung M393A2K40CB2-CVF |
| RAID: | 1 x Intel® RAID modules RMS25CB080 |
| HBA: | LSI SAS 9300-8E SGL, 8-Port ext. 12Gb/s SFF-8644, SATA + SAS |
| Network connection: | 2x Broadcom NetXtreme E-Series P210TP, Dual-Port 10GBase-T |
| Boot: | 2x 128GB SSD/MLC SATA DOM (Mirrored – RAID 1) |
| Write Log: | RAM/SSD tiered |
| Read Cache: | RAM/SSD tiered |
Storage configuration
| JBOD: | 2HE SuperChassis 216BE2C-R741JBOD |
| SSD: | 13x Toshiba PM5-R 1920GB SSD, TLC, SAS 12Gb/s, 2.5’, 15mm, 1 DWPD |
| Zpool: | 2x Raid Z2 (2 Gruppen a 6 Disks) + 1x WarmSpare |
| ISCSI Multipathing: | Active-Active configuration |
| Storage capacity: | 12,58TB usable in two Zpool’s |
