The Genstor HA cluster solution is a high availability, business continuous, load balancing solution that allows two servers to be connected forming a cluster. The function of this cluster is to keep data flowing even if one of the servers fails. This failover mechanism is built on redundancy and employ functions such as:

• Redundant resource management
• Failure detection
• Isolation and recovery
• Logging and notification tasks

Genstor HA based clusters can be configured for one of two operational modes: Dual Active and Hot Standby.

Dual Active: Both servers share the workload. If one of the servers fails, the remaining server automatically starts carrying the full workload.

Hot Standby: One server is online carrying the full workload and the other server is in a hot standby mode. If the online server should fail, the standby server will automatically come online to take over the workload.

Genstor now offers a “Cluster-In-A-Box” solution. Included in the cluster is 2 server boards connected to a common pool of 16 drives. The system employs a dual 10GbE connection between the server boards via the midplane. If one server board fails, the other server board can take over control and access the drives, keeping the system up and running. The two controllers can also work in Dual Active mode. This feature is available with optional storage software.

Features

• Automatic Detection: Allows two servers to continually monitor each other for breakdowns and/or irregularities. If one of the servers should fail, the other one will automatically take over.
• Automatic Switch-over: When HA detects an irregularity, recovery routines are immediately implemented. If the irregularity cannot be resolved locally, operations are switched to the other server.
• Automatic Recovery: Once the failed server is repaired, the administrative person can choose to initiate either an automatic or manual recovery process.
• Load Balancing: With HA configured for "Dual Active" configuration, both servers will share in the network workload resulting an increase in network efficiency.
• Automatic or manual failover: In addition to Automatic failover, Genstor HA Solutions can be configured for manual failover. This feature frees up one of the servers to perform functions such as trouble shooting, upgrading and/or scheduled maintenance without causing long interruptions.
• Automatic Notification: When irregularities or a breakdown occur, Genstor HA Solution will automatically notify pre-selected personal via page or fax messages.

HPC Cluster Solutions

Genstor HPC cluster solutions are customized to customer requirements and budget. Our HPC clusters are based on a 1u/4 node arrangement, using Intel 3000 series Xeon CPUs or AMD 6100 series CPUs for up to 48 cores and support for up to 512GB of memory per node. Our solutions offer a variety of interconnects allowing low latency for cost-effective performance, ensuring we can provide the best solution while remaining within your budget.

Memory architecture is versatile. Large DIMM sizes are more cost-effective now making caching and diskmem using memory extremely appealing. We offer a variety of GPU options to broaden our cluster capabilities and offload many processes from the main CPUs.

Hardware

• Host: Intel® or AMD® based, with optimized settings.
• Interconnect: Gigabit Ethernet, and Infiniband.
• Memory: Up to 512GB per node.

Operating Systems supported

• BSD (most versions)
• Linux(most versions)
• Microsoft Windows

Database

• Customer spec.