02-04-13 | Blog Post
How to Configure a High Availability Rack
For most small to medium-sized Michigan businesses, a high availability IT infrastructure is a necessity to conduct business, generate revenue, and ensure customer satisfaction. How do you reduce the risk of downtime? The answer lies in the anatomy of a high availability rack, designed and managed in a Michigan data center. High availability racks are uniquely designed to withstand single points of failure with redundant power and Internet connectivity.
We have several resources for those of you interested in understanding the anatomy of a high availability rack and how to properly build one. Online Tech’s High Availability White Paper illustrates the redundant power supply and redundant Internet connectivity that ensure a greater level of protection against loss of power or network connectivity.
Online Tech’s Ask an Expert: How to Configure a High Availability Rack video takes a step-by-step approach on how to build a rack for high availability. Detailed instructions are also noted below.
How to Configure High Availability Power
To configure high availability power, you need two rack PDUs – the primary and the secondary. The primary PDU is fed by the primary UPS and the secondary PDU is fed by the secondary UPS. Each UPS should be backed up by their own generators.
How to Configure High Availability Cabling
The primary power supply should be cabled back to the primary rack PDU. The secondary power supply should be cabled back to the secondary rack PDU. This prevents a failure from one of the PDUs affecting the server.
Please note: When a server has two power supplies, running both at the same time, it is drawing its current from each power supply simultaneously – half from No. 1 and half from No. 2. This means half the power is fed from the rack PDU No. 1 and half the power is fed from rack PDU No. 2. So you want to make sure that, for example, if you have a 20 amp circuit, 16 amps is the usable limit, which means you can only draw 8 amps off of each PDU. If the primary PDU fails, the secondary PDU has to now support 16 amps, which is the usable limit.
The primary NIC should be cabled to the primary switch and the secondary NIC should be cabled to the secondary switch.
High Availability Networking
The primary switch should be powered by the primary rack PDU, has connections to the primary NICs and all the servers, and with the uplink to the primary core router. The secondary switch should be fed by power from the secondary rack PDU and have connections to the secondary NICs and all the servers and uplinked to the secondary core router.
By configuring your rack with a redundant design, if the primary switch fails, the secondary switch takes over. If the primary NIC fails, the secondary NIC takes over. If the primary power supply fails, the secondary power supply takes over. If the primary rack PDU fails, the secondary rack PDU takes over.
For more information on high availability, view/read our interview, Online Tech’s video Ask an Expert: What is High Availability?