When you install servers into a data centre, there are some important things to consider;
- How important is the application my server is running ? Can the server be offline and not impact my business greatly or must it be online 24/7
- Does my server have redundancy built in such as RAID and secondary power supplies
- Does my server support accessing ‘out-of-bands’
- Is my server covered by vendor warranty and support if parts fail
Many people collocate equipment for different reasons, so the importance of the above questions really does depend on the individual and their requirements. In some cases it may be best to discuss this with your data centre provider.
As a general ‘rule of thumb’ recommendations;
- Install equipment that has built in redundancy such as RAID and dual power supplies. The most common failure we see are failed hard drives, followed by failed power supplies ( although much more rare ). Failure of motherboards / CPU / RAM is extremely rare.
- Make sure the equipment has ( and is configured ) out-of-bands. This could be known as iLO ( HP ), iDrac ( Dell ) or IMM ( IBM ). This allows you to remotely control the server just as if you were at the equipment physically ( power on/off, install operating system etc )
- Ensure you have alerts configured – if a component fails, it is best to know about it so you can have a replacement ready
Article by Paul Arch, based on experiences operating data centre and cloud infrastructure in Perth Western Australia