You may or may not have heard of the term RAID when applied to hard-drives. RAID is an acronym for Redundant Array of Independent Disks, and is a system where a number of identical hard-drives can be utilised in a number of pre-defined configurations for the purposes of reinforcing data storage and/or extending the capacity of a single hard-drive across a number of drives while the operating system sees them all as one single drive. For more information on the types of RAID configuration available, see this article.

In my opinion; running a RAID array on your home system configured as a workstation isn’t really a viable idea. It can add a considerable amount of cost to your hardware budget; even with the falling cost of larger hard-drives, it is not much less prone to failure than a standard setup in such circumstances, and it can use considerably more wattage, putting extra strain on the PSU and spitting in the eye of environmental campaigners.

I won’t go into all the details of the various types of RAID array and their individual advantages and disadvantages at this point: Time and space constraints don’t allow me to write the eBook that I’d need to write in order to do the subject justice.

Server systems, however; particularly public servers, are a completely different kettle of fish with regard to using RAID; particularly in terms of backup:

As you may be aware, certain RAID configurations, such as RAID 1, automatically copy the data content of one disk to the other/another, in a technique known as “mirroring”. – Therefore, should one of the discs develop a data-error of some kind; the RAID-controller can instantly pick up the fact that the two discs have a data discrepancy, after which it can check the parity of the relevant areas of both discs individually, and discern which of the two has failed; then replace the data on the erroneous disc with the uncorrupted data on the non-erroneous disc. This saves the server system fro  having to call on a probably already busy technician to attend to the issue and run chkdsk or some similar tool to resolve the data error on a single drive.

In addition to this; should there be an error caused by irretrievable bad sectors on one of the drives, the original mirrored data is stored in its original form on another disc, and the server system signals a repair technician to replace the malfunctioning drive, after which the system restores the non-corrupted data to the new drive from the still-working mirrored drive.

The upshot of this is that the system can keep running with minimal if any interruption, and the workload on the technician is minimised. The last thing a server centre would wish for is avoidable downtime where a data error has to be manually repaired by the duty tech.

The above applies, not only to public internet servers, but to a slightly lesser extent to business and home servers; whether connected to serve a domain or as part of a Local Area Network or LAN.

In conclusion, then, a RAID array is a most useful thing to run on any type of server; particularly RAID 1 or similar. – However for a home or office computer used as a workstation I wouldn’t really consider setting up a RAID array as it just gives rise to unnecessary complications.

- Having said that; there may be circumstances where a RAID array on a terminal or workstation might be a good thing. Do you, the reader, have any ideas where this may be applicable? if so then please feel free to comment further to the above.

As a last minute addition, I’ll just add, as I’ve also said elsewhere, Linux is always a better operating system for servers in my opinion. Although I normally use Windows only on my desktop, with my Linux usage being so infrequent as to be negligible in that respect; I would nevertheless recommend Linux (Red Hat and others.) way more than I would recommend Windows for use with a server system.

In short; and you have every right to disagree if you wish, my preference is Windows for workstations, Linux for servers. – And yes it is possible to run RAID in Linux.

Advertisment: -