The article focuses on common components, such as SATA disk drives in a 7-drive RAID-5 (or RAID-6) array, each at 2 TB gross capacity. People who have administered RAID over time understand that the scariest time, where risk is greatest, is right after a failure. The RAID controller is trying to restore the drive it lost. Subsequent failure(s) can mean the entire array is lost. The author focuses on a common problem - read failures. A humble little issue that could cause, in this example, a loss of 12 TBs of data.
The more spinning disk drives you have, the greater your chance of one drive failing. In the example, seven drives, the risk of a drive failure increased and left it vulnerable while it tried to restore some 2 TBs of data. Enterprise configurations have additional protections, but are just less vulnerable. Rising disk capacities without a corresponding increase in throughput adds to this risk of total loss of all data on the array. Thank you, Mark Snyder -----Original Message----- Not too long ago there was a thread on this list about RAID where some argued that the MTBF was so high on drives that RAID was obsolete. Now this article says basically RAID is obsolete because of such high MTBF. Neither camps who seem to not like RAID much offer anything whatsoever to replace it. I'm not sure why in the case of RAID, the idea is that if it fails at all, the entire idea of RAID is a failure. If this logic were extended, none of us would be driving cars. While this article might be interesting in some shallow way, in the end the real point is not even really stated, just alluded. RAID was never meant as a complete backup solution. Those who backup will be rewarded, those who don't..enjoy your suffering. Mike ************************************************************************* ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *************************************************************************
