On 1 May 2014 09:39, Carlos Marcano wrote: .... Acknowledges surprise on the result regarding > temperature as a failure motivator but fails to recognize it's still > important incidence. > > http://static.googleusercontent.com/media/research.google.com/en//archive/d= isk_failures.pdf The importance to note is that AS PRESENTED temperature is roughly inversely correlated with failure rate up to two years of age, anything up to 40C is OK at 3 years, with 40-45 being bad and over 45 very bad, and 15-30 best, then for 4 years old drives (that have survived) it swaps to about linear with temperature except that 30-35 is slightly better than 15-30. Pretty clearly [tm] the data needs more beating than these graphs allow BUT it seems that running drives in the 35-40 C range overall produces best results. Figure 4 confirms this - but what it means needs to be understood. The columns are proportion of drives at observed temperatures and the line graph with range markers show failure rates at these various temperatures. They do not say how or where the drive temperatures are measured. Presumably this is reported by the drive itself - and that still does not say exactly what that means - but that information will be available. It does SEEM that running drives "not too cold" may be in order. Of course, taking too much notice of specific impressions formed from such an article is dangerous, as not only is the manner in which it was acquired and processed not fully known, but the usage patterns of Google may be different enough from 'your' usage to make a difference. Russell --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .