On 2014-08-27, Geoff Steckel <g...@oat.com> wrote: > This paper: > http://download.intel.com/support/motherboards/server/sb/enterprise_class_versus_desktop_class_hard_drives_.pdf > describes the different features and intended uses of > enterprise drives vs. desktop drives. The hardware requirements > for a (good) enterprise drive far exceed those of a desktop drive.
Well, it's not clear how much of that is actual technology and how much is marketing. > Most important to users: different error recovery philosophy! Most important: difference in data integrity (factor 10 if you believe the data sheets). > Desktop: do whatever is necessary to correct a read error, > no matter how long it takes. The software is not time sensitive > and may not be able to recover from a single sector error. > > Enterprise: disk must stay on line. Perform simple error > recovery and depend on higher level software to repair > or replace the bad sector. The irony is that my most recent drive failure, which I mentioned here on this list, involved a Hitachi Ultrastar 7K3000--not a desktop drive--and the most remarkable thing was how tenacious the little beast proved in recoverying the data, although it became VERY slow. Of course, if you read that Intel document, they are saying: "A typical desktop drive command timeout can take many minutes and no disk access is allowed while the system attempts to retry the command." "Typical timeout for an enterprise class drive is 7 to 15 seconds and retries are limited to a few attempts." "Desktop-class drives with timeout values exceeding 30 seconds should not be used in an enterprise-class system." These are ridiculously large numbers, but if you take them at face value, then this is another argument against desktop drives, as running into a bad sector will badly lock up your system and having a multitude of bad sectors will render the rest of the drive effectively unrecoverable. -- Christian "naddy" Weisgerber na...@mips.inka.de
