It's not a matter of not understanding when MY OWN experience is that I have had more hard then soft errors. You seem to have a consistent problem of believing your experience equals facts. Just because in your experience you have had more soft errors then hard, doesn't make it true for everyone else. If I have a client who lost an entire days work because I chose to implement a solution that only solved soft errors I would be doing him a disservice. I'm not sure how I would tell him that at the cost of thousands of dollars of inactivity while down, was worth not spending a few hundred more when the server was being built. It was just a few weeks ago that this was exactly the case, a hard drive had failed hardware wise and the engineering company would have been completely down had it not been for a RAID. Instead he never noticed a problem at all, his company kept making money and with zero down time the HD was replaced. This solution from the 80's and 90's as you say, kept him running in 2008 when your so called 98% solution would have cost him a lot more in down time.
Mike On Sat, Sep 6, 2008 at 9:55 AM, Tom Piwowar <[EMAIL PROTECTED]> wrote: > >Are not we better off trying to do our best protecting all types of > errors? > > Not necessarily. This is a great example of the importance of good > judgement. There are 2 types of errors with "soft" occurring 98% of the > time, and "hard" 2%. You want to make a major effort to protect against > the 2% and do nothing about the 98%. To me that is just silly. RAID is a > silly waste of resources. > > >All your problems are soft related, I've seen more hardware related. > > I'm sorry the you don't understand that things change and the > significance of MTBFs. You keep implementing tech solutions from the '80s > and '90s. > > ************************************************************************* ** List info, subscription management, list rules, archives, privacy ** ** policy, calmness, a member map, and more at http://www.cguys.org/ ** *************************************************************************
