[<<]Message[>>] [<<]Author[>>] [<<]Subject[>>] [<<]Thread[>>]
Number : 4944 Date : 2003-07-04 Author : Bob Weir Subject : Re: Backup and Disk replacement --- just a thought Size(KB) : 1
Kan said: - > Most of us in computer management business, try to set up > and carry out a periodic backup regimen to avoid catastrophic > disk failures that we all know happens with rather almost > predictable fashion. Yes, but the normal failure pattern is a "bathtub" shape with a modest number of early failures followed by a long stable period with few failures before the terminal failure. SMART should predict most failures and trigger replacement at a cost effective time. With hardware, "If it ain't bust don't fix it" is a good motto. There may be apparent merit in replacing one of a pair of disks used in an "XX cloned array" half way through it's expected life but replacing the pair when the first fails achieves the same result with less intervention. > 1. If we accept the notion that wear-and-tear is somewhat > proportional to the amount of usage, then, it leads to > a conclusion that a daily-backup operation actually > increases the likelihood of failure. I think this will be a marginal effect. Unlike other schemes the great (unsung) advantage for xxcopy is the incremental nature of /CLONE (or similar xxcopy switch groupings) that minimise disk activity during routine holistic backup. The decision to backup after certain periods or tasks is based on the value of the disk contents and the probability of that content becoming corrupted (more often by software than from breakdowns or cosmic rays?). I'm not in computer management but, for me, protecting the disk contents from corruption by software (e.g. by the use of routine XX cloning to a second disk) far outweighs any modest increase in the overall cost of ownership caused by increased wear and tear on the disks. Regards, Bob Weir
This message if part of XXCOPY's message Archive. The archive contains all the messages posted at Yahoo!Groups: XXCOPY.