In practice, you can RMA any drive, even one that is perfectly fine.
Also certain numbers of reallocated sectod do signal a failing drive,
and a steady increase over a certain rate does to. Not with maximum
reliability, but SMART also is about early warning, so a bit of
paranoia is a good idea.
Not at all. Seem to me you have limited experience with SMART.
Arno
Google's paper, Failure Trends in a Large Disk Drive Population:
http://labs.google.com/papers/disk_failures.pdf states:
Our results conï¬rm the ï¬ndings of previous smaller population studies
that suggest that some of the SMART parameters are well-correlated
with higher failure probabilities. We ï¬nd, for example, that after
their ï¬rst scan error, drives are 39 times more likely to fail within
60 days than drives with no such errors. First errors in
reallocations, offline reallocations, and probational counts are also
strongly correlated to higher failure probabilities. Despite those
strong correlations, we ï¬nd that failure prediction models based on
SMART parameters alone are likely to be severely limited in their
prediction accuracy, given that a large fraction of our failed drives
have shown no SMART error signals whatsoever. This result suggests
that SMART models are more useful in predicting trends for large
aggregate populations than for individual components. It also suggests
that powerful predictive models need to make use of signals beyond
those provided by SMART.
So S.M.A.R.T. is really not proven technology, like IBM's glass
platter drives. [1] Anyone else concur?
[1]
http://www.datacent.com/datarecovery/hdd/ibm