[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Disk performance deteriated to unbearable levels

Stan Hoeppner wrote:
On 11/9/2011 1:34 AM, Bob Proulx wrote:
Miles Fidelman wrote:
try smartctl -A /dev/sda

that will give you a much longer list of statistics collected by the drive

the one I always look at first is the absolute value of "raw read
errors" - if that's higher than 0, the drive is starting to fail,
and its internal code is spending more and more time trying to read
and re-read data off the media
Excellent information!
Not really.  The numbers are not absolutes.  And they differ among
manufacturers.  S.M.A.R.T. is a data format standard for "drive health",
but it does not dictate "value" standards for the field contents.

For the end user, attempting to interpret some S.M.A.R.T. data for some
makes of drives is hit or miss.  I would say as a general rule that one
should contact the drive manufacturer if s/he sees S.M.A.R.T. error
counts rise.  The drive may or may not be failing.

Specifically for "raw read error" - in the format

  1 Raw_Read_Error_Rate     0x000b   088   088   062    Pre-fail  Always       -       3342368

I've found, across many different dries, that the "VALUE" field (which is a relative field calculated over time) is generally useless, while the RAW_VALUE field is almost always indicative of pending failure if it's value is anything other than 0.

In theory, there is no difference between theory and practice.
In<fnord>  practice, there is.   .... Yogi Berra

Reply to: