Re: Dying hard drive?
On 2013-10-18 11:43 -0400, Miles Fidelman wrote:
> Do a smartctl -A /dev/sd[abcd] - look for non-zero raw read errors
> and reallocated sector counts.
And Current_Pending_Sector !
> I've found, at least for the WD drives I use in my servers -
> anything other than a 0 raw-read-error count is a sign of
> near-term disk failure.
This is highly manufacturer dependent. In my experience, with
Seagate drives, it's normal to have Raw_Read_Error_Rate in the
tens or hundreds of millions.
> Note that this particularly applies if you're not using an
> enterprise-class drive. Standard drives try very hard to read from
> the medium, and take a long time before they give up. Enterprise
> drives assume they're part of a RAID array and just give up,
> throwing an error.
Really ? Interesting, didn't know that.
--
André Majorel <http://www.teaser.fr/~amajorel/>
lists.debian.org, a spammer's favourite.
Reply to: