Smartd error

[Marko Vojinovic]

On Mon, Sep 13, 2010 at 2:59 PM, Suvayu Ali <fatkasuvayu+linux at gmail.com> wrote:
> On Monday 13 September 2010 05:38 AM, Timothy Murphy wrote:
>> Yesterday I got an internal email from smartd telling me
>> of an error on one disk (on my server):
>> ------------------------------
>> Device: /dev/sdb, 1 Currently unreadable (pending) sectors
>> ------------------------------
>> I'm getting a copy of this email every hour.
>>
>> I'm not sure how serious a warning like this is.
>
> A few weeks back I saw a similar warning and my disk started showing bad
> signs within a few days. RMA it as soon as possible! You can take a look
> at the archives for my posts under "understanding smart logs" around the
> 15th of August.

FWIW, I had a drive which gradually (over a period of 2 years, iirc)
raised the number of pending unreadable sectors all the way up to 8558
(last time I checked), but the drive still seems to work, several
years now.

I don't keep anything serious on it (some multimedia stuff which can
be recovered from backup if the drive fails), but nevertheless,
recently I transfered the drive to another box, reformatted it without
issues, dropped a large load of data on it, and it still Just Works.
Of course, smartd regularly informs me about those 8 thousand bad
sectors, but I don't seem to be feeling any visible consequences yet.

I mean, if you have critical data on the drive then yes, do replace it
immediately. But if not, you can consider just keeping an eye on the
smartd log and making sure this number doesn't grow more than 1000 per
year. ;-)

On some drives 1 bad sector means that within couple of days the whole
drive will fail completely, but some can be used for several years on
without much trouble. YMMV.

HTH, :-)
Marko