Re: [smartmontools-support] Which attributes can estimate number of bad blocks?

[Bruce A. <ba...@uw...>]

Hi Gandalf,

> On 02 Aug 2016, at 08:59, Gandalf Corvotempesta <gan...@gm...> wrote:
> 
> Il 01 ago 2016 22:33, "Tim Small" <ti...@se...> ha scritto:
> > Although it's relatively rare, I've had drives which have been happily
> > operating for years with a reallocated sector count of a few tens,
> > without any further problems.  I've put these down to a one-off event or
> > very localised manufacturing defect within the drive.
> 
> Currently i have a sas drive with grown list set to 16 from some months ago.
> everything else seems to be stable. Would you replace this drive with 16, stable,  bad sectors?

I would not replace this, if the number of bad sectors is not growing.  16 is not a large number.

> with sata disk i have some drives with 3 or 4 logged errors 4 years ago and nothing more after that. Would you replace?

Probably not.  What types of errors.

> I also have a sata disk with raw read error rate set to 84, smart long test running weekly doesn't catch any issue and is able to run properly every time. No reallocation or bad sectors, only this 84 raw read error rate.

I would not worry about this.  It could have to do with the normalization of this raw value, which might render the actual value meaningless.  Do you have other drives which are identical hardware and firmware to compare with

> Raid controller osd able to run weekly consistency check and patrol reads (it will read the whole raid looking for bad sectors) with no issue at all.
> Would you replace this disk?

No, I would not.

Cheers,
	Bruce