Rich Freeman on 14 Jun 2019 07:03:02 -0700

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

Re: [PLUG] Is My Disk Toast?

On Fri, Jun 14, 2019 at 9:10 AM Louis K <> wrote:
> I've got a HDD that starting throwing errors to dmesg (below). It happened on about 8 different sectors recently, and happens repeatedly. I've been googling around and learning about smart, but the diagnostic seems (also below) to indicate the disk is good (Reallocated_Sector_Ct and Reallocated_Event_Count have raw values 0).
> Is this disk dying, or can I run a tool to reallocate the bad sectors?

I have had many hard drives with persistent unrecoverable read errors
where SMART never gave any hint that any reallocation is taking place.
I suspect some vendors just don't report the truth in this data.

If the OS sees errors, some piece of hardware isn't doing its job
correctly.  That is probably the drive, though I've seen issues with
controllers and cables too.  98% of the time it will be the drive.

You're probably already losing data unless you have some kind of
redundancy/etc.  Since the error is being reported if you do have
redundancy then the data is probably being corrected and the sector is
probably getting rewritten.  However, if the drive is out of
reallocation sectors or for whatever reason isn't doing that, then
you'll keep getting new errors.

I generally consider any kind of persistent error like this grounds
for replacement.  If you can capture errors in the SMART testing log
(do an offline test) then chances are the manufacturer will RMA it
under warranty if it still qualifies.

Stepping back, ALL storage devices WILL fail, sometime.  So today is
the best time to be ready for your next failure.  It is nice when you
get some errors and can potentially replace the drive while it still
can serve redundant data for most of its content, but drive failures
can also come without any warning at all and be total in nature.  For
casual use you probably don't need spare drives lying around, but you
should have enough redundancy that it is unlikely that you will have
data loss, and that should include offline backup for anything that
you can't re-create.

Philadelphia Linux Users Group         --
Announcements -
General Discussion  --