[mythtv-users] Needs second opinion, is my HD failing?

[Mark Cooke]

On Mon, 2009-03-09 at 10:10 -0400, John Drescher wrote:
> >
> > 2009/1/29 Richard Shaw <hobbes1069 at gmail.com>
> >>
> >> I've been having some quirky issues lately and decided to take a look
> >> at the SMART data for the disk. There seems to be a large count of
> >> errors in some of the categories.
> >
> > [snip]
> >>
> >>  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail
> >> Always       -       955

[snip]

> Good point. I have not thought of that before even though I manage
> several hundred hard drives. I'd better start looking at that. At home
> all of my spinning drives appear to be good. Looks like having a
> (realloc sector) value other than 0 is not normal with seagate 7200.10
> and 7200.11 drives.

Anything more than 0 is a concern - what you can do to help easy your
mind about a (small) non-zero count is get the drive to do an internal
SMART test.  This leaves the drive accessible for recordings while the
test is done (but high io load means the test takes longer to complete).

For IDE disks:

  smartctl -t long /dev/wonky_hd

For SATA:

  smartctl -t long -d ata /dev/wonky_hd


You'll hopefully (eventually) get a section like the following in your
smartctl -a output:

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining
LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed without error       00%     30768
-


If you get:

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining
LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed: read failure       90%     16265
114569

it is time for you to fill out the warranty returns form and make sure
your backups are current and working...


Nearly final note - configure smartd to email you when critical SMART
parameters change.  That might be a very useful early warning.


Final note - not all disk controllers support raid, so you might have to
play switcharoo with your cables.  (Eg, the jmicron's on the ex58-ud5
don't with the centos kernel).


Cheers,

Mark