[CentOS] CentOS and Marvell SAS/SATA drivers

Mon Mar 7 20:19:43 UTC 2011
compdoc <compdoc at hotrodpc.com>

> My scepticism regarding SMART data continues ... the flaky drive
>showed no errors, and a full test and full zero-write using the WD
>diagnostics revealed no errors either.  If the drive is bad, there's
>no evidence that would cause WD to issue an RMA.


I've been having a rash of drive failures recently and I have come to trust
SMART.

One thing's for sure - SMART is not implemented the same on all drives or
controllers. Recently one older Seagate drive showed no SMART capability in
linux using the gnome-disk-utility, but I could read the SMART data from the
drive in Windows with HD Tune.

It isn't infallible, but SMART is certainly one tool you can use in the
diagnosis. I wouldn't ignore Reallocated Sector counts or Current Pending
Sector counts, for instance.

Working for a customer this weekend, I replaced an older 60G WD drive that I
knew for months to have bad sectors, but the Reallocated Sector Count was
still 0. After a scan for errors with HD Tune, the Current Pending sector
count showed 13, but the Reallocated Sector Count never grew.

There is still a lot for me to learn - like the relationship between SMART
within the drive and the controller's support of SMART. You would think they
are independent of each other, but I wonder...