freebsd stable - Jul 2007 - Help! My laptop drive may be dying.

Hi,

My laptop drive might be dying. It is a Samsung MP0804H which I have 
used for around 28 months without issue.

Every now and then it will click and sound as though it is thermally 
recalibrating itself.

I ran SMART diagnostics from smartmontools, and Samsung's own diag tools 
which all report the drive is OK (a full captive surface test). I see 
nothing untoward in the SMART info pages.

Whilst Windows is able to tolerate the retrying of ATA commands which 
this click appears to be inducing, FreeBSD can easily get sick and just 
hang, which majorly gets in the way of real work.

I am always running X without exception when this happens, so I can't 
get meaningful ATA error messages. Of course these happen before the 
buffer cache is flushed, so they don't show up in /var/log/messages, if 
anything is showing up at all.

Many thanks for any help you can provide...

regards,
BMS

At 09:43 AM 7/22/2007, Bruce M Simpson wrote:
>I ran SMART diagnostics from smartmontools, and Samsung's own diag 
>tools which all report the drive is OK (a full captive surface 
>test). I see nothing untoward in the SMART info pages.
Hi Bruce,
         These symptoms (the OS reporting errors, the drive saying 
all A-OK) remind me of a bad cable/connector.  Have you tried a new cable ?

         ---Mike

Bruce M Simpson wrote:
> Hi,
>
> My laptop drive might be dying. It is a Samsung MP0804H which I have 
> used for around 28 months without issue.
>
> Every now and then it will click and sound as though it is thermally 
> recalibrating itself.
>
> I ran SMART diagnostics from smartmontools, and Samsung's own diag 
> tools which all report the drive is OK (a full captive surface test). 
> I see nothing untoward in the SMART info pages.
Well, sounds like the drive is indeed dying, recalibrating noises like 
that is a bad sign, its most likely having real trouble reading certain 
areas of the medium.

If your SMART output tells anything about read retries or number of 
remaps that could be an indicator of upcoming problems, however not all 
drives has that info in the SMART pages.

In the real world scenario SMART can only tell you about the problem 
when the drive has given up on the data, there is almost newer any real 
prewarns to failure.
>
> Whilst Windows is able to tolerate the retrying of ATA commands which 
> this click appears to be inducing, FreeBSD can easily get sick and 
> just hang, which majorly gets in the way of real work.
Depends on whats happening, you could try to up the timeout in 
ata-disk.c and see if it survives the errors that way, to at least try 
to save the data before its too late.

-S?ren