smart(8) Call for Testing

[Mark Saad]

On Mon, Mar 26, 2018 at 4:39 PM, Chuck Tuffli <chuck at tuffli.net> wrote:
> On Mon, Mar 26, 2018 at 1:25 PM, Mark Saad <nonesuch at longcount.org> wrote:
>> Mike / Chuck
>>   Silly question , how does it handle a failed disk ?
>
> It depends on how failed the device is. The application uses CAM to
> send pass-through commands. So, if the device doesn't appear in
> camcontrol devlist, you'd get "no such device" or a message from CAM
> indicating you might need to load a module. If the device is still
> registered with the system but not processing commands, I'd expect
> (hope?) the commands to timeout. In either case, there may not be any
> out put, but the exit status should be non-zero.
>
> On the other hand, if the disk is in some sort of Monty Python-eske
> "I'm not dead" state and responds to SMART / admin commands, you could
> see the output.
>
> --chuck

I was thinking more along the lines of when smartctl reports the overall-health
which is always a good time to read; and very Monty Python-eske

SMART overall-health self-assessment test result: FAILED!
Drive failure expected in less than 24 hours. SAVE ALL DATA.

https://serverfault.com/questions/114196/smart-warns-me-but-i-dont-trust-it

-- 
mark saad | nonesuch at longcount.org