Hallo all,

I currently working on instruments for early recognition of hardware
errors. One of the biggest problems are hard drives (i know, you know
it)... i can't use smart because it completely  fails in some cases.
Good indicator is kernel log, but it make no sense to parse it all the
time.

My question is: are there any sysfs interface to check for io errors on
some drive? I see there is  /sys/block/<dev>/stat interface, is it
possible to add it there or add a new file?


-- 
Regards,
        Alexey