I use smartd to monitor the disks in several large RAID arrays, trying
to find out in time if anything suspicious is going on. Particularly,
I am interested in unreadable sectors (I don't care, if a disk says
it's "healthy", if it is not able to retrieve the stored data).
Currently, I am running smartd with a configuration file, that
has a line like this for every single disk:
/dev/twa0 -a -I 190 -I 194 -R 5 -d 3ware,0
It seems like current disks never work too long without running into
media errors, so I tolerate some unreadable sectors, as long as the
number is stable. Unfortunately, with the above configuration, if
a disk has some unreadable sectors, smartd will regularly repeat
the syslog message about the problem.
Is there any way to get smartd to monitor everything of interest,
but report any detected problem only _once_?