Hi Bruce,

 thank you for having replied in a such short time.

Well I thought smartd run some tests infact I coulnd't explain the existence of smartctl except than reading attributes.

As you told I've to run  "smartctl -t long /dev/hda" by crontab...but what should be the most right scheduling...I mean 4 hours or more? Doesn't this generate some problems (conflicts)  with the self test which is made by the hard disk itself? (or the hard disk doesn't perform any kind of self test?).

Let me see if I understood the flow of operation:

1. I've to run smartctl in order to query the hard disk and smartctl updates the attributes

2. smartd will read the updated attributes and will notify what changed is.

Another question....(sorry :o) ) I can't understand when I should use the directive -S:


       -S VALUE

              Enables  or disables Attribute Autosave when smartd

              starts up and has no  further  effect.   The  valid

              arguments   to  this  Directive  are  on  and  off.


This is also a command line parameter we can find on using smartctl so this seems  I can use smartd instead of smartctl, but we have told that these programs make different things...so I feel a little confused on finding duplicated command line parameter/directive on both smartctl and smartd.... :o( when using the one and the other?


Thank you,


 Bruce Allen <ballen@gravity.phys.uwm.edu> wrote:

Hi Fabrizio,

> I've just installed smartd and configured smartd.conf as reported below:
> /dev/hda -H -o on -f -l error -l selftest -m fabriziodimeo@yahoo.it -t \
> -R 1 -R 3 -R 5 -R 7 -R 11 -R 13 -M daily

This looks OK.

> executing the command smartctl -c /dev/hda I get this message:

You should probably use
smartctl -a /dev/hda
to get a complete picture.

> I run the self test manually and next I abort it, but seems no more
> tests have been executed by smartd.....because the message reported
> just a few rows above...is this right?

smartd does not run self-tests. It only monitors the self-test log for
signs of self-tests that showed errors. If you want to run self-tests on
a regular basis I suggest that you schedule a cron job to do this,
smartctl -t long /dev/hda
for example.

> My questions are:
> 1. how can I check out what smartd is doing (because no message is
> written in /var/log messages every 30 minutes - this is the default
> run-time right?)

smartd only writes messages into /var/log/messages when a disk Attribute
has changed. Until that happens, there are no messages.

> 2. is the configuration directives right as I wrote it in smartd.conf?

It looks fine. You might want to add -M test to get a test mail on
startup to be sure that is working OK.


Yahoo! Cellulari: loghi, suonerie, picture message per il tuo telefonino