2

After going through the man pages of smartctl, smartd and smartd.conf, I am no clearer on how S.M.A.R.T. is supposed to work on a scsi device. I have a default configuration file that says:

DEVICESCAN -H -m root -M exec /usr/libexec/smartmontools/smartdnotify -n standby,10,q

The above does not work for my case as DEVICESCAN is not able to detect my scsi disk on RAID 1 configuration. I ran the following command and it says that the next internal SMART test will run in 47 minutes time:

# smartctl -d scsi -A /dev/sg1 === START OF READ SMART DATA SECTION === Current Drive Temperature: 34 C Drive Trip Temperature: 68 C .... Vendor (Seagate/Hitachi) factory information number of hours powered up = 11183.48 number of minutes until next internal SMART test = 47

From the man page of smartd, I understand that the default interval between disk checks is 30 minutes. Now, my question is, how does the 47 minutes come about?

Also, from the man page of smartd.conf, there are actually very few options that are valid, namely the -l self-test (even -H is for ATA only). Do I have to explicitly schedule a self-test (short or long) prior to the disk check?

For the above question, my point is, if I just have the following in my config:

DEFAULT -d scsi -l selftest -m root -M exec /usr/libexec/smartmontools/smartdnotify /dev/sg1 /dev/sg2

would it be of any use at all?

Extra Info: I am using Dell PERC H200 RAID controller on two Seagate ST3300657SS disks.

Improve this question
4
  • "number of minutes until next internal SMART test" is based on factory settings in the device. smartd can initiate a self-test as well, either periodically (with the -i option, which is where you saw the 30 minute default number) or on demand (via smartctl). See smartmontools.sourceforge.net/smartmontools_scsi.html#selftest Commented Jul 22, 2014 at 16:23
  • @MarkPlotnick, is smartd capable of monitoring failures for scsi device? From the man page of smartd.conf, it seems that the important -f and -H options are for ATA only. The background scan perform by disk does not seem to rely on smartd and smartd (smartctl can) does not seem to be able to read results from it. Furthermore, default self-tests are not logged. It appears to me that if self-tests are not performed, smartd would not be able to detect any failures, since the log which it relies on will not be updated. Commented Jul 23, 2014 at 4:30
  • Can you add your disk and RAID controller (and HBA, if you're using an external RAID) info to the question? Someone with experience with them can comment on their quirks. Some RAID controllers will restrict the low-level info you can get from drives. Some that I used would cause smartd to report "soft errors" on a RAID-5 only when a disk died. I realize you're using sg devices to get at the raw drives, but the controller may still affect what you can do with the individual drives. Commented Jul 23, 2014 at 12:52
  • @MarkPlotnick, its done. That is precisely why I need to know how it works :) Commented Jul 23, 2014 at 13:19

1 Answer 1

Reset to default
4

A SCSI device doesn't have anything exactly like SMART.

A SCSI device is mostly monitored with LOG SENSE commands for its supported log pages and some of them contain information similar to what SMART provides, or at least a similar concept.

The main equivalent is the "Informational Exceptions" log page 0x2f which provides an ASC/ASCQ pair that is zero if all is good and is non-zero if there is a warning or downright error.

The temperature and uptime are provided in different pages, the uptime is provided in the Background Media Scan log page. You can find the full log page definitions in the SCSI documents SPC and SBC.

Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.