S.M.A.R.T. Monitoring Tools

Kapil,

Sorry for the slow reply.  I don't know of a straightfoward way to modify 
smartd to do what you want.  I would suggest that you write a simple 
script which uses smartctl and checks (say once per minute) to see if a 
self-test is still in progress.  Then log that data to a file with systime 
time/date timestamps and later do a comparison with the smartd logs.

Cheers,
 	Bruce

On Fri, 3 Aug 2007, Dukle, Kapil (GE Healthcare) wrote:

> Hi all,
>
> I m trying to use smartd to run periodic SMART tests on the SCSI drives
> for the Linux blades.
> I setup smartd so it logs information to /var/log/smartd.log
>
> Here is what /etc/smartd.conf looks like:
> # To schedule a long Self-Test everyday at 3 am, 9 am, 2 pm and 6 pm
> /dev/sda -a -s L/../.././(03|09|14|18)
> /dev/sdb -a -s L/../.././(03|09|14|18)
>
>
> I see messages in the log file that indicate when the test starts. My
> question is:
> Is it possible to have smartd report when the self test actually
> finished?
>
> I am trying to run the test multiple times during the day to test how
> long it takes
> during varying degrees of disk usage.
>
>
> Here is the /var/log/smartd.log output:
>
> Aug  3 02:55:03 vre smartd[28148]: smartd version 5.37
> [x86_64-redhat-linux-gnu] Copyright (C) 2002-6 Bruce Allen
> Aug  3 02:55:03 vre smartd[28148]: Home page is
> http://smartmontools.sourceforge.net/
> Aug  3 02:55:03 vre smartd[28148]: Opened configuration file
> /etc/smartd.conf
> Aug  3 02:55:03 vre smartd[28148]: Configuration file /etc/smartd.conf
> parsed.
> Aug  3 02:55:03 vre smartd[28148]: Device: /dev/sda, opened
> Aug  3 02:55:03 vre smartd[28148]: Device: /dev/sda, is SMART capable.
> Adding to "monitor" list.
> Aug  3 02:55:03 vre smartd[28148]: Device: /dev/sdb, opened
> Aug  3 02:55:03 vre smartd[28148]: Device: /dev/sdb, is SMART capable.
> Adding to "monitor" list.
> Aug  3 02:55:03 vre smartd[28148]: Monitoring 0 ATA and 2 SCSI devices
> Aug  3 02:55:03 vre smartd[28150]: smartd has fork()ed into background
> mode. New PID=28150.
> Aug  3 03:25:03 vre smartd[28150]: Device: /dev/sda, starting scheduled
> Long Self-Test.
> Aug  3 03:25:03 vre smartd[28150]: Device: /dev/sdb, starting scheduled
> Long Self-Test.
> Aug  3 09:25:03 vre smartd[28150]: Device: /dev/sda, starting scheduled
> Long Self-Test.
> Aug  3 09:25:03 vre smartd[28150]: Device: /dev/sdb, starting scheduled
> Long Self-Test.
>
> Running smartctl -l selftest doesn't seem to have this information
> either:
>
> [root@...]# /usr/sbin/smartctl -l selftest /dev/sda
> smartctl version 5.37 [x86_64-redhat-linux-gnu] Copyright (C) 2002-6
> Bruce Allen
> Home page is http://smartmontools.sourceforge.net/
>
>
> SMART Self-test log
> Num  Test              Status                 segment  LifeTime
> LBA_first_err [SK ASC ASQ]
>     Description                              number   (hours)
> # 1  Background long   Completed                   -    7970
> - [-   -    -]
> # 2  Background long   Completed                   -    7964
> - [-   -    -]
> # 3  Background short  Completed                   -    7481
> - [-   -    -]
> # 4  Background long   Completed                   -    5499
> - [-   -    -]
> # 5  Background long   Completed                   -    5493
> - [-   -    -]
> # 6  Background short  Completed                   -    5485
> - [-   -    -]
> # 7  Background short  Interrupted ('-X' switch)   -    4861
> - [-   -    -]
> # 8  Background short  Completed                   -    4861
> - [-   -    -]
> # 9  Background short  Interrupted ('-X' switch)   -    4861
> - [-   -    -]
>
> Long (extended) Self Test duration: 1560 seconds [26.0 minutes]
>
>
>
>
>
> Thanks,
> Kapil
>
>

Hi all,
=20
I am wondering how the SMART long selftest responds to an increase in
disk activity. For example, after kicking off a long selftest, I start
using the drive for continuous streaming
reads or streaming writes (using dd or sg_dd), how does the SMART
subsystem handle this?
=20
Does the SMART subsystem temporarily pause the selftest? Does the
selftest resume after the activity ends (or is reduced), or does one
need to manually restart the test?
=20
I couldn't find any detailed explanation on this. Please let me know if
this is documented someplace -  I couldnt find it in the man pages...
=20
Thanks,
Kapil
=20

On Wed, 10 Oct 2007, Dukle, Kapil (GE Healthcare) wrote:

> Hi all,
>
> I am wondering how the SMART long selftest responds to an increase in 
> disk activity. For example, after kicking off a long selftest, I start 
> using the drive for continuous streaming reads or streaming writes 
> (using dd or sg_dd), how does the SMART subsystem handle this?

I think that this depends upon the individual disk's firmware.  Since GE 
Healthcare probably buys a lot of disks, I suspect you can ask the 
engineers at the disk drive company and get a clear answer.

> Does the SMART subsystem temporarily pause the selftest? Does the 
> selftest resume after the activity ends (or is reduced), or does one 
> need to manually restart the test?

I think the self-test does not need to be manually restarted, but simply 
slows down.  This might even be part of the ATA specs (which you can find 
linked from the smartmontools home page).

> I couldn't find any detailed explanation on this. Please let me know if 
> this is documented someplace - I couldnt find it in the man pages...

It can not be documented.  SMART implementation is up to the disk drive 
manufacturer, and varies significantly from company to company.  There is 
even quite a bit of variation between different drive families from a 
given manufactuer.

Cheers,
 	Bruce