Re: SMART Status Query

Re: SMART Status Query

[Stefan Richter]

(Cc'd LSML in case I write nonsense)

Sean Bruno wrote:
> Out of curiosity, I tried to query the smart status from disks in a
> FW400 enclosure.  I was prompted that the drives do not support SMART.
> 
> The Seagates in the enclosure do indeed support SMART and I suppose that
> there is a missing bit in the firewire stack or in my enclosure that
> would allow these commands to be passed down to the disks.
> 
> Does this situation have a resolution?  Is there a way around it?
> 
> I do find it interesting, that the the actual disk information is
> returned and not the "spoofed" enclosure disk names.
> 
> I tried "smartctl -a /dev/sdd" and was presented with:
> [sean@home-desk ~]$ sudo smartctl -a /dev/sdc
> smartctl version 5.38 [x86_64-redhat-linux-gnu] Copyright (C) 2002-8
> Bruce Allen
> Home page is http://smartmontools.sourceforge.net/
> 
> Device: ST380811 AS               Version: 3.AA
> Serial number:             5PS13NWB
> Device type: disk
> Local Time is: Mon Feb  2 22:36:07 2009 PST
> Device does not support SMART
> 
> Error Counter logging not supported
> Device does not support Self Test logging

AFAIK SMART requires "ATA pass-through", i.e. the tunneling of ATA
commands through SCSI.

The FireWire drivers happily send any command to the target which the
Linux kernel or an application emits.  The missing piece here is ATA
pass-through support in the firmware of the FireWire-to-SATA bridge chip
of the drive enclosure.

I have not yet heard of FireWire-to-SATA bridges which do support ATA
pass-through.
-- 
Stefan Richter
-=====-==--= --=- ---==
http://arcgraph.de/sr/

Re: SMART Status Query

[Sean Bruno]

On Tue, 2009-02-03 at 09:06 +0100, Stefan Richter wrote:
> (Cc'd LSML in case I write nonsense)
> 
> Sean Bruno wrote:
> > Out of curiosity, I tried to query the smart status from disks in a
> > FW400 enclosure.  I was prompted that the drives do not support SMART.
> > 
> > The Seagates in the enclosure do indeed support SMART and I suppose that
> > there is a missing bit in the firewire stack or in my enclosure that
> > would allow these commands to be passed down to the disks.
> > 
> > Does this situation have a resolution?  Is there a way around it?
> > 
> > I do find it interesting, that the the actual disk information is
> > returned and not the "spoofed" enclosure disk names.
> > 
> > I tried "smartctl -a /dev/sdd" and was presented with:
> > [sean@home-desk ~]$ sudo smartctl -a /dev/sdc
> > smartctl version 5.38 [x86_64-redhat-linux-gnu] Copyright (C) 2002-8
> > Bruce Allen
> > Home page is http://smartmontools.sourceforge.net/
> > 
> > Device: ST380811 AS               Version: 3.AA
> > Serial number:             5PS13NWB
> > Device type: disk
> > Local Time is: Mon Feb  2 22:36:07 2009 PST
> > Device does not support SMART
> > 
> > Error Counter logging not supported
> > Device does not support Self Test logging
> 
> AFAIK SMART requires "ATA pass-through", i.e. the tunneling of ATA
> commands through SCSI.
> 
> The FireWire drivers happily send any command to the target which the
> Linux kernel or an application emits.  The missing piece here is ATA
> pass-through support in the firmware of the FireWire-to-SATA bridge chip
> of the drive enclosure.
> 
> I have not yet heard of FireWire-to-SATA bridges which do support ATA
> pass-through.

That makes sense, unfortunate, but not unexpected.

Can you tell me how my system is getting the Seagate drive information
then?  I've looked over the data available to me on the Firewire bus,
and don't see where it's being pulled from.

Sean

Re: SMART Status Query

[Stefan Richter]

Sean Bruno wrote:
> On Tue, 2009-02-03 at 09:06 +0100, Stefan Richter wrote:
>> Sean Bruno wrote:
>> > I tried "smartctl -a /dev/sdd" and was presented with:
>> > [sean@home-desk ~]$ sudo smartctl -a /dev/sdc
>> > smartctl version 5.38 [x86_64-redhat-linux-gnu] Copyright (C) 2002-8
>> > Bruce Allen
>> > Home page is http://smartmontools.sourceforge.net/
>> > 
>> > Device: ST380811 AS               Version: 3.AA
>> > Serial number:             5PS13NWB
>> > Device type: disk
>> > Local Time is: Mon Feb  2 22:36:07 2009 PST
>> > Device does not support SMART
>> > 
>> > Error Counter logging not supported
>> > Device does not support Self Test logging
>> 
>> AFAIK SMART requires "ATA pass-through", i.e. the tunneling of ATA
>> commands through SCSI.
>> 
>> The FireWire drivers happily send any command to the target which the
>> Linux kernel or an application emits.  The missing piece here is ATA
>> pass-through support in the firmware of the FireWire-to-SATA bridge chip
>> of the drive enclosure.
>> 
>> I have not yet heard of FireWire-to-SATA bridges which do support ATA
>> pass-through.
> 
> That makes sense, unfortunate, but not unexpected.
> 
> Can you tell me how my system is getting the Seagate drive information
> then?  I've looked over the data available to me on the Firewire bus,
> and don't see where it's being pulled from.

According to a quick look at the smartmontools source, these data come
from the SCSI INQUIRY command.

FireWire firmwares insert all sorts of data into their INQUIRY
responses.  Sometimes the response is hardcoded by the firmware author,
sometimes it is built from data which were transparently obtained from
the attached SATA disk --- like with your Seagate disk.
-- 
Stefan Richter
-=====-==--= --=- ---==
http://arcgraph.de/sr/