Skip to contentSkip to conversation

IDENTIFY DEVICE not working any more?

DISCUSSION Dieter Stüken 3 participants 5 messages Aug 15, 2005 libata

lore
Share & tools
lore
#0 lore raw

I used smartmontools for sata devices quite successfully awhile. Since 2.6.12.2 however, I get an ioerror on "DENTIFY DEVICE" errno=25 [Inappropriate ioctl for device] and smartctl gives up.

So, is the prelimnary SMART support over ATA commands broken again and I have to wait patiently until smart over SCSI commands will work some day.

I feel like flying blind with my 3.5TB SATA data without SMART....

Dieter.

Signature · 4 lines
Dieter Stüken, con terra GmbH, Münster
    [email protected]
    http://www.conterra.de/
    (0)251-7474-501
#1 lore raw
Quoted 7 lines · Dieter Stüken

I used smartmontools for sata devices quite successfully awhile. Since 2.6.12.2 however, I get an ioerror on "DENTIFY DEVICE" errno=25 [Inappropriate ioctl for device] and smartctl gives up.

So, is the prelimnary SMART support over ATA commands broken again and I have to wait patiently until smart over SCSI commands will work some day.

Over here it works fine (sata_promise driver + 2.6.12.3 + smartmontools).

Also afaik the .1/.2/.3 patches did not contain changes to the libata or scsi layer.

#2 lore raw
Quoted 11 lines · Erik Slagter
Quoted 3 lines · Dieter Stüken

I used smartmontools for sata devices quite successfully awhile. Since 2.6.12.2 however, I get an ioerror on "DENTIFY DEVICE" errno=25 [Inappropriate ioctl for device] and smartctl gives up.

Over here it works fine (sata_promise driver + 2.6.12.3 + smartmontools).

Also afaik the .1/.2/.3 patches did not contain changes to the libata or scsi layer.

strange. I Have written a tiny test with: rc = ioctl(fd, HDIO_GET_IDENTITY, &id); and definitely got: "/dev/sda: Inappropriate ioctl for device" with 2.6.12.2. I did not test any other 2.6.12.x, but it works fine for an other 2.6.11 system.

Unfortunalely I can't install an other kernel today, as all my systems are in production. Is there any new kernel config I missed? The only difference ist, that my 2.6.12 system is a x86_64, where as the 2.6.11 system is an i686. I'll try some constellations later at home, where I have an other x86_64 system.

Dieter.

Signature · 4 lines
Dieter Stüken, con terra GmbH, Münster
    [email protected]
    http://www.conterra.de/
    (0)251-7474-501
#3 lore raw

Were the 2.6.11 kernels you have included with a distribution such as ubuntu, fedora, etc? If it was, there's a good chance that it has the passthru libata patch, that allows you to access drives using smartctl, with the special ata flag.

Tyler.

Quoted 26 lines · Dieter Stüken
Quoted 7 lines · Erik Slagter
Quoted 3 lines · Dieter Stüken

I used smartmontools for sata devices quite successfully awhile. Since 2.6.12.2 however, I get an ioerror on "DENTIFY DEVICE" errno=25 [Inappropriate ioctl for device] and smartctl gives up.

Quoted 5 lines

Over here it works fine (sata_promise driver + 2.6.12.3 + smartmontools).

Also afaik the .1/.2/.3 patches did not contain changes to the libata or scsi layer.

strange. I Have written a tiny test with: rc = ioctl(fd, HDIO_GET_IDENTITY, &id); and definitely got: "/dev/sda: Inappropriate ioctl for device" with 2.6.12.2. I did not test any other 2.6.12.x, but it works fine for an other 2.6.11 system.

Unfortunalely I can't install an other kernel today, as all my systems are in production. Is there any new kernel config I missed? The only difference ist, that my 2.6.12 system is a x86_64, where as the 2.6.11 system is an i686. I'll try some constellations later at home, where I have an other x86_64 system.

Dieter.

Signature · 3 lines
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.338 / Virus Database: 267.10.8/71 - Release Date: 8/12/2005
#4 lore raw
Quoted 1 line · Tyler

Were the 2.6.11 kernels you have included with a distribution such as ubuntu, fedora, etc? If it was, there's a good chance that it has the passthru libata patch, that allows you to access drives using smartctl, with the special ata flag.

That may be the explanation: I used prebuild kernels from SuSE and got the illusion the SMART enabled libata went into the recent kernels since 2.6.9 by default. So I'll try with the libata patch explicitely again.

Thanks for the tip.

Dieter.

Signature · 4 lines
Dieter Stüken, con terra GmbH, Münster
    [email protected]
    http://www.conterra.de/
    (0)251-7474-501