Limited SMART output on Samsung PM1633a when upgrading from smartctl 6.5

[yhemery]

Hi,

When upgrading smartctl from version 6.6 to 7.0, most of the SMART informations are no longer displayed on a Samsung PM1633a. The inquiry is still working, but there is no selftest results, no phy details, and so on.

The server setup is :

• Supermicro 2028R-DE2CR24L, with an embedded LSI 3008 HBA + an LSI SAS3x40 expander
• 24x Samsung MZILS3T8HMLH/007 (PM1633a)
• Debian 8 + custom kernel (4.19.67)

smartctl is compiled with the following arguments :

--with-nvme-devicescan --with-gnupg=no --with-initscriptdir=no --with-exampledir=no --with-drivedbdir=no --with-smartdscriptdir=no --with-smartdplugindir=no --with-systemdsystemunitdir=no

I added the output of the following commands, for both version 6.6 and 7.0 :

• smartctl -x /dev/sda
• smartctl -x /dev/sda -r ioctl,2

Feel free to ask if you need more details.

As a side note : Ticket #1225 might be the same issue, but I'm not sure of that.

[yhemery]

[6.6] smartctl -x /dev/sda

[yhemery]

[6.6] smartctl -x /dev/sda -r ioctl,2

[yhemery]

[7.0] smartctl -x /dev/sda

[yhemery]

[7.0] smartctl -x /dev/sda -r ioctl,2

[yhemery]

It might be related to an issue with the "Supported log pages and subpages" command :

# sg_logs --page=0x00 /dev/sda
    SAMSUNG   MZILS3T8HMLH/007  GXL0
Supported log pages  [0x0]:
    0x00        Supported log pages
    0x02        Error counters (write)
    0x03        Error counters (read)
    0x05        Error counters (verify)
    0x06        Non-medium errors
    0x0d        Temperature
    0x0e        Start-stop cycle counter
    0x0f        Application client
    0x10        Self-test results
    0x11        Solid state media
    0x15        Background scan results
    0x18        Protocol specific port
    0x1a        Power condition transition
    0x2f        Informational exceptions (SMART)
# sg_logs --page=0x00,0xff /dev/sda
    SAMSUNG   MZILS3T8HMLH/007  GXL0
Supported log pages and subpages  [0x0, 0xff]:
    0x00,0xff   Supported log pages and subpages
    0x34,0xff   ??

I'll check with Samsung if that's a known behaviour.

[Christian Franke]

Possibly related: ticket #1225.

[Doug Gilbert]

Subversion revision 4958 (of smartmontools version 7.1) should fix this problem, I hope. The firmware problem that caused this has been reported to Samsung.

[yhemery]

Hi,

That's fixed with r4958 indeed, thanks !

I also reported this to Samsung last week, but got no answer yet.

[Christian Franke]

Fixed in r4958. Thanks!

[Christian Franke]

Tickets #1225 and #1258 have been marked as duplicates of this ticket.

[yhemery]

Hi,

I have a very similar issue with the Samsung PM1643 : Even with the fix r4958, some LOG pages like the selftest results are not displayed. Looks like the "Supported log pages and subpages" page is incomplete, as for the PM1633a.

# sg_logs /dev/sg10
    SAMSUNG   MZILT15THMLA/007  GXF2
Supported log pages  [0x0]:
    0x00        Supported log pages
    0x02        Error counters (write)
    0x03        Error counters (read)
    0x05        Error counters (verify)
    0x06        Non-medium errors
    0x0d        Temperature
    0x0e        Start-stop cycle counter
    0x0f        Application client
    0x10        Self-test results
    0x11        Solid state media
    0x15        Background scan results
    0x18        Protocol specific port
    0x1a        Power condition transition
    0x2f        Informational exceptions (SMART)
# sg_logs --page=0x00,0xff /dev/sg10
    SAMSUNG   MZILT15THMLA/007  GXF2
Supported log pages and subpages  [0x0, 0xff]:
    0x00,0xff   Supported log pages and subpages
    0x0d        Temperature
    0x0d,0x01   ??
    0x0d,0x02   ??
    0x0d,0xff   Temperature and subpages
    0x15        Background scan results
    0x15,0x01   Pending defects
    0x15,0xff   ??

Doug, have you got any feedback from Samsung on this topic ? Maybe they already have a fix for this model.

Also, feel free to ask if you want me to open a new ticket.

Yannick.