[smartmontools-support] "Unexpected sense" errors logged on Dell PERC H700 controllers

[Christian Franke]

Terry Kennedy wrote:
>    All systems are FreeBSD 12 (amd64) and all have smartmontools 7.0. All of
> the systems also have Dell PERC H700 controllers. That last part is import-
> ant, as those controllers are apparently more picky than the drives about
> what sort of commands they will pass through.
>
>    At smartd startup and during the nightly periodic job that checks the
> health of the drives, the controller logs a message like this for each
> drive:
>
> mfi0: 14428 (614612552s/0x0002/info) - Unexpected sense: PD 02(e0x20/s2) Path 5000cca02a444485, CDB: 4d 00 40 ff 00 00 00 3e fc 00, Sense: 5/24/00
>
>    Running smartmontools with the "-r ioctl,4" option to generate a trace
> shows a difference between the systems that log the unexpected sense errors
> and ones that don't.
>
>    System with the problem:
>
> [SNIP]
> === START OF INFORMATION SECTION ===
> Vendor:               HITACHI
> Product:              HUS156030VLS600
> Revision:             E774
> Compliance:           SPC-4
> User Capacity:        300,000,000,000 bytes [300 GB]
> ...
>   [log sense: 4d 00 40 00 00 00 00 00 12 00 ]
>    CAM status=0x1, SCSI status=0x0, resid=0x0
>    Incoming data, len=18:
>   00     00 00 00 0e 00 02 03 05  06 0d 0e 0f 10 15 18 2f
>   10     30 37
>    status=0x0
>   [log sense: 4d 00 40 ff 00 00 00 3e fc 00 ]
>    CAM status=0x8c, SCSI status=0x2, resid=0x0
>    Incoming data, len=16124 [only first 256 bytes shown]:
>   00     00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00
> ...
>   f0     00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00
>    sense_len=0x20, sense_resid=0x0
>    >>> Sense buffer, len=32:
>   00     70 00 05 00 00 00 00 18  00 00 00 00 24 00 00 cf
>   10     00 03 00 00 f8 23 00 00  00 00 00 00 00 00 00 00
>    status=0x2: sense_key=0x5 asc=0x24 ascq=0x0
> Log Sense for supported pages and subpages failed [unsupported field in scsi command]
> scsiGetSupportedLogPages: number of unreported (standard) log pages: 1 (sub-pages: 0)
> [SNIP]
>
>    System without the problem:
>
> [SNIP]
> === START OF INFORMATION SECTION ===
> Vendor:               HITACHI
> Product:              HUS154530VLS300
> Revision:             B598
> Compliance:           SPC-3
> User Capacity:        300,000,000,000 bytes [300 GB]
> ...
>   [log sense: 4d 00 40 00 00 00 00 00 12 00 ]
>    CAM status=0x1, SCSI status=0x0, resid=0x0
>    Incoming data, len=18:
>   00     00 00 00 0e 00 02 03 05  06 0d 0e 0f 10 15 18 2f
>   10     30 37
>    status=0x0
> scsiGetSupportedLogPages: number of unreported (standard) log pages: 1 (sub-pages: 0)
> [SNIP]
>
>    The difference is that the failing systems all do a third LOG SENSE
> command requesting a subpage of 0xff.
>
>    Does anyone have any ideas as to what is going on here, in particularly
> why smartctl is doing a third LOG SENSE for some drives and not for others?
> Note that while the drive models are different, they report the same pages
> as available.
>

The difference is that the HUS156030VLS600 advertises compliance with 
SPC-4 (SCSI Primary Commands - 4) which adds log subpages. Then smartctl 
 >= r4679 queries "Supported Log Pages and Subpages log page" via LOG 
SENSE page/subpage 0x00/0xff. This may fail in this case because drive 
firmware does not implement it (which would possibly violate SPC-4) or 
controller pass-through is not SPC-4 compatible.

Related patch: https://www.smartmontools.org/changeset/4679
I forwarded your report to the patch author.

Thanks,
Christian