[smartmontools-support] Differences in test report results: xselftest vs selftest

[Christian Franke]

Daniel Fishman wrote:
> When smartctl displays results of hdd's selftest, contents of the report
> differ, depending on whether you ask for an extended or a regular self 
> test.
> For extended test (smartctl -l xselftest /dev/sdc), it says:
>
> ---------------------------------------------------------------
> SMART Extended Self-test Log Version: 1 (1 sectors)
> Num  Test_Description    Status                  Remaining 
> LifeTime(hours)  LBA_of_first_error
> # 1  Extended offline    Completed: read failure       90% 
> 4899         281474968262399
> ...
> # 8  Extended offline    Completed: read failure       90% 
> 4863         7336633448
>
> ---------------------------------------------------------------
>
> For regular test (smartctl -l selftest /dev/sdc), it says:
> ---------------------------------------------------------------
> SMART Self-test log structure revision number 1
> Num  Test_Description    Status                  Remaining 
> LifeTime(hours)  LBA_of_first_error
> # 1  Extended offline    Completed: read failure       90% 
> 4899         4286519039
> ...
> # 8  Extended offline    Completed: read failure       90% 
> 4863         3041666152
> ---------------------------------------------------------------
>
> Also, 'fdisk -lu /dev/sdc' says:
>
> ---------------------------------------------------------------
> Disk /dev/sdc: 3.64 TiB, 4000752599040 bytes, 7813969920 sectors

0xffffff7f16ff = 281474968262399 (48-bit LBA is out of range)
0x0000ff7f16ff = 4286519039 (truncated to 32-bit field of old ATA-5 log)
0x0001b54c2468 = 7336633448 (48-bit LBA is not out of range)
0x0000b54c2468 = 3041666152 (truncated to 32-bit)
0x0001d1bfb7ff = 7813969919 (last valid LBA)

Looks like a firmware bug: LBA fields in the self-test logs are not set 
properly.


> From all this it seems as though *both* reports are incorrect: one 
> complains
> about a non-existent sector, another complains about a good readable 
> sector.
>
> How can I interpret those reports and use their results to fix the
> reported problem?

Are any read error reported in the SMART error logs?

If not, run a read test with badblocks or GNU ddrescue. If read errors 
occur, the last ones should appear in the error logs. Errors detected by 
self-tests only usually do not.

-- 
Regards,
Christian