[smartmontools-support] SAMSUNG HD103SJ failures - Spin_Up_Time + IDNF

[Christian Franke]

Ohad Schneider wrote:
> I have a couple of Samsung HD103SJ hard drives with a few smartmontools failures (using the GSmartControl wrapper on Windows).
> The important details follow, and please find below the full output for both hard drives.
>
> ------------
> Failure 1
> ------------
> Spin_Up_Time            PO---K   071   001   025    Past 8895
>
> This only happens in one of the hard drives, and I suspect it might be a real issue.
> However, I wonder how seriously I should treat it since it only happened in the "Past"?

I don't know, but ...


> 9 Power_On_Hours          -O--CK   100   100   000    -    85921

... ~10 years of power on time suggest that the drive should be replaced :-)


> ------------
> Failure 2
> ------------
> 10 -- 51 00 20 00 00 74 70 6d af e0 00  Error: IDNF at LBA = 0x74706daf = 1953525167
> 10 -- 51 00 20 00 00 74 70 6d b0 e0 00  Error: IDNF at LBA = 0x74706db0 = 1953525168
>
> These I get in both drives, and I suspect some OS / driver is at fault (but I could be wrong of course, far from an expert in this area).
> The reason for my suspicion is that these are the absolute last sectors on the disk, so it seems unlikely to be coincidence for only the very last 2 sectors to fail.

You could possibly ignore these errors as the last occurrence was 
reported more than 4 years ago:

> Error 3 [2] occurred at disk power-on lifetime: 48476 hours (2019 days + 20 hours)

The last few sectors are normally not used by regular partitions, but 
may be used for RAID Metadata (e.g. Intel RST RAID) or a secondary GPT.


> Executing "wmic diskpart" on seemingly supported this theory, as it stated the drive only has 1953520065 total sectors (less than the reported 1953525167 / 1953525168).
> However, executing "smartctl -I" and dividing the amount of "user capacity" by "bytes per sector" I get the exact number of the IDNF: 1000204886016 / 512 = 1953525168

Apparently WMI reports a fake total sectors value compatible with 
traditional CHS addressing.

1953520065 = 121601 * 255 * 63 (C * H * S fake values)

"diskmgmt.msc" should report "931.51 GB" (which actually means GiB) for 
this drive which matches the value printed by smartctl.


> ...
> --------
> TESTS
> --------
> Both tests pass the short & extended self-tests.
> The strange thing is that after I executed the tests, the Lifetime(hours) value of the newly executed tests was LOWER than the value for past tests (as if the past tests ran AFTER the test I just executed).

 From smartctl man page:
"[Note: this time stamp wraps after 2^16 hours, or 2730 days and 16 
hours, or about 7.5 years.]"


> ,,,
> P.S. is this the "old" mailing list? If so, what is the new one?

This is the new one.


> Have you considered moving to a more modern Q&A platform like StackExchange (even use an existing one like ServerFault)?

See similar discussion 9+ years ago:
https://www.smartmontools.org/ticket/217


Thanks,
Christian