Interpreting SMART Details - Hitachi

[Luke]

Rather than clutter a recent similar discussion concerning Seagate HDDs this thread concerns the SMART data for a Hitachi HDD that I've just installed in an HDR-FOX T2.
The 2TB Hitachi HDD was sourced from a Sky recorder.
EDITED
The main question has been removed from this post as the thread immediately turned into a thread about the wording of the add on question.
A separate thread has been started for just the main question.

Also is the Power_On_Hours suspect? The Life Left interpretation is 94% but if this were a 2TB Seagate I'd expect the Life Left for 48504 to be 47% which is very different from the 94% for this Hitachi

 

[Black Hole]

Any thoughts on how the Power_Cycle_Count and Power-Off_Retract_Count became so different?

It's essential to check Hitachi's exact interpretation of those attribute numbers. There is significant variation between manufacturers (and even maybe models), so the default smrtctl output might not be valid.

Also is the Power_On_Hours suspect? The Life Left interpretation is 94% but if this were a 2TB Seagate I'd expect the Life Left for 48504 to be 47% which is very different from the 94% for this Hitachi

I don't understand that; the Seagate PoH wraps around at 65536, so I don't see how 48504 can be 47% (unless the value stored internally is 32-bit but smrtctl only reads the bottom 16).

 

[Luke]

It's essential to check Hitachi's exact interpretation of those attribute numbers. There is significant variation between manufacturers (and even maybe models), so the default smrtctl output might not be valid.

Yes you are correct. As I stated in the the starting post as this is a Hitachi I am starting a new thread so that it doesn't get confused with the Seagate interpretations.
As I have failed to come up with anything concerning Hitachi, or even non-Hitachi, attributes for this that would explain it that is why I am asking.

I don't understand that; the Seagate PoH wraps around at 65536, so I don't see how 48504 can be 47% (unless the value stored internally is 32-bit but smrtctl only reads the bottom 16).

I am comparing to non-wrap round Seagates that I own and using .. To do a comparison with a wrapped round Seagate without taking into account the wrap round would be meaningless, and also wouldn't add anything.

 

[prpr]

the Seagate PoH wraps around at 65536

Not universally true. In fact I would say it's almost completely untrue. I have had personal experience of dozens of Seagate drives with well over 80000 hours and they all reported hours correctly.

 

[Black Hole]

What can I say? My always-on HDR1 purchased around 2010 currently says PoH = 10515 (through WebIF >> Diagnostics >> Disk Diagnostics). I have never seen a value greater than 65535. (As a rule-of-thumb: 10,000 hours ≈︎ 1 year.)

 

[MymsMan]

What can I say? My always-on HDR1 purchased around 2010 currently says PoH = 10515 (through WebIF >> Diagnostics >> Disk Diagnostics). I have never seen a value greater than 65535. (As a rule-of-thumb: 10,000 hours ≈︎ 1 year.)

Not doubting your figures but it is bizarre that any manufacturer would use such a small field for such a key statistic - the design lifetime for a hard disk is well in excess of 6 years!

 

[Black Hole]

it is bizarre that any manufacturer would use such a small field for such a key statistic

Maybe they haven't – maybe it's a limitation of the smrtctl version on HDR-FOX reading, interpreting, or presenting the data. My point is that without further information we don't really know what these figures represent. A lot of it is supposition, which has gained currency by correlating with observation.

 

[prpr]

maybe it's a limitation of the smartctl version on HDR-FOX reading, interpreting, or presenting the data. My point is that without further information we don't really know what these figures represent.

Power_On_Hours is a pretty clear measure though.
The smartctl version supports up to 24 bit native display for this quantity and will also display another 24 bits (as three lumps of 8) if they are non-zero, so that's easily up to 16.7 million. I just tested it by re-compiling it having injected some dummy values:
9 Power_On_Hours 0x0032 075 075 000 Old_age Always - 16777215 (255 255 255)
It's safe to assume therefore that the data is coming from the disk like that.
Is it a standard Pipeline 5900.2 500GB one? I don't have one old enough.

 

[Black Hole]

It's safe to assume therefore that the data is coming from the disk like that.

Well there ya go. However clear "power on hours" might seem to be, it still carries terms and conditions. That's my whole point: without an actual spec from each individual manufacturer and possibly each model, the interpretation of the smart attributes remains hearsay.

Is it a standard Pipeline 5900.2 500GB one? I don't have one old enough.

My three active HDRs retain their original HDDs and must all be at least 10 years old. I've taken to turning one of them off when not specifically in use, but only recently. Edited highlights:

HDRFOX1# smartctl -a /dev/sdb

Model Family:     Seagate Pipeline HD 5900.2
Device Model:     ST3500312CS

  9 Power_On_Hours          0x0032   088   011   000    Old_age   Always       -       10530

Powered_Up_Time is measured from power on, and printed as
DDd+hh:mm:SS.sss where DD=days, hh=hours, mm=minutes,
SS=sec, and sss=millisec. It "wraps" after 49.710 days.

^^^ I have no idea what this means! ^^^

vvv If we assume the "Error" count below is logged sequentially, clearly the "power-on lifetime" isn't monotonic! vvv
                                                                                                                                                                          
Error 11 occurred at disk power-on lifetime: 10298 hours (429 days + 2 hours)

Error 10 occurred at disk power-on lifetime: 7590 hours (316 days + 6 hours)
Error 9 occurred at disk power-on lifetime: 56214 hours (2342 days + 6 hours)
Error 8 occurred at disk power-on lifetime: 56214 hours (2342 days + 6 hours)
Error 7 occurred at disk power-on lifetime: 56214 hours (2342 days + 6 hours)

And yet, to disprove my hypothesis:

HDRFOX3# smartctl -a /dev/sdb

Model Family:     Seagate Pipeline HD 5900.2
Device Model:     ST3500312CS

  9 Power_On_Hours          0x0032   017   017   000    Old_age   Always       -       73126

HDRFOX4# smartctl -a /dev/sda

Model Family:     Seagate Pipeline HD 5900.2                                                                                                                              
Device Model:     ST3500312CS                                                                                                                                             

  9 Power_On_Hours          0x0032   021   021   000    Old_age   Always       -       69753

...so why the heck is that??? There's no difference in the version of smartctl.

 

[prpr]

There's no difference in the version of smartctl.

What about the disks' info.?

Anyway, your blanket statement in #2 about wrapping is clearly wrong. What is it you say about extrapolating from a sample size of 1?

 

[Black Hole]

What about the disks' info.?

They're all firmware version "SC13". What else would you like to know?

Anyway, your blanket statement in #2 about wrapping is clearly wrong. What is it you say about extrapolating from a sample size of 1?

Fair enough, but you can't make a blanket statement that they don't wrap either. Until we know how to explain it, we can make no assessment whether it might be a common occurrence.