TechSpot

What things doesn't SMART do?

By rodion15
Oct 11, 2015
Post New Reply
  1. I've heard something about SMART offline scans, I wonder what they do, are they scanning the whole surface the drive for sector readability? I don't think so, otherwise you wouldn't need deep scan apps such chkdsk or HDDscan right?

    I'd appreciate advice on a good deep article explaining SMART (besides Wikipedia).
     
  2. jobeard

    jobeard TS Ambassador Posts: 9,311   +617

    S.M.A.R.T is onboard ecc data collected by the drive itself. However, the wiki I found (using your SMART notation) is entirely a different thing.
     
    cliffordcooley likes this.
  3. rodion15

    rodion15 TS Booster Topic Starter Posts: 130

    LOL, funny TS ambassador, read that article head to tail, absolutely boooring!
     
  4. jobeard

    jobeard TS Ambassador Posts: 9,311   +617

    Hmm, what were you looking for / expecting ?
     
  5. jobeard

    jobeard TS Ambassador Posts: 9,311   +617

    As an example, S.M.A.R.T data recorded by the HDD Driver includes:

    ID Attrib Name real current worst Threshold Raw
    1 Read Error Rate 0 100 100 51 0x000000000002
    2 Throughput Performance 0 252 252 0 0x000000000000
    3 Spin-Up Time 1867 ms 94 90 25 0x00000000074B
    4 Start/Stop Count 639 100 100 0 0x00000000027F
    5 Reallocated Sectors Cnt 0 252 252 10 0x000000000000
    7 Seek Error Rate 0 252 252 51 0x000000000000
    8 Seek Time Performance 0 252 252 15 0x000000000000
    9 Power-On Hours (POH) 76d 20h 100 100 0 0x000000000734
    10 Spin Retry Count 0 252 252 51 0x000000000000
    11 Recalibration Retries 183 100 100 0 0x0000000000B7
    12 Power Cycle Count 644 100 100 0 0x000000000284
    191 G-sense Error Rate 57 100 100 0 0x000000000039
    192 Power-off Retract Count 2 100 100 0 0x000000000002
    194 Temperature 34 °C 64 59 0 0x000000120022
    195 Hardware ECC Recovered 0 100 100 0 0x000000000000
    196 Reallocation Event Cnt 0 252 252 0 0x000000000000
    197 Pending Sector Count 0 252 252 0 0x000000000000
    198 Uncorrectable sctr cnt 0 252 252 0 0x000000000000
    199 UltraDMA CRC Error cnt 0 200 200 0 0x000000000000
    200 Multi-Zone Error Rate 14409 100 100 0 0x000000003849
    223 Load/Unload Retry cnr 183 100 100 0 0x0000000000B7
    225 Load/Unload Cycle cnt 54449 95 95 0 0x00000000D4B1 ​

    All of which is recorded in real-time, without running any diagnostic whatsoever.
     
  6. jobeard

    jobeard TS Ambassador Posts: 9,311   +617

    Btw:
    mathematically, it's impossible to prove a negative.
     
  7. rodion15

    rodion15 TS Booster Topic Starter Posts: 130

    I just need to know what those offline scans do, for example: do they scan the whole surface of the disk hunting bad sectors? If that is so, it'd be needless to use a surface scan such as chkdisk can do.
    Thanks for your answers once again (y)
     
  8. jobeard

    jobeard TS Ambassador Posts: 9,311   +617

    Maybe you know something I don't, but I've never seen SMART associated with surface scanning. Examining the counters shown above, there's only one that imo *might* imply surface status, "Reallocated Sectors Cnt", but it is not.

    All the counters are of the type "I tried this (xyz) and the return status was not normal."
    eg Recalibration --> I asked for sector x but got back other than x, invoke calibration.

    If you need surface testing, then by all means invoke CHKDSK C: /F /R
     
  9. cliffordcooley

    cliffordcooley TS Guardian Fighter Posts: 8,549   +2,894

    We may be getting confused with Disk Utilities that also monitor SMART stats.
     
  10. jobeard

    jobeard TS Ambassador Posts: 9,311   +617

    Perhaps, but SMART is onboard diags totally within the HD itself, ergo NO surface scanning,
     
  11. rodion15

    rodion15 TS Booster Topic Starter Posts: 130

    OK, sorry I didn't expose my doubts properly, I'll try to explain myself as thoroughly as I can: What I learnt is that SMART performs "offline scans" (that's what Wikipedia says to name one source) and keeps a log of bad sectors found. So it's not strange to imagine that those scans are intended to detect bad sectors on the surface of the drive (meaning on the platters of the hard drive, where else?), thereby my doubt. My idea is that perhaps the bad sectors are only detected by SMART when they're used (read or written), and therefore, if there's a big chunk of the hard drive that hasn't been read/written for a long time and there were bad sectors in that area (due to a head crash on that area that would produce hard bad sectors or to a mismatch between the info in the sectors and the EEC, which would produce the so called soft bad sectors), they wouldn't be logged by SMART until those sectors where used some day in the future or if you use some utility that scans the whole surface of the drive or simply if you zero all the hard drive sectors or perform a full erase. But I'm not sure, thus my doubt: does the offline scan look for bad sectors? does it perform a scan of the whole surface?
    Thanks anyway.
     
    Last edited: Oct 21, 2015
  12. cliffordcooley

    cliffordcooley TS Guardian Fighter Posts: 8,549   +2,894

    What you are looking for is a utility to replace SMART functionality, when you should be thinking of a utility that triggers the Smart error. SMART will trigger if like you said read/write can't be performed. What you need is a utility that forces a bit flip of all addresses, so that SMART flags the bad locations. Personally I don't see why, when the locations will be triggered automatically on attempted standard use.
     
    jobeard likes this.
  13. rodion15

    rodion15 TS Booster Topic Starter Posts: 130

    You said you don't see why you may need to check the whole surface of a drive? For example: you want to create a hard drive with several operating systems and test tools for testing purposes on a used hard drive you have at the workshop, if you don't check the disk thoroughly before installing these tools in it, you may regret it later if there're bad sectors in it.
     
  14. cliffordcooley

    cliffordcooley TS Guardian Fighter Posts: 8,549   +2,894

    Fill up the drive, format drive (easier than deleting everything), then fill up drive again with different data. If that doesn't trigger errors, the drive is still good.

    or sit back and watch the following work
    and for reference an older thread on this topic
     

Similar Topics

Add New Comment

You need to be a member to leave a comment. Join thousands of tech enthusiasts and participate.
TechSpot Account You may also...