Jump to content











Photo
- - - - -

Tool to read the entire SMART table regardless of manufacturer


  • Please log in to reply
3 replies to this topic

#1 Uneitohr

Uneitohr

    Frequent Member

  • Advanced user
  • 219 posts

Posted 10 February 2017 - 03:43 PM

Hello,

 

There are times when I need to make sure that a drive is failing. Usually if there are no signs of it's failure, I go look at the SMART table using either HDTune or HDD Sentinel. The problem is that some of these HDDs and SSDs have proprietary SMART elements that cannot be read by any other applications other than the ones made by manufacturer. It is very tedios and stupid to install one tool for each HDD manufacturer you have on a system.

 

I'm looking for something that can give me all SMART elements regardless of manufacturer. Does something like this exist, and can you elaborate further if it does?

 

Thank you



#2 steve6375

steve6375

    Platinum Member

  • Developer
  • 7566 posts
  • Location:UK
  • Interests:computers, programming (masm,vb6,C,vbs), photography,TV,films
  •  
    United Kingdom

Posted 10 February 2017 - 07:15 PM

I haven't fully tested it, but PassMark's diskcheckup is maintained and I would expect it to have been tested on a wide range of drives.

http://www.passmark....diskcheckup.htm

 

In my experience the Status (OK/FAIL) and threshold settings set by the manufacturer can be extremely lenient.

I have seen very sick drives which have not shown any smart values below the threshold values (i.e. 'Pass').

Even a few brand new drives (which may have been mis-handled in our factory) showed 100% SMART values but had defects.

 

Drives have lots of ways of repairing themselves (by mapping out bad sectors with good ones), retries, error correction, etc.

So as long as there are plenty of spare unused replacement sectors, SMART will not fail a drive.

 

PassMarks approach of monitoring the SMART values is best, because you can see if you are getting more defects or errors with time.

If you wait for the SMART values to go below the threshold, it will be too late!

 

At the end of the day, you still will need a regular backup regime (e.g. offline USB 3 HDD or cloud storage), so I tend not to worry about SMART monitoring (I can usually hear if the hard drive is having problems!).


  • Uneitohr likes this

#3 Uneitohr

Uneitohr

    Frequent Member

  • Advanced user
  • 219 posts

Posted 10 February 2017 - 11:55 PM

Very interesting, Steve.

So, what would be a better way to test if a drive will fail or not? Assuming there are no weird noises, and no miss-handling of any kind.

My thinking is bad sector test will be useless as long it remaps those with spare ones.

What about a speed test? I've seen drives that go from 0MB to 66MB.



#4 steve6375

steve6375

    Platinum Member

  • Developer
  • 7566 posts
  • Location:UK
  • Interests:computers, programming (masm,vb6,C,vbs), photography,TV,films
  •  
    United Kingdom

Posted 11 February 2017 - 10:12 AM

Since HDD controllers will remap sectors, you can only check a drive by looking at such things as the reallocated sector count over time.

That is what PassMarks s/w does.

You can look at the reallocated raw sector count - but this varies widely depending on model and usage! For instance a high-density 2.5" drive in a laptop would have a lot more reallocated sectors after 6 months than a 10,000 rpm 5.25" Server HDD.

 

You can read contiguous disk sectors and look for any sector reads which take longer that usual.

This may indicate bad sectors which have to be retried or even ECC corrected by the HDD controller - however, some of the sectors may actually be reallocated sectors on a different track - in this case you would expect the read to take longer because the head has to go to a different track for that block read. So it is normal for an old 2.5" laptop HDD to have many 'longer' reads.

 

The DiskDoctor is a Windows .exe and is part of RMPrepUSB (CTRL-D). It can do sequential scans of a HDD and look for sectors which take longer than the normal time to be read. However, for an 'old' drive, this does not necessarily mean it is currently 'bad' because it may simply be reading reallocated sectors. On a brand new drive, you should see very few 'excess time' read warnings.

 

I am afraid predicting an HDD failure is difficult. The most common problem is to have a head crash due to mishandling. In this event you will have lost data. So making regular backups is the only real solution.

 

And use your ears - if you hear it whistling/whining or sometimes the head seems to recalibrate, grate, stutter or step slowly during accesses, then that is a good warning!






2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users