Hi, All
I don’t know whether asking this question here is appropriate.
It is about hard disk drive S.M.A.R.T (Self-Monitoring, Analysis and Reporting Technology).
I can get the smart information(such as attribute value,attribute thresholds…)
I know that if no attribute value is less than or equal to their corresponding attribute thresholds,
then device reliability status is OK.
But I want to do an accurate description of the hard disk’s health and performance.
[I want to display it in digital form,such as:100 means the best,and 1 means an impending degrading]
Is there a standard algorithm to describe the hard disk’s health and performance?
Or I can design it myself(use the relevant attribute value and attribute thresholds to do some arithmetic)?
Any help is appreciated.Thanks.
Best Regards
Zhou ChengJun
xxxxx@yahoo.com.cn wrote:
I don’t know whether asking this question here is appropriate.
It is about hard disk drive S.M.A.R.T (Self-Monitoring, Analysis and Reporting Technology).
I can get the smart information(such as attribute value,attribute thresholds…)
I know that if no attribute value is less than or equal to their corresponding attribute thresholds,
then device reliability status is OK.
But I want to do an accurate description of the hard disk’s health and performance.
[I want to display it in digital form,such as:100 means the best,and 1 means an impending degrading]
Is there a standard algorithm to describe the hard disk’s health and performance?
Or I can design it myself(use the relevant attribute value and attribute thresholds to do some arithmetic)?
I’m delving into philosophy here, but I’m wondering whether this is
really useful. In a sense, disk drive health is Boolean: either there’s
nothing to worry about, or I need to replace the disk NOW.
(OK, I’ll allow three states: healthy, sick, dead.)
If I told you your disk drive was 80% healthy, what would you do with
that information? Don’t you really just want to be told “you need to
replace this drive”?
–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.
Well,the hard disk’s performance degrading with the growth of the use time.
I want to evaluate the rate of the degrading.
There are several relevant attribute and attribute thresholds
about the health and performance of HDD.
I want to do an overall evaluation of it.
zhou chengjun wrote:
Well,the hard disk’s performance degrading with the growth of the
use time.
I wasn’t aware that the physical performance of a hard drive degrades over time. Sure, the filesystem can become defragmented and so on, but is this really the case?
> I wasn’t aware that the physical performance of a hard drive degrades over time.
Well, if you think of it not as of a whole drive but as of collection of sectors, the OP’s statement makes a perfect sense. For example, consider the scenario when the power goes off while you are writing to the disk - in this case sectors that you are writing to may become unreadable, but the disk as a whole is still fully workable. Therefore, the damage may be just accumulating over time, and, at some point, reach the level when disk becomes unusable…
Anton Bassov
Now… I’m not a disk expert, but according to the storage experts I’ve worked with in the past couple of years, this hasn’t been true for a long time.
When the power is lost to a disk, there are two sources of power used to properly retract the heads: (a) the residual power remaining in the capacitors of the drive, (b) the current generated by the disk which is still spinning as a result of inertia – the electric motor serves as a mini generator during spin-down after power failure.
Together, these provide plenty of power to complete the write in progress and safely retract the heads.
Peter
OSR
I believe that most current hard drives do automatic sector replacement when
an in-use sector goes bad. I don’t know where all the ‘extra’ sectors are
located and that may vary by manufacturer and model, but if they are located
at the end of the drive, each time a ‘bad’ sector is going to be read it
will take time to seek to the replacement.
I have heard that the magnetic media becomes resistant to change over time.
Maybe that is a degrade from the OP POV. Maybe the drive’s spinning
mechanism begins to introduce vibration that requires retries more often as
the drive ages.
I suspect only the big manufacturers of hard drives can answer this question
completely, but since Seagate offers a five year warranty on most of their
drives it can’t be too bad. If you remember the old 8" 5MB (yes MB not GB)
drives that hung on the door of a Cromemco System running Cromix and CP/M,
that you can buy a 1TB SATA 300 drive for less than $150 (USD) is quite
astounding. I remember a 1.8GB SCSI drive for around $1000.
wrote in message news:xxxxx@ntdev…
> zhou chengjun wrote:
>
>> Well,the hard disk’s performance degrading with the growth of the
>> use time.
>
> I wasn’t aware that the physical performance of a hard drive degrades over
> time. Sure, the filesystem can become defragmented and so on, but is this
> really the case?
>
Peter Viscarola wrote:
Together, these provide plenty of power to complete the write in
progress and safely retract the heads.
Wasn’t this sold as a “feature”, called “autopark” or similar, maybe 10-15 years ago?
I remember having to run some utility called “park.com” on my old ST-225 and ST-251 drives before turning off my machine (after a session of Leisure Suit Larry…)
xxxxx@hotmail.com wrote:
> I wasn’t aware that the physical performance of a hard drive degrades over time.
>
Well, if you think of it not as of a whole drive but as of collection of sectors, the OP’s statement makes a perfect sense. For example, consider the scenario when the power goes off while you are writing to the disk - in this case sectors that you are writing to may become unreadable, but the disk as a whole is still fully workable. Therefore, the damage may be just accumulating over time, and, at some point, reach the level when disk becomes unusable…
But even if that did occur, that’s “protocol” damage. It wouldn’t be
detected by S.M.A.R.T.
–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.
xxxxx@yahoo.com.cn wrote:
I don’t know whether asking this question here is appropriate.
Give it also a try on the NTFSD forum (if you didn’t already).
There should be more file system developers reading and some of these
probably know a lot about hard drive internals.
Good luck!
> I believe that most current hard drives do automatic sector replacement when an in-use sector goes bad.
AFAIK, there are 2 types of “bad sectors”, i.e non-readable and non-writable. What you are speaking about is the latter type. When disk hardware detects non-writable sector it just replaces it and never uses it again, but data does not get lost. However, in addition to that sector may be unreadable because of electrical/mechanical/etc damage…
Anton Bassov
Well,SMART attributes includes:Raw Read Error Rate,Spin Up Time,Reallocated Sectors Count,
Seek Error Rate,Spin Retry Count and so on.Some of the attributes reflect the hard disk’s health and performance.I don’t know how to give an overall assessment…