SMART monitoring programs?????

[JohnJAdamson]

Hello.
What is the verdict regarding the accuracy of these applications?
I decided to run both SiGuardian and Active Smart on a computer with 2 IDE
hard drives. A Seagate drive which came with a new computer in 1999 and a
Western Digital 40GB drive that I bought second hand about a year ago.
I've not had any problems with them and they pass various diagnostic scans
with WD Lifeguard and scandisk but the SMART monitoring programs keep
reporting frequent changes (about twice a day) in the status, especially
giving a high 'seek error rate' with SiGuardian predicting a fail in both
drives sometime in October or Sept this year! (it keeps fluctuating it once
predicted the WD drive to fail in 2014)
Should I be worried by this? How accurate and likely is it that these drives
are about to die and do people pay much attention to these programs?
Thanks.

 

[Rod Speed]

What is the verdict regarding the accuracy of these applications?

There is no application accuracy with some of them like Everest
which just reports the drive SMART data and leaves you to
interpret it apart from what the drive says is acceptible etc.

I decided to run both SiGuardian and Active Smart on a computer with 2 IDE
hard drives. A Seagate drive which came with a new computer in 1999 and a
Western Digital 40GB drive that I bought second hand about a year ago.

I've not had any problems with them and they pass various diagnostic scans
with WD Lifeguard and scandisk but the SMART monitoring programs keep
reporting frequent changes (about twice a day) in the status, especially
giving a high 'seek error rate' with SiGuardian predicting a fail in both
drives sometime in October or Sept this year! (it keeps fluctuating it once
predicted the WD drive to fail in 2014)

Should I be worried by this?

Yes, seek errors arent desirable.

How accurate and likely is it that these drives are about to die

Post the Everest report.

and do people pay much attention to these programs?

I dont, I use the Everest listing myself.

 

[JohnJAdamson]

There is no application accuracy with some of them like Everest
which just reports the drive SMART data and leaves you to
interpret it apart from what the drive says is acceptible etc.




Yes, seek errors arent desirable.


Post the Everest report.


I dont, I use the Everest listing myself.

Hello there.
Here is the Everest report of the SMART data for the 2 drives:-

--------[
SMART ]-------------------------------------------------------------------------------------------------------

[ ST317242A (7BP00G4Z) ]

01 Raw Read Error Rate 0 84 81 146090271 OK:
Always passing
03 Spin Up Time 0 70 70 0 OK:
Always passing
04 Start/Stop Count 20 100 100 309 OK:
Value is normal
05 Reallocated Sector Count 36 100 100 1 OK:
Value is normal
07 Seek Error Rate 30 77 60 50852826 OK:
Value is normal
09 Power-On Time Count 0 87 1 0 OK:
Always passing
0A Spin Retry Count 97 100 100 0 OK:
Value is normal
0C Power Cycle Count 20 98 98 2570 OK:
Value is normal
C7 Ultra ATA CRC Error Rate 0 200 200 1 OK:
Always passing

[ WDC WD400EB-11CPF0 (WD-WMAAT7322572) ]

01 Raw Read Error Rate 51 200 200 0 OK:
Value is normal
03 Spin Up Time 21 98 91 2458 OK:
Value is normal
04 Start/Stop Count 40 99 99 1142 OK:
Value is normal
05 Reallocated Sector Count 140 200 200 0 OK:
Value is normal
07 Seek Error Rate 51 100 253 0 OK:
Value is normal
09 Power-On Time Count 0 95 95 3814 OK:
Always passing
0A Spin Retry Count 51 100 100 0 OK:
Value is normal
0B Calibration Retry Count 51 100 100 0 OK:
Value is normal
0C Power Cycle Count 0 99 99 1096 OK:
Always passing
C4 Reallocation Event Count 0 200 200 0 OK:
Always passing
C5 Current Pending Sector Count 0 200 200 0 OK:
Always passing
C6 Off-Line Uncorrectable Sector Count 0 200 200 0 OK:
Always passing
C7 Ultra ATA CRC Error Rate 0 200 200 0 OK:
Always passing
C8 Write Error Rate 51 200 200 0 OK:
Value is normal

 

[Rod Speed]

Sounds like it got seriously confused if you labelled the data below
properly, its the Seagate that has the very high seek error rate.

Hello there.
Here is the Everest report of the SMART data for the 2 drives:-

--------[
SMART
]-------------------------------------------------------------------------------------------------------

[ ST317242A (7BP00G4Z) ]

01 Raw Read Error Rate 0 84 81 146090271 OK:
Always passing
03 Spin Up Time 0 70 70 0
OK: Always passing
04 Start/Stop Count 20 100 100 309
OK: Value is normal
05 Reallocated Sector Count 36 100 100 1
OK: Value is normal
07 Seek Error Rate 30 77 60 50852826

Urk, thats clearly what they are worried about, thats an obscene number.

OK: Value is normal
09 Power-On Time Count 0 87 1 0
OK: Always passing
0A Spin Retry Count 97 100 100 0
OK: Value is normal
0C Power Cycle Count 20 98 98 2570
OK: Value is normal
C7 Ultra ATA CRC Error Rate 0 200 200 1
OK: Always passing

[ WDC WD400EB-11CPF0 (WD-WMAAT7322572) ]

01 Raw Read Error Rate 51 200 200 0
OK: Value is normal
03 Spin Up Time 21 98 91 2458
OK: Value is normal
04 Start/Stop Count 40 99 99 1142
OK: Value is normal
05 Reallocated Sector Count 140 200 200 0
OK: Value is normal
07 Seek Error Rate 51 100 253 0
OK: Value is normal
09 Power-On Time Count 0 95 95 3814
OK: Always passing
0A Spin Retry Count 51 100 100 0
OK: Value is normal
0B Calibration Retry Count 51 100 100 0
OK: Value is normal
0C Power Cycle Count 0 99 99 1096
OK: Always passing
C4 Reallocation Event Count 0 200 200 0
OK: Always passing
C5 Current Pending Sector Count 0 200 200 0
OK: Always passing
C6 Off-Line Uncorrectable Sector Count 0 200 200 0
OK: Always passing
C7 Ultra ATA CRC Error Rate 0 200 200 0
OK: Always passing
C8 Write Error Rate 51 200 200 0
OK: Value is normal

This drive is clearly fine.

I'd check which drive does have the high seek error rate and replace it myself.
It might actually be fine, but drives are so cheap I wouldnt risk it myself.

 

[Arno Wagner]

Hello.

What is the verdict regarding the accuracy of these applications? I
decided to run both SiGuardian and Active Smart on a computer with 2
IDE hard drives. A Seagate drive which came with a new computer in
1999 and a Western Digital 40GB drive that I bought second hand
about a year ago. I've not had any problems with them and they pass
various diagnostic scans with WD Lifeguard and scandisk but the
SMART monitoring programs keep reporting frequent changes (about
twice a day) in the status, especially giving a high 'seek error
rate' with SiGuardian predicting a fail in both drives sometime in
October or Sept this year! (it keeps fluctuating it once predicted
the WD drive to fail in 2014) Should I be worried by this? How
accurate and likely is it that these drives are about to die and do
people pay much attention to these programs?

Usually a SMART monitor is absolutely accurate, because it only
reports whet the disk tells it. If ist starts to predict things
not told to it by the disk (e.g. your 2014 value), it is in the
realm of wild speculation. Only the disk manufacturer can make
some presictions and even these would be speculative.

That said, if some SMART parameter is outside the manufactieres
specified "good" range (part of the respective SMART value),
then there is some serious problem. What exactly it means is
not clear and I have had drives that were near death yet had
a SMART status of "PASS".

As for your case of seek errors, that can have reasons outside
of the drive, like vibration, insufficient hard mounting od a
PSU that does not deliver stable voltages. The PSU theroey
sounds very likely to me, since two different drives seem to
display the same symptoms. Unlikely without a common problem,
i.e. one not found in the disks themselves.

And yes, I think you should be worried about this. A defect PSU
can damage or kill drives.

Arno

 

[Joep]

Hello.
What is the verdict regarding the accuracy of these applications?

Accuracy of apps is not an issue, all they should do is display what SMART
has to say. Some of the apps however take several samples and based on
trends they try predicting when SMART thresholds will be reached. IMO it's
bull.

It is SMART itself that tries to predict failure, If a treshold is reached,
then SMART predicts the disk will fail soon and that you need to replace it.
That's how it should work.

Of course there are disk that die without SMART ever reaching thresholds and
there's the disks SMART tells they're dying while there really is nothing
wrong (estimation, about 20%).

I decided to run both SiGuardian and Active Smart on a computer with 2 IDE
hard drives. A Seagate drive which came with a new computer in 1999 and a
Western Digital 40GB drive that I bought second hand about a year ago.
I've not had any problems with them and they pass various diagnostic scans
with WD Lifeguard and scandisk but the SMART monitoring programs keep
reporting frequent changes (about twice a day) in the status, especially
giving a high 'seek error rate'

For some disks, seek errors seem to be pretty 'standard'. Monitor SMART
values more closely, try to determine what is normal behavior for this disk,
and what is out of the ordinary. The seek error threshold can typically move
towards the threshold to recover later on.

SMART tools that jump to conclusions based on attribute values that change
back and forth from time to time will only make you nervous and they do add
a lot. They try to outsmart SMART and they can't. For example, if some
sectors are reallocated at some point, and another few half a year later,
they will draw a line towards a point in time the disk will fail if every
half year a few sectors are reallocated.

Studies show that many failed disks show a sudden decrease on seek errors,
grown defects and ecc error correction attributes. If you see one of those 3
drop down dramatically even when they do not reach thresholds I'd back up
some data to be on the safe side.

 

[Rod Speed]

Accuracy of apps is not an issue, all they should do is display
what SMART has to say. Some of the apps however take
several samples and based on trends they try predicting
when SMART thresholds will be reached. IMO it's bull.

Maybe, its got more potential than
the tiny little brain in the drive has tho.

It is SMART itself that tries to predict failure, If a treshold
is reached, then SMART predicts the disk will fail soon and
that you need to replace it. That's how it should work.

And it may well be possible to get some useful
info from the change in the data over time too.

Of course there are disk that die without SMART ever reaching
thresholds and there's the disks SMART tells they're dying while
there really is nothing wrong (estimation, about 20%).

For some disks, seek errors seem to be pretty 'standard'. Monitor
SMART values more closely, try to determine what is normal behavior
for this disk, and what is out of the ordinary. The seek error threshold
can typically move towards the threshold to recover later on.

SMART tools that jump to conclusions based on attribute values that
change back and forth from time to time will only make you nervous
and they do add a lot. They try to outsmart SMART and they can't.

Bullshit, of course they can.

For example, if some sectors are reallocated at some point, and
another few half a year later, they will draw a line towards a point in
time the disk will fail if every half year a few sectors are reallocated.

Doesnt have to be done that crudely.

Studies show that many failed disks show a sudden decrease on
seek errors, grown defects and ecc error correction attributes. If
you see one of those 3 drop down dramatically even when they do
not reach thresholds I'd back up some data to be on the safe side.

And a SMART monitor can advise the owner of the
drive based on that approach to drive replacement too.

 

[Joep]

Maybe, its got more potential than
the tiny little brain in the drive has tho.

I have not seen SMART monitors that do. Most just do hocus pocus with SMART
data trying to make the user feel a lot of intelligent analysis is going on.

And it may well be possible to get some useful
info from the change in the data over time too.

Yes: if attribute xx keeps developing towards the threshold in the same
avarage rate till now then the disk will fail in yy:mm:dd. There's to many
if's there to be useful.

The rest of the info such a tool will give is funny to look at and that's
about it.

Bullshit, of course they can.

Not by trying to predict treshold exceed conditions.

reallocated.

Doesnt have to be done that crudely.

In essence, yes. Of course the reallocated sector attribute won't get better
after it dropped, an attribute like seek error rate can improve again. So
there you have yet another system tray annoyance nagging an attribute
dropped, oh no, it improved again ... etc.

And a SMART monitor can advise the owner of the
drive based on that approach to drive replacement too.

Which ones do that? I must admit is has been a few months since I tried
about every SMART monitor I could find and none of those could. It was very
obvious that if they tried to predict TEC, this was implemented by a couple
of simpletons.

 

[Rod Speed]

I have not seen SMART monitors that do.

Irrelevant to whether its obviously possible to do what
you are doing yourself but implemented in a monitor.

Most just do hocus pocus with SMART data trying to
make the user feel a lot of intelligent analysis is going on.

Irrelevant to whether its obviously possible to do what
you are doing yourself but implemented in a monitor.

Yes: if attribute xx keeps developing towards the threshold in
the same avarage rate till now then the disk will fail in yy:mm:dd.

That aint the only possibility, what you say you do
at the bottom is obviously a useful approach too.

There's to many if's there to be useful.

Nope. Perfectly feasible to automate what you claim
to do at the bottom, for those who dont know enough
about the basics to work that stuff out for themselves.

The rest of the info such a tool will give
is funny to look at and that's about it.

Thats just what the current monitors do, not what is obviously possible.

Not by trying to predict treshold exceed conditions.

Thats just one approach.

In essence, yes.
Nope.

Of course the reallocated sector attribute won't get better
after it dropped, an attribute like seek error rate can improve
again. So there you have yet another system tray annoyance
nagging an attribute dropped, oh no, it improved again ... etc.

Doesnt have to do it like that either.

Which ones do that?

Irrelevant to what is clearly possible and whether that is
better than what the tiny little brain in the drive itself can do.

I must admit is has been a few months since I tried about
every SMART monitor I could find and none of those could.
It was very obvious that if they tried to predict TEC, this
was implemented by a couple of simpletons.

Irrelevant to what is clearly possible and whether that is
better than what the tiny little brain in the drive itself can do.

 

[Joep]

Irrelevant to what is clearly possible and whether that is
better than what the tiny little brain in the drive itself can do.

Ah I see. Yes I was commenting on *existing* monitors and I assumed OP was
asking about monitors currently available: "What is the verdict regarding
the accuracy of these applications?". IMO current monitors do not add a lot
of intelligence and all do pretty simple linear predictions which are not
reliable at all, they just nag the heck out of you.

And yes I agree, I think it is possible analyse trends and behavior of SMART
attributes and 'say things' based on that analysis better than current
monitors do.