On Tue, 18 Nov 2003 18:25:55 +0100, "Folkert Rienstra"
> > I installed HDD Health Monitor and had forgotten all about it until it
> > started showing popups every day telling me my drive was going to fail
> > within days. That was about 6 months ago.
> > My solution was to uninstall the program
> Aah, the 'ignorance is bliss' solution: ignore it and it just goes away.
Appears that way. Like I said, it suddenly was supposed to fail within days
and it's been 6 months already.
> Unfortunately that will not stop that counter counting down.
What counter counting down? I never saw any change in my attribute values.
The 'value' readings are still all at 100 except for temp and spin-up
(higher) and CRC error rate (200). I think I rightfully concluded that in
my case Health Monitor was FOS and that nothing was wrong with the drive.
(I did do a fresh backup at the time though, just in case
> > - my drive is fine and passes all SMART tests.
> That's now.
> Who's to tell whether it still does that on 8-14-2004 when
> that counter presumably will have reached the threshold?
The fail date was almost 6 months ago for me - never came to be.
> > The idea that a program can tell you the exact date that
> > a drive will fail seems a little silly to me.
> It is not an exact date, it's the predicted date of when the threshold
> will be reached, that is calculated from the current rate of countdown
> when continued in the same pace.
I wonder what the program would do if the threshold for an attribute is 50
and the reading for it is 100 one day and 90 the next (for one or a few
readings in a row)? Nothing I would hope, or would it immediately tell me
that my drive would fail in 4 days? It *should* IMO just keep track of the
readings for a couple days before getting the user all excited unless the
countdown starts to pick up steam.
MBM5 will sometimes show my 2500 RPM fan as having a low of 0 RPM and a high
of 5000 RPM, and a voltage or two as having a low of 0 volts due to bad
readings. So maybe one or more bad readings can screw up Health Monitor too
and cause false warnings. The program should have enough smarts to throw
out nonsense readings but apparently it doesn't.
What is a 'countdown' of an attribute value counting anyway? For my drives,
Start/stop count, Power cycle count, Power-off retract count and Load/unload
cycle count all have a raw value of 30h for one drive that I updated the
firmware for I guess 48 days ago, and values of 17Fh, 17Eh, 25Ch and 25Ch
for the 2nd drive. Probably all would be 17Eh (382 decimal) for the 2nd
drive if I hadn't had it set it for auto power off after 2 hours of idle
time for a while (I only wanted that for one drive, not both, and
unfortunately you can't do it).
So if there's an extra Load/unload cycle one day for some reason (one not
due to a power cycle), what does that do to the attribute value that's
normally 100? Would it drop to 99, -1 for each extra cycle? Would it stay
at 99 forever if there's no more extra load/unload cycles, or would it be
reset back to 100 after a number of power cycles with no more problems?
Since for Jani's drive the Load/unload value went from 100 to 93, and the
program is saying the drive will fail in 9 months (which I assume means it
will reach the threshold of 50), the time it took to drop from 100 to 93
must be about 6.3 weeks (or about one extra cycle every 6.3 days) if I did
the math right. Is that the way it works? If so, what's to say the drive
can't keep working for many YEARS with an extra load/unload cycle about
every 6 days? It may never fail, or it may fail in a week.
If Jani's problem is showing up in the raw values (he didn't say), unlike me
with my screwy mother board he may have a REAL problem. >> Stay informed about: Load/unload count