current status on HDDs?

[PCBONEZ]

Re: current status on HDDs?

Thinking back my first drive upgrade to that thing was a 1Gb Seagate.
It was on sale at Fry's for $360 I think it was.

[Logistics]

Re: current status on HDDs?

I miss Micropolis.

[Wizard]

Re: current status on HDDs?

Utter garbage!!

I once received pile of 2GB and 4GB version of Micropolis (all HH 3.5"), around 12 in total and they were from novell rack and owner said is hard to put in and pull out of their cases so they get bumped. This what destoryed them all because the platters were not tightly clamped (clamp ring is pressed onto the spindle spool and platter have too much play on the spindle.

The QC is utterly in the dirt. Screws not even torqued at all yes even on the spindle motor to the case!

What a total loss of everything I tried to use. :P (Grrr to the seller who don't understand the problems). GOOD RIDDENCE TO THE MICROPOLIS.

Cheers, Wizard

[PCBONEZ]

Re: current status on HDDs?

OOOh.
I think that was a nerve.. LOL

[gdement]

Re: current status on HDDs?

That's probably right. I remember the characteristic vacu-formed like shape.
I still have a 420MB (I think it is) one of those that works last I checked.

I remember that it definitely was Seagate who bought Conner.
I looked up revised jumper settings for my Conner 40MB on the Seagate site around that time. Don't have that info anymore though, and the settings labeled on the drive don't work with anything modern. (Not that it matters, it's just a museum piece now)

I have Conners in that 40MB and a 250MB size, both work fine. The 250MB ended up having more miles put on it as it remained useful for DOS stuff. It deserves honors for lasting forever, and I kept going back to it long after obsolescence.


Recently I got one of those CompactFlash -> IDE adapters, and hooked up an 8GB CF card to an old Pentium motherboard. I installed MSDOS 5.0 on it and it just works, no fuss at all. I don't think I'll use real hard drives for DOS anymore.

[dood]

Re: current status on HDDs?

Okay, so since the 9th of january, The maxtor drive POH count has increased by 6,000. Something tells me it's not counting hours!

[jpdoe]

Re: current status on HDDs?

@dood
Some firmwares keep track of time differently. And given enough time, some of them will start counting from zero again. What I do when I'm in doubt is download smartmontools (there's a package for win32 too) and run a short selftest.
If it's an IDE hard disk: smartctl -t short /dev/hda
if it's a SATA hard disk: smartctl -t short /dev/sda

After a couple of minutes:

If it's an IDE hard disk: smartctl -A /dev/hda
if it's a SATA hard disk: smartctl -A /dev/sda

at the end of the output, there will be the results of the selftests, and each of them will contain the hour of the hard disk's lifetime it was run on.

[dood]

Re: current status on HDDs?

I've read that SMART will keep track of time in either hours, minutes, or seconds. But for this drive, none of those options seem to jive up. Unless, like you said, it's in minutes and has reset itself back to 0 somewhere along the way.

Now that I've done the math... that has to be what it is. 4 days x 24 hours x 60minutes is 5760 minutes.

[jpdoe]

Re: current status on HDDs?

The thing with SMART is that it was thought to give a simple GO/NO GO status. There's apps that let you look inside the black box, but since that's something that theoretically should be done only with the manufacturer provided software, it's not unusual to have inconsistencies.

Apart from what you've mentioned about different time units, the hard disk firmwares sometimes have bugs in the way they store the parameters. Sometimes for certain parameters, the bits will be in the wrong order, for example. That's why smartctl has a parameter (-F) to allow the user to ask for an output with these bugs corrected. But this parameter uses a database to know which firmware versions need correcting, and may not be up to date.

[Gianni]

Re: current status on HDDs?

current status of HDD
Reading this thread I become curios so I searched for tools to read my SMART status.

I tried 3 tools: smart.exe, HDD Health and Western Digital Data Lifeguard Tools.
The last one doesn't give any information about HDD health even if my system HDD is WD; ok it seems to have some useful tools... but what about the HDD life status?

smart.exe read the parameters and everything seems OK.
HDD Health said my WD HDD is critical

What tool do you use to check the status of your HDD?
Is SMART function useful to keep you away from loosing your stuff?

I hope this post is not too "off topic", if it is I hope a moderator can delete it.

Ciao
Gianni

P.S: it is curios that Datalife can't read firmware revision while the other tools can

[dood]

Re: current status on HDDs?

For ease of interpretation, I use SpeedFan. It has a tab in the program for drive SMART info. There's then a button to get the detailed report. It sends the data to a web program that analyzes it and puts it in an easier-to-read chart.

However, from the looks of it, I would back up your critical data now!

[PCBONEZ]

Re: current status on HDDs?

Recently I got one of those CompactFlash -> IDE adapters, and hooked up an 8GB CF card to an old Pentium motherboard. I installed MSDOS 5.0 on it and it just works, no fuss at all. I don't think I'll use real hard drives for DOS anymore.

Interesting......

I wonder if you can RAID CF cards?
~~ In DOS for that matter.
.

[i4004]

Re: current status on HDDs?

>Is SMART function useful to keep you away from loosing your stuff?

nope

read this
https://cdn.badcaps-static.com/pdfs/...4611a820c2.pdf

"Despite this high correlation, we conclude that models
based on SMART parameters alone are unlikely to be useful
for predicting individual drive failures. Surprisingly, we found
that temperature and activity levels were much"

offcourse, you don't have to read it all.

[jpdoe]

Re: current status on HDDs?

@Gianni

It seems that smart_hd.exe and WD "hdd health" are giving conflicting information. If the WD utility is the most recent, I would believe its output over the 3rd party utility output.
Look at parameter ID 03: "Spin up time":
smart_hd.exe
value: 95
threshold: 21
hdd health:
value: 95
threshold: 200
So for smart_hd this parameter is OK, but for hdd health, it is a sign of error.

The SMART table consists of a list of parameters. Each of them has three values associated.
value: actual value of the parameter.
worst: worst value the parameter took on the past.
threshold: lower value the parameter can take without signaling a failing state.

All these numbers are on the 0-253 range.
The "raw" column gives the actual value of the parameter before being normalized into this range. For some parameters it has a real meaning. For example, the Power On Hours Count gives the time the hard disk was switched ON.

[jpdoe]

Re: current status on HDDs?

Hi, i4004.
It's true that SMART isn't foolproof. But it is an useful utility, especially when mantaining other people's PCs. It seems most people wait until they run their PCs to the ground to call someone for help. There's an option on most BIOSes I've seen to tell the BIOS to check the SMART status of the HDDs on each startup. When an HDD is failing its SMART test, it gives an ominous "HDD failing. Backup your data NOW" message. It's very effective at making people get up their asses and call you.

The most interesting parameters, those I usually monitor, are "Reallocated event count" and "Current pending sector". They give you an idea of the status of the platter's surface. SMART tests will only fail when the HDD runs out of spare sectors to use instead of ruined sectors. But there certainly is a difference between a hdd in which these parameters are zero, and one in which these are in the hundreds. Granted, sometimes these parameters will indicate the HDD is failing when the fault lies somewhere else (PSU). But they give a way to test for hard disk integrity without even opening the case.

Another good one is CRC error count. On some HDDs it increases constantly, without nothing being wrong. But on others it's the perfect tool to know when the data cable is giving trouble.

And finally, the selftests are invaluable. They test the hdd without involving anything else (not even the data cable).

[i4004]

Re: current status on HDDs?

did you have any success with maxtor(esp. older ones..) drives with smart enabled?

google numbers suggest only small numbers of hdds could predict failure via smart, and they have helluva lot of drives to test it on...

for me problem is not reallocation at all, because bad sectors are detectable(they show in event viewer of windows too), so when hdd failes in such gradula way, it's excellent!
what the REAL problem is are drives that fails totally, and without any warning anywhere.
i had 2 such maxtor drives.

[PCBONEZ]

Re: current status on HDDs?

I think you guys are missing something.

- SMART is SMART.
It's not just a counter.......

To explain:
If you get an error (over the threshold) during the BOOT-time self test it adds it to a log and reports a problem to the system on BOOT.
But also if that over the threshold value goes back below the threshold [and stays that way] the log will eventually updated to the new lower highest value.

In essence the log shows the worst case of the last 'so many' BOOT-time self tests.
[I don't know how many.]
The self test counts are stored in the Firmware (where SMART is) in a log file and not on the platters.
And the report screen only shows the worst case of the last 'so many' self tests.
[Firmware can only hold so much data.]

So, lets say you have a drive with a butt-ton of CRC errors.
You write zeros to [wipe] the drive and that fixes them.

The report screen will still be reporting the worst case of the last so many self tests in it's log,, but as the old logs are replaced with new logs the highest value will eventually be corrected to the new lower number of errors.

This is also why a one-time error [false positive] will show during boot a few times and then go away.

This is from a long discussion I had with WD Tech support.

.

[jpdoe]

Re: current status on HDDs?

IIRC, Maxtor had a nice, complete, SMART implementation.
But I don't see much of them, and I have to say you're correct in the fact that SMART is useless when dealing with sudden deaths. I usually chalk them up to cheap PSUs.

The google study is interesting. The most interesting find (besides the SMART uselessness for hdd crash prediction) is the dissasociation between hdd lifetime vs hdd working temperature. I certainly didn't see that one coming... I always thought cool HDDs were happy HDDs.

[i4004]

Re: current status on HDDs?

>I usually chalk them up to cheap PSUs.

enermax 365 with the 2nd. not cheap at all, infact pretty damn expensive and good psu.

bonez, i wouldn't call that smart.
infact, i wouldn't call anything any computer did thus far smart..hehe
sure it's counter, only it's not the simplest counter there is...
comparing few readings within defined time frame(and alarming if it crosses the threshold)..hardly smart...
and hardly usable in most important cases where heads crash and u lose everything in one instance...

so no, not a fortune telling crystal ball at all.

[jpdoe]

Re: current status on HDDs?

@PCBONEZ

Are you referring to the "worst" parameter values? If that's affirmative, it's quite interesting. I thought once the HDD had a failure event, it would fail SMART tests for good. But your observations make sense. It would allow the firmware to deal with false positives caused by external causes (PSU, data cable).

Regarding the zero wipes, what I observed with several firmwares is that in order to cause an increase in "Reallocated event count" and a decrease in "Current pending sector" I had to zero wipe the HDD. It would seem that triggers the reallocation on several firmwares.

@i4004
Nice PSU... I've only seen them in photographs. The close thing to a good PSU I've seen is a Topower a friend of mine bought a couple of years ago. Around 450W and USD100, IIRC. He was extremely dissapointed when he lost a HDD because of it, and I made him open it up. Full of crap caps, most of them already bloated.