Maximum PC

Been meaning to ask this and keep forgetting (didn't see a better forum to post this in tho). Seems that on most MB's (or at least the ones I've seen) that S.M.A.R.T. is disabled by default. Being a monitoring tool for hard drive failure, I would think it should be enabled? Or does it negatively impact on the computer performance?

It can get really annoying.

For example, I've got a Toshiba drive that's waaaaaaay past the SMART threashold for Power On Hours. If I run any SMART monitoring software it'd constantly say my drive was about to fail which gets really annoying.

It does in fact slow down the performance of your system. I don't know by how much, but it is the reason why most motherboards ship with it turned off.

Gotcha... kinda figured that might be the case, but just wanted to be sure. Thanks guys!

It really only useful if your running a server and want some indication to replace it before total HD failure thus causing long downtimes and possible data loss. It's a preventive maintenance feature.

For the average user, the feature is useless. We run hardware until we want the next best thing or until we drive it into the ground.

When it's disabled, boot up is a tiny bit faster because it doesn't have to display it on post.

I'm almost positive its included in the onboard chip on the hd itself so the bios one doesn't need to be turned on. Reason I say this is mine was turned off in bios yet still while I was in windows, smart came up and said a drive was about to fail. It was a 4 yr old WD which I didn't really care about and didn't have any important stuff on. But it died 3 days after. So it worked even with the bios turned off and it was accurate.

If you do not use S.M.A.R.T.-aware utilities or if you do not need that level of real-time reporting, disable HDD S.M.A.R.T. capability for a slight overall performance.

This entire thread... ugh.

steve771: S.M.A.R.T. is disabled by default on some boards because their vendors are morons.. There's no reason not to enable it, except...

Flytrap7: Toshiba hard drives are the worst ever made, so that's unsurprising. With a proper S.M.A.R.T. implementation (including properly normalized raw values) it is not annoying.

GPFontaine: You don't know how much because you can't tell. I don't think it's even possible to notice a performance impact because it's so negligible.

tmartinez2000: It's not useless for regular users. If you don't want to know how many bad sectors have been found or how many are pending, then that's fine, but I don't think you should give the advice that it's useless and everyone should ignore it. I use it daily (I work on 5 to 10 failing hard drives every week). The part about your bootup being faster because of your BIOS not displaying the status -- that's up to your BIOS vendor. Most don't report S.M.A.R.T. on startup except to say that it is enabled on a drive. There's no measurable performance loss from a single ATA packet being sent to S.M.A.R.T.-enabled drives on startup and displaying a line of text on the screen about it.

Some BIOSes don't report S.M.A.R.T. status on startup, or even poll it at all. It shouldn't slow down your boot process as the processor on the hard drive doesn't really have to do much work in order to talk to S.M.A.R.T. and it is independent of your actual CPU, etc.

the_tripp: It is in fact performed by the CPU of the hard drive itself. Your motherboard BIOS can support reading from it or controlling the hard drive's S.M.A.R.T. functionality. Some BIOS vendors allow you to disable reporting but won't force a drive to disable S.M.A.R.T. Unfortunately, some BIOSes force-disable S.M.A.R.T.

daveyd: there is no performance increase from having it off. Everyone should start using S.M.A.R.T. utilities such as HDTune, unless you don't like being warned in advance that your drive is about to fail.

http://wiki.djlizard.net/Data_recovery#smartctl


This is the correct forum to post threads like this one:
http://www.maximumpc.com/forums/viewforum.php?f=6

and you're completely entitled to your opinion, despite the fact that it has nothing to do with what I was talking about. I've got three 30GB 2.5" HDDs here that have been running 24/7 since May 2001 and have yet to fail, so much for 'worst ever made', it's too bad I cant say the same for Fujitsu drives.

As for the values, I've yet to see a program that allows the user to change the threshold value so that the SMART monitoring programs will finally STFU.

I think (if I remember correctly, that is) that the threshold ratio is calculated and provided by the drive itself, so there might be something wrong with the S.M.A.R.T. module in your drive's System Area. If you are actually interested in the gritty details, check out hddguru.com's forums - those guys are hard to the core and can explain in more detail (but in a lot of broken English).

There may have been a firmware change after your model for fixing just such a problem, but hard drive manufacturers never release updates to hard drive firmware to users unless it's life and death of the company (which is another thing that bothers me about the industry). Out of curiosity, what are the exact models? Are they all MK*GAX / MK*GAK?

I don't think I've actually worked with any Toshiba drives under 40GB, so you may have me there.

I can say that 9 out of 10 Toshiba drives that are >= 40GB that have arrived in my shop one way or another were failing, even if they weren't actually brought in for hard drive replacement / crashing problems. I have to test every hard drive before doing anything else (virus scans, etc) so I am quite disappointed in the entire hard drive industry, not just Toshiba.

Almost all of the Toshiba hard drives I've found failing had bearing lubrication problems (one of the two exact problems the "freezer trick" is for), and sometimes during recoveries the head stack would completely collapse onto the platters at 4200+ RPM. It sounds like someone eating a bowl of Grape Nuts in your ear. It's the sound of complete and utter data loss, and it makes me cringe to know that it's all over.

I'm definitely not out to call an entire company trash without a method to my madness (and I mean you no offense directly about the Toshiba stuff), but suffice it to say I am not happy with what I see day-to-day. I hate working on hard drives - it's almost all I do anymore, and I wish they would all stop failing so much. I'd be ridiculously happy if I never saw a failing hard drive again.

I agree with you on the Fujitsu, though - they have to be second worst, if you will still bear with me while I continue to give Toshiba the #1 spot. Fujitsu has the worst S.M.A.R.T. implementation I've ever seen in their modern models, where almost all attributes are un-normalized and none of the major databases (such as smartmontools) are able to normalize them. Normalization sometimes requires a bit of cooperation from the implementor, and you know how hard drive companies are about their stuff - it's all hush hush. But yeah... Fujitsu sucks pretty hard. Hitachi has also not impressed me, despite the rave reviews I keep seeing / hearing about.

I'm a Seagate guy because I know what to expect from them: normalized values, several additional S.M.A.R.T. attributes, lengthy white papers for technical information, and a decent warranty. Now they make "enterprise" drives that have a longer MTBF and I'm all over them.

I understand your vision with s.m.a.r.t. but I've never had the need for using such utility. I still have a 2GB WD hard drive that runs Windows 95, and the Quantum Bigfoot 6.4GB TS for it's backup. Mind you the latter is a 5 1/4" hard disk. Both have lasted much longer than their expected shelf life.

Most hard drive failure result from power problems, or the user disconnecting the drive while it's being formatted, or being written to. If you feel that a particular brand is messed up try changing over to Western Digital, Seagate, or Quantum if they're still around.

Most? Where is everyone getting all of this misinformation?!

I don't think any of my customers have ever opened up the side panel of their computer and removed the power cable to their drive while it was formatting or being written to, unless I completely misunderstand you. Drives can handle sudden power loss, too - they immediately park on the landing zone when power is removed. You think the manufacturers simply didn't plan for such an occurrence and allow the head stack to collapse where it is on the platter while it's spinning at 7200 RPM?

Yes, low-RPM low-density 5 1/2" hard drives are going to last practically forever. Their platters aren't even made out of the same materials as they are today; yours is probably glass or aluminum, whereas today's platters are steel or titanium foil (depending on the manufacturer). I have no idea what you're on about with ridiculously small hard drives and Windows 95.

Platter sector density, high RPM, high heat, and cheap manufacturing causes hard drive failures. Hard drives that last forever are not beneficial to hard drive manufacturers. The same can be said for batteries and light bulbs: the potential is there for extremely long-lasting equipment but no manufacturer wants to drop their bottom line by creating decent hardware. That, and end-users don't want to buy $1,000 hard drives when you can get a huge one right now for $80. Manufacturers would have to increase the quality of their manufacturing equipment and procedures, and there's no way they would want to sacrifice their cheap factories, labor, and high profit margin in order to make a better drive.

batteries, and light bulbs are classified as expendable items. If cheap is what you're looking for, expect the worse. Cheap is often associated with the lack of quality assurance during the manufacturing process, or having little to no tech support at all.

the side panel need not be open. - While your hard drive, or thumb storage device is being formatted, pull the the USB connection out, and report back your findings. Tell me if it's not dead.

Is there a certain dollar amount required before something is no longer expendable? Hard drives are expendable; they last 1-5 years (max) these days, and I'm constantly replacing them left and right - it's like thousands of light bulbs going out. Hard drives could be made better, but they're not. Light bulbs didn't have to be expendable either, but they are.

All of these items are cheap because their manufacturers want them cheap. It's a matter of bottom line.



I'll bite on this anyway because apparently I am a glutton for punishment: Why do you think it would die? Barring a +5v arc of electricity trying to jump the gap while you're unplugging the connector causing it to suffer a huge shock, there is no reason that this should kill an external hard drive. It's possible this could kill the enclosure (by shocking the controller chip) if it's cheap and unprotected from basic arcing, but it should not kill the hard drive inside of the enclosure.

However, look at thumb drives - they're cheap and are now expendable, and they don't have room inside of them for lots of power-related protection. Sure, you could ruin a cheap thumb drive controller chip by removing it while accessing it, but this isn't guaranteed to happen and it certainly isn't the norm.

Also note that S.M.A.R.T. is not accessible via USB.

Yes. It's called SCSI.

They don't come with 5 year warrenties just because of the name




5 volts, hell even 12 volts, is not going to arc across the distance of the two conductors found on either a USB type A connector or a standard molex power connector, there simply isn't enough voltage to cover the distance.

Even classic SCSI is being phased out for SAS (which is still a SCSI command set but over serial and with a better MTBF than IDE and higher voltages, etc). It's all about the drive's internals rather than the connector or language they speak, namely the head stack, platters, CPU, and firmware/System Area.

Most major manufacturers have changed their warranties for IDE drives to 5 years (give or take), though it hasn't increased their reliability one iota.

With that comment I was mainly wondering how come light bulbs can be expendable (when they don't have to be) but hard drives can't be considered expendable (when they are).


I'm not going to argue here because my expertise stops when things reach the electrical level, but what daveyd said about drives dying if you unplug them while they're writing... well, that's just not usually the case. I suppose it could happen, but it's not the first thing I think of when I think of a hard drive's points of failure, most of which rest squarely in the manufacturer's lap. Drives can be made better than they are, and end-user drives are simply pathetic in reliability.

are you that absent minded? Forgive me if you are ten years of age or younger, but where did I mention arc?

Stick to pencil, and paper, since you have 5 to 10 failing hard drives every week.

You say heat is what causes a hard disk to fail, and continue with... "Light bulbs didn't have to be expendable either, but they are."

Besides electricity, how do bulbs produce light?

Okay, where did I say you mentioned arcing? I'm trying to figure out how removing a USB connector while a drive is writing could kill the drive and electrical shock is all I can think of (even though USB is shielded for just such an occurrence). Flytrap7 has basically stated that's wrong, so why don't you tell us what you're talking about? I was talking about internal hard drives up until you said most hard drives are killed by power problems or removing the power while writing. Well if neither are the problem (and removing the power doesn't apply) for all of my customers, how then are their hard drives dying? Do you work for a hard drive manufacturer? You seem to be defending them.


I'm a bench tech - replacing hard drives constantly comes with the territory.


I don't know what you're trying to say here, so I can't really respond. Heat is not the only factor, and Google's research claims that heat has no bearing on a drive's failure rate at all which is probably true except that a few garbage drives (like Maxtor Diamondmax Plus 8 slimline, for instance) which produce an excessive amount of heat and have absurd failure rates. It's mostly head stacks, platters, and the controller boards that cause premature failures. They could be made better and they are not.

Oh this is just fucked up.


Several days after this thread the goddamn heads on my Toshiba MK6021GAS crashed.


BLARG!!!!