Sign in with
Sign up | Sign in
Your question

Avoiding the Heartbreak of Notebook Hard Drive Failure

Last response: in Laptops & Notebooks
Share
November 15, 2006 12:22:46 PM

Your notebook's hard disk drive will fail some time and there's nothing you can do about it. Calvin Chu tells you what you, hard disk designers and operating system architects can do to lessen the impact of drive failure.
November 15, 2006 12:59:13 PM

Basically a good common sense article. Make sure to keep those reciepts with your warranty. Backing up data is a no-brainer for any important data. I'll give this 3 stars 8)
November 16, 2006 2:37:55 AM

What about using something like Norton Ghost, to create a complete back up of your Lap top onto your main rig? Although I am not sure how to do it though?
Related resources
November 16, 2006 5:03:46 PM

Unrelated to OP post, but MTBF doesn't mean that your drive will last that long. 1 million hours that's laughable. It is a relative term and is computed by failures over multiple drives, I'm not sure how many but it is a hell of a lot. It is used to compare to other drive's MTBF. I am not sure if I got the math right but 1,000,000/24/365=114.155 years, those drives must really kick ass.
November 16, 2006 7:59:11 PM

That idea about drive-level built in raid is an extremely smart one.

/runs off to patent office

/rushes back in

Because I don't see any real reason that the hard drive could not do the parity calculations and whatnot itself. Basically, it would be a limited RAID card built into every drive and instead of the RAID functionality being applied at the drive level, it would be at the platter level.

Seagate, WD, anybody else out there who has not been bought by the other two----GET ON THIS!

I'll buy a few, I promise. Especially if because of perpendicular recording you don't lose too much capacity going to RAID 1. I don't download nearly enough pRon to buy a 1 TB drive, but I sure would like 300GB that I know has RAID 1 and won't crap on me.

Plus with the utter lack of PCI-e devices out there and the insistence to keep building motherboards that support more PCI-e and less PCI, I can't fit a RAID card in my current rig anyways.
November 16, 2006 8:51:49 PM

The only useful thing I got out of this article was the MTBF & Temperature graph.
November 16, 2006 9:23:36 PM

built in raid-1. well, double edged sword here. why? well, if the drive itself is two drives in one sharing the same case, chances are both platters will probably go through the same heat, the same rough conditions, the same head movements (to keep things cheap and compact, obviously) and, continuing with this concept, they will probably die at the same time.

and theres a couple ways for hard drives to die. but the main reason IMO is heat. my friend has a hdd and its data got corrupted massively because of heat. after cleaning the 1 cm thick amout of dust that was on the drive, it worked again. but for how long, i dotn know.

ive had a couple drives die on me. three maxtors, one seagate and one wd. most due to heat, but one of them, the read head died or something.

the only good way of doing this raid-1 thing would probably to seperate the two drives within a single drive, but that implied having two head mechanisms, two motors, etc as well as possible magnetic interference (i wont go into that because im not all that savvy in that respect) and possibly twice the heat (two motors).

so i dont know about this single drive raid stuff....

besides, wasnt there another simili raid-1 expandsion card reviewed not long ago?

i think the best way would be periodic drive backup to an external drive or dvds. but even those have their flaws. dvds are optical media and external drives are just that: external drives (that can fail as well)

ah well. hope one day we manage to make a drive that never dies. (im thinking what about SSD? from samsung. and there are some 64 gb flash drives too - but then we go into read/write cycles and all that)
November 16, 2006 11:21:36 PM

Reliability is an issue with all hard drives, notebook, desktop and external as well.

Everybody knows that hard drives are now the slowest component of a modern pc, yet hard drive manufacturers have yet to really address the issue of speed.

PMR does boost capacity nicely, but it has not been around long enough to determine how reliable it really is.

And then comes the third issue of reliability itself. Manufacturers may neglect to address this in order to keep replacement unit sales. Sure, replacing one drive in five years is reasonable, but replacing a drive a year for five years in a row is hardly fair for consumers.

Manufacturers NEED to take the lead now and give us drives that are higher capacity, go beyond the 1-2 years lifespan and can work 24/7 on demand. This is especially relevant for notebooks because they always tend to get the slowest/smallest drives, must work with greater heat/shock environments and compromise on many desktop drive features.

Yes, all this may be asking a lot, but R&D/technology has revolutionized practically every other aspect of the PC. Just as the days of the floppy drive have now gone, its time to give the HDD a much needed improvement.
November 17, 2006 12:30:29 AM

well, tbh, ive been getting 3-5 years from my hard drives. atleast those that have died. i still have two quantum fireballs (4 gb and 3 gb) that are still working after x years. (dont know when they were made - dont feel like checking)

but then again, i get my 3-4 years because i do sort of take care of my hard drives. ever since my first samsung died within a year, and figuring out heat can lead to death, ive always shoved a fan or atleast gave the drives some breathing room (you cant breathe in the old 3.2 inch bays, specially if its two slot and theres a floppy on top of the drive)

so now i have two wds being cooled by a 120mm fan. should last me atleast 3 years. if not, off to rma they go.

i am considering seagates though. 5 year warranty. :D 

heat is really an issue in laptops though. so in that environment, i think ssd might be the way to go.
November 17, 2006 6:43:05 PM

Quote:
The only useful thing I got out of this article was the MTBF & Temperature graph.

Do you whine about everything? :lol: 
November 18, 2006 1:40:40 AM

Actually, drives will not necessarily fail simultaneously under identical conditions -- this scenario is exactly that experienced by millions of drives in hundreds of thousands of servers and SANs, and if your theory were correct, I can assure you it would attract a lot of attention. This also holds for IDE drives, which are now used in certain SAN implementations.

The real issue is the fact that if multiple platters are contained within a single hard drive enclosure, a failed platter cannot be replaced to bring the drive-array back to full redundancy. This will only be a real benefit if the drive platter and head components are individually self-contained, allowing for replacement (online or offline -- doesn't matter) and subsequent array rebuild to the new platter.
November 24, 2006 4:54:34 AM

A really nice basic article reminding all users be aware. While a good safe backup is the only way to be sure of hard drive failure protection, the use of SanTools, as stated in the article, may give a early warning of failure. SanTools may give some users too much information and control of low level HD S.M.A.R.T. features, I would recommend users look at the new Active Smart v.2.51.0.2176 (www.ariolic.com) as a really nice alternative to SanTools and less expensive.
February 12, 2009 8:33:59 PM

While I agree that you can't replace a failed platter, it can at the very least provide you with a) data protection -- your data is not lost because its still contained on a secondary, good, platter and b) notification and time to replace the drive without any loss of said data. -- the key is, how will the hard drive notify you that one of the platters has failed, and if you are aware how likely are you to bother replacing the drive before it fails completely? i suppose ultimately S.M.A.R.T. monitoring was supposed to be the answer for all this, but how many people actually use it?

joebrainer said:
Actually, drives will not necessarily fail simultaneously under identical conditions -- this scenario is exactly that experienced by millions of drives in hundreds of thousands of servers and SANs, and if your theory were correct, I can assure you it would attract a lot of attention. This also holds for IDE drives, which are now used in certain SAN implementations.

The real issue is the fact that if multiple platters are contained within a single hard drive enclosure, a failed platter cannot be replaced to bring the drive-array back to full redundancy. This will only be a real benefit if the drive platter and head components are individually self-contained, allowing for replacement (online or offline -- doesn't matter) and subsequent array rebuild to the new platter.

February 12, 2009 9:12:38 PM

3 words for you guys

solid state drives.

The smart media stuff will catch up to the platter media sooner or later. And that will be some cool stuff. Think of a raid 5 stripe of 6 or 8 - 72gb smart cards. That would be some fast stuff hooked straight into the bus. One of the medias fails?? no prob its raid 5, eject it and put a replacement in whiile you are working. Best of all no motors = less noise.
a b D Laptop
February 13, 2009 3:56:17 PM

This thread is 3 years old, Solid State drives were so expensive that they were out of the question.
February 13, 2009 4:07:49 PM

ahh I didnt even check the orig thread date. Someone dredged it up before me though!
!