Hard Drive Life Span

[Harry]

Anyone have an idea what the normal life span is for a hard drive before it starts developing bad sectors or needs to be replaced for some other mechanical reason.

Someone told me he thinks it is 5 years at 8 hours per day of use.

[Daboo]

This probably depends on the make. I have had two drives that went after 3 years with an average use of 6 hours per day (one was WD, the other Conner). Others just keep on going, and going, and going, ... I guess the better the company that made it, the better your CHANCES of it running longer.

[kraken]

Hard drives, like the motor in your car, last longer when they are driven continuously at a regulated pace (as against stop/start wear and tear). The initial startup of a cold hard drive is when most wear and tear occurs. When using power management to turn off your hard drives (motor) when not being used actually increases the chances of failure. The only time that you should power down your hard drive is when you have finished your computing session (turning the computer off). Having a setting that turns off your hard drive after 15 minutes of idleness is asking for trouble. When using Power management make sure the "turn off hard disks" is set to "never". Being green and turning off hard disks after 5 minutes to save a few watts of power, will cost you a lot in the long term in broken drives and lost data.

Five years sounds about right, but it is variable dependant on how it is used.


[Edited by kraken on 12-15-2000 at 11:50 AM]

[glc]

Excellent points, kraken. Note that if you allow your computer to go into "standby" it also spins the drives down. The computer itself, in addition to all components in it except maybe for the fans (which are cheap) will last longer if allowed to just run. I have no power management whatsoever - my machine runs 24/7 chewing SETI units unless I am out of town. I do have it hooked to a UPS which is very important. My electric bill is very low except in the summer when the A/C is running - computers don't draw that much juice just sitting there running.

You really can't generalize on hard drive life except for some notoriously cheap brands (such as JTS, which is now out of business). I have seen drives die within 72 hours, and have seen them last 15 years. Generally, IDE drives have a 3 year warranty and SCSI drives have a 5 year warranty. I have yet to have a drive die in my machine, but I generally replace them with larger drives periodically. I have used Maxtor, WD, and Fujitsu all with good results.

Here is a blurb from the Western Digital website on MTBF:

MTBF is Mean Time Between Failure and is typically expressed in hours. The hours are
calculated by dividing the total number of failures into the total number of operating
hours observed. For Western Digital, MTBF represents the average number of hours a
field population of drives will work before a failure occurs.

The Western Digital MTBF specification is 300,000 hours. If a disk drive is operated for
24 hours a day, 365 days a year, it would take an average of 34 years before this disk
drive will fail.

[Harry]

Yeah but isn't there a difference between drive failure and bad permanent clusters because a drive will still run a long time with bad clusters and it seems as though bad clusters will develop long long before you get drive failure.

[glc]

Not necessarily - if I see a drive with *any* bad clusters I would replace it. It's only gonna get worse. Bad clusters are grounds for warranty replacement. Personal experience with customers shows that a drive with bad clusters just can't be trusted.

If you have bad clusters, you can try zeroing out the drive and repartitioning and reformatting it - if the format report shows any bad clusters, it's time to get a RMA if it's still on warranty. If you want to try to keep using the drive, use it as an extra drive just to store files on - it's not trustworthy enough to be your boot drive.

[kraken]

Thanks for the vote of confidence glc, my last few posts in other threads have been a bit iffy (memory lapses).

As far as the power management is concerned, I find that setting the monitor only to go on standby (off) after 15 minutes is more usefull than a screensaver.

The MTBF factor that is used, is based on 1000 drives and the average amount of time before one drive fails. In the case of a MTBF rating of 300,000 hours, this works out to one drive failing every 300 hours (12.5 days) on average. (1000 drives running for 300 hours = 300,000 hours). This rating is averaged for the number of failed drives in an allotted timespan. In the first year (8,760,000 hours) only 30 (29.2) drives out of the 1000 will have failed or 3% if the rate of failure is constant. This failure rate will increase with time.

[Edited by kraken on 12-15-2000 at 01:26 PM]

[kraken]

Bad clusters (sectors) breed like rabbits. Get a couple and before long you have a whole family of them. Bad sectors are the early warning sign of impending doom. A drive may last for a year or more with bad sectors but if some important info is contained within a sector that has gone bad, you may have lost it forever. If the MBR sector goes bad, goodbye drive.

[bdunn]

I got a call from a friend about a dead drive. He has been using the same drive day in and day out for many years it finally died. It was an ST-225 MFM drive.

Thanks to SpinRite I was able to get much of his data off of it and onto a zip disk. That drive is about 15 years old. All that was on it was DOS 3.x NotaBene and a bunch of documents.

[luisr]

There are probably more factors affecting hard drive life span than we can count here.

From my own experience (strictly Western Digital drives so far) they can last a very long time.

I have a 850 MB Caviar series hard drive I purchased back in 1994 which is still running like new and has no bad sectors. Six years and counting (knocking on wood just in case).

I have a second 3.2 GB Caviar series since april 1998.

This system used to be on during the day and off during the night until 1997 when I started leaving it on continuously. And I still keep it on all the time. I also have power management set to turn off the drives after 1 hour of inactivity - which only happens once or twice in a normal day.

Another bit of info might be of help. I live in the tropics and the inside of the case can get quite hot during August and September.

[kraken]

Where I work we have a lot of old drives which have to be tested for placement into refurbished PCs (pcsforkids.org). Most of these drives are 5 years and older. We have a big bin with lots of hard drives with big black Xs on them (dead). About 30% of the drives we test,test good with no bad sectors. About 20% run with a few bad sectors (minor, usually one bad track). The remaining drives (Xs) just don't work or have major problems.

These drives usually come from office environments where they get a good work out.

[courto]

I've got a 4.3gig Samsung harddrive (less than 3 years old)that gave me a primary disk failure last week. I took it out tapped it a few times, then decided to reconnect it and try it out again. Well the thing booted up and has been going OK ever since. Strange but true.
Chris.

[Kent]

I have had several Quantum Fireball drives die recently. They were about 4 years old. They all died the same way: the motor control chip burned up, and I mean that literally. It has a black burned line in the case. I read that it is caused by the bearings getting tight. If you have any of those drives, I recommend you replace them before it is too late. They give very little warning.
I guess that the warranty on drives might tell us how long the manufacturer expects them to last. Seagate gives 5 years, most others give 3, and a few give 1 year.

[Ghostbusters34574]

Don't know what the average life span is, however, just reformatted a ibm 3 gig hard drive from abou 1986 or 7 and it is still going strong.........

[fatboyjim]

Harry posted this thread 4 years ago you realise... hard drive dead yet?

[luisr]

This thread has a longer lifespan than many hard drives.

[Kent]

I did not notice the date. Wow. And still no good answer, because there are too many variables.
A Micronet rep told me that the life of a drive is cut in half for every 10 degrees F in the ambient temperature. That would apply over some range, of course, but fans could make a huge difference.
But I had a Maxtor 80GB drive die after 7 months, with a fan blowing on it.
I listen to old drives. When they get old, the bearings start to go bad, and they start to sound like tiny circular saws. You will know it when you hear it.
The best answer is that your drive will die tomorrow, so back up your files today.
To get off subject, I believe drives cannot operate in a vacuum. The heat generated by fast drives is because of air resistance, but without air, the head will hit the platter. Is that right?

[Craig100]

Originally Posted by Kent To get off subject, I believe drives cannot operate in a vacuum. The heat generated by fast drives is because of air resistance, but without air, the head will hit the platter. Is that right?

Yep, kent, that is correct. There is a small cushion of air that accumulates once the drive begins to spin.

Craig