How to know when to replace it.
This is a pretty common-sense question.
You would think there would be a common-sense answer.
There isn’t one.
Let me explain why that is, and what you need to pay attention to instead of age.
Become a Patron of Ask Leo! and go ad-free!
Hard drive age isn't important
Hard drive failures are unpredictable. Some may show warnings, but many will not. Failures can be sudden, complete, and permanent. The only protection is a backup strategy that protects whatever you keep on your hard drives.
Hard drive failures are unpredictable
The root of the problem is that drive failures are almost impossible to predict.
Many hard drives seem to last forever. They often outlast the useable life of the computer they’re in.
Some hard drives die quickly. Manufacturers try to identify those by running them for some amount of time before they sell them; the early failures never see the light of day.
Some hard drives give you warning quite literally, showing pop-up warnings that something is in the process of going wrong and that the drive should be replaced soon.
Some hard drives just get slower or have other performance-related issues. They work, kinda, but get slower and slower over time.1 The most common cause is bad sectors on the drive. This can sometimes be alleviated, at least for a time, by running CHKDSK /R on the drive.
Failures can be sudden
Most importantly, and I’d say even more commonly, hard drives just die with no warning at all.
One moment, all is well. The next moment, it’s not.
If you’re lucky, you can salvage the drive. With a little more luck, you might salvage some or all of the data.
More often, though, the drive is toast, never to be heard from again. (Or perhaps never to be trusted again, if you manage to salvage anything.)
You know where I’m going
This is why backing up is so critical.
Your computer is under constant threat of catastrophic failure. Hard disks are just one example — the failure could be anywhere. Be it hardware, software, or malware, things can suddenly go poof.
I’m not saying that catastrophic failure is likely; I’m saying it’s possible. It happens. And it can happen without warning.
Thus, backing up so that a sudden failure isn’t catastrophic is critical to using your computer with confidence.
A more common “failure”
Honestly, the most common “failure” among hard drives these days isn’t hardware, it’s software.
Or rather, capacity.
What was once a hard disk with large capacity becomes “smaller” as Windows, your programs, your data, and other things expand. Your hard disk hasn’t actually gotten smaller, it just seems like it because everything you store on it has become larger. Eventually, you run out of room.
Short of replacing the entire computer (usually completely unnecessary), the pragmatic solution at that point is to replace the disk with something larger.
Has it failed? Not really. You might move it to a different, older computer, or throw it in an external drive enclosure, or repurpose it in some other way. But it’s “failed” in that it no longer meets your growing needs.
Do this
Realize that your hard disk can fail suddenly and without warning at any time. The probability is low, but definitely not zero.
I’ve said it before, and I’ll keep saying it because it’s critically important.
The only way to protect yourself is to make sure you’re backing up completely and regularly.
It’s happened to me, and it could happen to you.
You can see me talk more about backups and other topics by subscribing to Confident Computing! Less frustration and more confidence, solutions, answers, and tips in your inbox every week.
Podcast audio
Related Video
Footnotes & References
1: Defragmentation used to also cause this, but since version 7, Windows has been defragmenting its hard disks automatically, once a week by default.
I saved the two 1TB internal hard drives from my last desktop computer. I installed them in my current desktop tower that came with a 1TB SSD and now this computer is 5 years old and used just for personal use. The two extra drives are being used for storage like you would use an external hard drive and each is about 75% full. I know they are mechanical and are spinning whenever I have my computer on which is about 16 hours a day. This has been handy since the music, data, and photos are easily accessed if I need them. Since they aren’t being accessed all of the time is there less of a chance of them failing?
I hope you have them backed up. Data in only one place is at high risk of failure.
I wonder if you have looked at SpinRite – https://www.grc.com/sr/spinrite.htm – or if you have any opinion on that?
Leo has written several articles in SpinRite. To list a few:
https://askleo.com/spinrite_repair_hard_disk_failures_and_recover_your_data/
https://askleo.com/how_do_i_fix_errors_on_my_hard_disk/
https://askleo.com/chkdsk_runs_on_every_boot_but_fails_and_my_system_is_unstable_what_can_i_do/
The search bar on every Ask Leo! page can help answer questions like that.
I used to recommend it long ago. To the best of my knowledge it’s not been updated in a very long time. I recall it had issues if BIOS support wasn’t correct, but I could be wrong.
Nothing is perfect, even backups. However, careful monitoring of HDD can help. Here at the office we have some computers which are up and running 24/7. And there is enough activity that the disk drives are pretty much running all that time. Other computers are used maybe a few hours a week, or even a month. The rest of the time they are powered off.
So we monitor the SMART, and run a test every so often. What typically happens is that we find that our large expensive HDD in our always-on, critical computers eventually start to wear out. They still work. They are still reliable. But they have enough wear on the bearings, and enough bad sectors, that they will NOT LAST another year or so in that environment. So they get replaced with newer bigger HDDs. But then the pulled drives get thoroughly scrubbed and retested. If good, then they go into the seldom used machines, where they will likely last for many years to come (with limited use).
With regular backups for security, this scheme has served us well and seems a good compromise between reliability and economy.
Leo – under ‘Footnotes & References’, I think you meant fragmentation and not defragmentation.
‘..since version 7, Windows has been defragmenting its hard disks automatically, once a week by default.’
Are you suggesting that spinning discs don’t need any defragmentation, because Windows itself does appear to make quite a poor job of defragmentation ?
Not sure how you’re defining “poor”, but indeed: I don’t believe fragmentation is something Windows 7 and later users need to spend time on.
I think by poor, he means by “Defragmentation used to also cause this” Windows used to do a poor job of defragmentation.
It should probably read:
“1: ̶D̶e̶f̶ Fragmentation used to also cause this, but since version 7, Windows has been defragmenting its hard disks automatically, once a week by default.”
Was surprised the article contained no mention of S.M.A.R.T. testing. I use CrystalDiskInfo now and then. Recently tested a USB drive that was transferring files rather slowly and the test confirmed issues in “Reallocated Sectors Count” and “Current Pending Sector Count.” Is this not at least a fair method of confirming a potential problem?
It’s unreliable. That’s the problem. Not all disk manufacturers implement it thoroughly. My point is that all these different things can help when they work, but there’s no guarantee that they’re working. So … back up.
The quickest easiest way to tell if a hard drive is beginning to fail is to install the Acronis Drive Monitor. It works with the SMART function for early detection. Note that it does not support all hard drives but works with the vast majority. I have it installed on all my machines and have used it as a quick test on others.
Surely this does not apply to Solid State drives, right? And, yes, hard drives are still out there in force especially for backup. So thank you for pointing me to SpinRite years ago – one of many great ideas here on Ask Leo – for which Many Thanks! Any suggestions as to monitoring potential fail on a SSD as a fail in only one virtual sector can trash the whole drive due the sequential read-write design, from what I understand? Many thanks and Happy Holidays 2022!
Wed.,11-23-2022 12:55
Leo –
What is a dencent backup service?
Barry Karas
I recommend using Macrium Reflect or Easus Todo, and backing up to an external hard disk.
I just check the SMART info from time to time. maybe I am semi-lucky but I have been on computers since 1995 and I have only had two hard drives fail on me, which is a 40GB IBM and a 80GB Maxtor (but the IBM one was a ‘Deskstar’ but was referred to as the ‘Deathstar’ since it was known to have a good chance of failing). so far I have not had any other hard drives fail and out of my 1TB and larger drives, of which I still use (I even use my hard drives smaller than 1TB for data backup to), my two oldest, a 1TB and 2TB Samsung, they still work and I want to say these are between 10-14 years old (likely between 2008-2012). they had plenty of hours on them as they used to be in my main PC running for years (and I generally leave my main computer running all of the time) but I currently use them for external backup (like connected to a USB3 docking station as I just slide them into place and power it up). the only internal hard drives I have running basically all of the time in my main PC is a 4TB Seagate/5TB HGST/6TB HGST, along with my boot drive which is a 250GB SSD.
jamesmknox said, “Nothing is perfect, even backups.”
Technically you are right, but with certain combinations of data backup the chances of data loss is low if not VERY low.
like for example having say a couple of backups on two different hard drives along with say two different types (specifically Verbatim and TY(Taiyo Yuden)) of DVD recordable media. the odds of both of the hard drives and media failing at the same time is very slim. like I would expect say a house fire to take it out (assuming all of that is stored in the same location) before the data would actually be unreadable from random failure of the hard drives/optical media themselves.
but I think as a general rule for basic data backup when factoring in reasonable level of protection against data loss vs convenience, I figure a bare minimum standard here is two copies of ones data on two different hard drives (like one copy one hard drive A and one copy on hard drive B). preferably one of the hard drives being external and only connected occasionally for data backup to further lower risk. with this, your chances of data loss should be on the lower side. probably low enough to where you should be fairly safe in general. but for more very high importance I would definitely have AT LEAST one more copy besides this usual two copies on two different hard drives standard, which boils down to optical media like DVD. because I think optical media is still the best alternative for long term storage at a reasonable price outside of hard drives.
A good backup routine is close to perfect
3 copies
2 different formats
1 copy kept offsite
How Do I Back Up My Computer?
My offsite backups are keeping all of my personal files on OneDrive (Dropbox is just as good) and occasionally taking one of my backups to a different location (I keep an encrypted copy at work)
@ Mark Jacobs ; one off-site physical copy is nice to protect against natural disasters and the like. but I am just betting on no natural disaster types of things will happen. because if that happened I would be basically screwed. but short of that type of stuff, the HDD and optical media combo (especially with the total of four copies (2 on two different HDD, 2 on two on different optical media)) should be near rock solid.
but I guess online storage can be additional protection to. but if I did this, depending on what kind of data it is, I would at least encrypt it with some sort of encryption before uploading it online. like depending on how sensitive the data is, if it’s something you would rather people not see but it’s not the end of the world if they do, then a simple program like 7-zip (.7z) can do some basic encryption that should be secure enough as I would assume that would deter most people (assuming you got a half way decent password/passphrase on it). but for more general encryption stuff I would probably opt for VeraCrypt.
but as you already know… many people slack off on general backup and have to learn the hard way, which is why I figure a minimal approach is a good balance of protection and ease-of-use since it’s decent protection but does not take too much of ones time, which pretty much boils down to a bare minimum of two different copies on two different devices (a copy on each device), of which hard drives are probably a safer choice than say SD flash storage since with flash, while it might be okay, if it does die I suspect it’s much more likely to be sudden where as with a hard drive you got a reasonable chance you will see warning signs before it just outright dies.
so while I realize many people, probably most people, don’t really use optical media anymore I would say it’s still one of the most reliable choices for data storage long term (say at least 10-20+ years minimum) as I got Verbatim etc media that’s 10-15 years old and still scans well with the disc quality check with programs like KProbe which shows a drives PI/PIF errors on a burned DVD for example and the discs basically scan as well or nearly as well since I first burned them etc.
As I said, I occasionally switch out my off-site drive. Sometimes it’s as long as a month between switching it out. But that’s probably good enough as it has everything from a month before in addition to all the data files I have in OneDrive. If I lost everything in my home, I could restore almost everything except for any programs I may have installed during the lapse.