Question Hard Drive Life Estimate

[AntoPL]

I've had this computer for a many years now and I'm wondering how long until the hard drive is done, here is a picture of the results from crystal disk info.

https://imgur.com/a/PfNQH7T

 

[USAFRet]

I've had this computer for a many years now and I'm wondering how long until the hard drive is done, here is a picture of the results from crystal disk info.

It will work, right up until the moment it dies.

There is no 'prediction' of remaining lifespan.
Keep good backups.

 

[Math Geek]

your guess is as good as anyone's. i've got drives over 10 years old still going strong and i've had drives die in 6 months.

all you can do is make sure you have good backups and if the drive dies, then nothing is lost. i have 2 hdd's in my system with almost as many hours, around 60k, and i have no concerns really.

 

[MWink64]

I'd run a full surface scan with something like HDDScan or Victoria and see if any more bad sectors are found. If not, it could go a while longer. As others noted, there really isn't a set lifespan for hard drives.

 

[Perene]

When the word CAUTION appears there, it's time to ditch any drive, HDD or SSD. Failure is imminent. It may die in the next boot.

 

[MWink64]

When the word CAUTION appears there, it's time to ditch any drive, HDD or SSD. Failure is imminent. It may die in the next boot.

Not necessarily. It's possible for a transient event (like a physical bump, while running) to result in a small number of non-spreading bad sectors. I've seen drives continue to work for years after such an incident. The thing is to make sure that they truly aren't spreading. That's why I suggested the full surface scan. If there's more than a dozen or so bad sectors, there's a good chance they will spread.

 

[LinuxDevice]

One thing Crystal Disk isn't showing is the timing of the reallocated sectors. If they just recently showed up, and occur in rapid succession, then it might be failing now. If they've been there for some time, then maybe they don't mean anything significant. I like @MWink64's description, as a tiny bump could be a game changer. Incidentally, testing in the factory for some ancient mainframe disk drives included a "tap test" while monitoring particle flow through an absolute filter (meaning only particles of a very tiny size would be shown). The particles breaking lose were a big clue as to whether the drive was going to last or not. Modern drives are a lot better, but once particles (such as from bad sectors of physical defect on the disk surface) break loose, they become an abrasive.

 

[Misgar]

I enjoyed the description of a "tap test" to monitor particles released from mainframe drives.

I must dig out an old computer with two 20MB Lapine Titan hard disks dating back to 1979. I also have a full height 5.25in SCSI drive somewhere.

 

[USAFRet]

I treat all storage devices as if they are to die in the next 0.25 secs. Even new out of the box.
And act accordingly.

 

[punkncat]

IMO, since it is showing a caution, back it up NOW and have a replacement on hand.


I have a 4TB drive that absolutely will not work in a PC properly. It makes noises, throws errors, all manner of bad things happen with the host PC slowing down and so forth. I put that same drive inside an "EST" style Linux format NAS and it has absolutely no issue at all there. Runs fine. Being used as the mirror drive in that enclosure. I have a replacement on hand and ready, but the drive has refused to give up yet.

I have a set of 250GB drives that are WAY WAY old backup drives that I used in my first solution for said. They still work fine. They have to be something in the neighborhood of 15-20 yo at this point. I have heard tell that HDD should be replaced at 5 years. I have rarely had any fail at or before that time frame.

 

[MWink64]

One thing Crystal Disk isn't showing is the timing of the reallocated sectors. If they just recently showed up, and occur in rapid succession, then it might be failing now. If they've been there for some time, then maybe they don't mean anything significant. I like @MWink64's description, as a tiny bump could be a game changer. Incidentally, testing in the factory for some ancient mainframe disk drives included a "tap test" while monitoring particle flow through an absolute filter (meaning only particles of a very tiny size would be shown). The particles breaking lose were a big clue as to whether the drive was going to last or not. Modern drives are a lot better, but once particles (such as from bad sectors of physical defect on the disk surface) break loose, they become an abrasive.

This is a very good point and we actually do have a potential clue. The fact that attribute C4 - Reallocation Event Count is also 4 indicates that they were separate events. If they were dealt with simultaneously, there would be only 1 reallocation event.

It may be possible to get an even clearer picture. You can use something like GSmartControl to check the Error Log. This could show at what point (Lifetime hours) the last few errors were detected. Compare that to the Power-On Hours attribute.