5 Signs Your Storage Drive is About to Fail

I once had 2 hard disks, same manufacturer and same capacity. One of them was making weird noises... it was clearly dying. So I took it and slammed it into the floor just for fun. Turned out I slammed the wrong disk into the floor. The healthy one. After which it stopped being healthy. Very sad story.

P.S. Yeah, I took out those nice neodymium magnets from both of disks. They didn't need them anymore.
 
Wait .... they missed one. I was told (by my grandfather in a bar) that they pack a special "smoke" inside of each drive so when it starts to come out you know you've got a sure fire problem .......
 
Noise is a very important metric. Even if your HDD isn't making any "weird" noises such as clicks or whirring, in my experience a simple, unusual increase in regular noise levels from normal operation (assuming the data is properly defragged) is a sign that the HDD isn't in good health, the data should be backed up asap and not used for anything important. Decreases in data transfer speeds are another red flag, when that isn't fixed by defragging and changing data cables.

>but but but muh SMART

I was gladly surprised to see this article not mentioning SMART like most articles on this theme do. SMART data isn't trustworthy and no one should rely on it too much. I've seen plenty of drives with perfect SMART health records suffering sudden and unexpected deaths with irrecoverable data loss. I find SMART more useful on SSDs because the wear leveling and reallocated sectors metrics at least can give some idea of how much useful life it has left.
 
I once had 2 hard disks, same manufacturer and same capacity. One of them was making weird noises... it was clearly dying. So I took it and slammed it into the floor just for fun. Turned out I slammed the wrong disk into the floor. The healthy one. After which it stopped being healthy. Very sad story.

P.S. Yeah, I took out those nice neodymium magnets from both of disks. They didn't need them anymore.

Holy crap, that blunder must've been traumatizing.

I always take out the magnets too :)
Also sometimes I keep the platters since they make great mirrors, lol.
 
Two of my (many years old) legacy HDDs suddenly started giving me long pauses every time I tried to access them from the File Explorer. "Stalling" is a sure sign of imminent failure. I replaced the larger of the two b/c losing it would be the bigger disaster, but I know I still need to replace the other.

BTW: I've had twice as many SSD's fail on me as HDD's. The reliability of early SSDs were just awful.
 
Nick,
I am very disappointed that you did not include any mention of SMART data, maintained by drive firmware, as an indicator of potential drive failure. For me, if a hard drive has a count greater than zero in any one of these SMART parameters, the drive is history and needs to be replaced: Reallocated Sectors Count (SMART element 5), Reported Uncorrectable Errors (187), Reallocation Event Count (196), Current Pending Sector Count (197), (Offline) Uncorrectable Sector Count (198). The simple free utility SPECCY provides SMART data for drives inside a system, but not for USB connected drives. The usage of SMART data by SSDs is problematic, as the manufacturers often maintain a minimal amount of data.

If you would like to read up about SMART data yourself, wikipedia is a good place to start.

https://en.wikipedia.org/wiki/S.M.A.R.T.
 
I've got a 120 GB Corsair Force SSD that's been running since 2011 with no problems. I've also found that flash drives are fine as long as they aren't overly used. I've had a 128K flash running for 5 years on the same computer as a backup and it's still good. Both of these are used daily although the use has been light recently.
 
I once had 2 hard disks, same manufacturer and same capacity. One of them was making weird noises... it was clearly dying. So I took it and slammed it into the floor just for fun. Turned out I slammed the wrong disk into the floor. The healthy one. After which it stopped being healthy. Very sad story.

P.S. Yeah, I took out those nice neodymium magnets from both of disks. They didn't need them anymore.
Oh.... I felt this to the depths of my soul.

I had a similar disaster once. My critical backup hard drive and main drive began to fail. So I bought another new drive and began this complicated Chinese fire drill swap to replace the failing drive. When I woke up the next day, it turns out the drive I had formatted was the critical backup and the main drive. Luckily I had other older backups and hard drives all hoarded away and within about a week I was able to basically patch things back together but the most recent settings and save data on the main drive. It was still a traumatic experience.

This is why I never disrespect or break an old hard drive, even one that might be broke. You never know when you might need it.

I've got a 120 GB Corsair Force SSD that's been running since 2011 with no problems. I've also found that flash drives are fine as long as they aren't overly used. I've had a 128K flash running for 5 years on the same computer as a backup and it's still good. Both of these are used daily although the use has been light recently.

Same. I just decommissioned a 120gb Corsair SSD I bought years ago, and as far as I can tell it's still running strong despite heavy daily use. I'm considering replacing one of the slow-*** HDDs in my old laptop with it.
 
Two of my (many years old) legacy HDDs suddenly started giving me long pauses every time I tried to access them from the File Explorer. "Stalling" is a sure sign of imminent failure. I replaced the larger of the two b/c losing it would be the bigger disaster, but I know I still need to replace the other.

BTW: I've had twice as many SSD's fail on me as HDD's. The reliability of early SSDs were just awful.

Stalling is sure a bad sign, however hard drives can also become slow or stall because of weak sectors. A low level format* might fix the problem. Sometimes the drive is indeed bad but I have fixed or at least improved a few slow or stalling hard drives by doing this.

The two tools I've used for this are HD Sentinel Pro ("Reinitialize Disk Surface" option) and HDD Low Level Format Tool. HD Sentinel Pro is payware (there is a freeware version but only the Pro and Enterprise versions have this functionality), and HDD Low Level Format Tool is shareware (the free version is capped at 50 MB/s throughput). Unfortunately I don't know of any freeware or open source tools that are able to do this as well as these programs. These tools will try to revitalize weak sectors, and mark them as bad sectors if they can't. With these tools even if the sectors can properly write and read data without corrupting it, they will be marked as bad sectors if they're too slow.

* Technically it's not really a low level format, which starting with IDE drives, can only be done at the factory. But LLF is what almost everyone calls this.

Latest drive I recovered by doing this was just a few months ago, a 2 TB Seagate Expansion portable HDD that was stalling and had some pretty slow disk areas that would read and write at less than 10 MB/s. I used HD Sentinel Pro with the Reinitialize Disk Surface procedure, and since then it's working like new. I think it's worth a shot.
 
I once had 2 hard disks, same manufacturer and same capacity. One of them was making weird noises... it was clearly dying. So I took it and slammed it into the floor just for fun. Turned out I slammed the wrong disk into the floor. The healthy one. After which it stopped being healthy. Very sad story.

P.S. Yeah, I took out those nice neodymium magnets from both of disks. They didn't need them anymore.
lol
 
If they could somehow put sensors in the NAND flash to measure the strength of program/erase cycles or something like that. I don't know though SMART may already have this kind of thing in place, but it would give a better idea of expecting failure down the line.

That said, I trust so far that SSDs are a lot more reliable than mechanical drives :)
 
Nick,
I am very disappointed that you did not include any mention of SMART data, maintained by drive firmware, as an indicator of potential drive failure. For me, if a hard drive has a count greater than zero in any one of these SMART parameters, the drive is history and needs to be replaced: Reallocated Sectors Count (SMART element 5), Reported Uncorrectable Errors (187), Reallocation Event Count (196), Current Pending Sector Count (197), (Offline) Uncorrectable Sector Count (198). The simple free utility SPECCY provides SMART data for drives inside a system, but not for USB connected drives. The usage of SMART data by SSDs is problematic, as the manufacturers often maintain a minimal amount of data.

If you would like to read up about SMART data yourself, wikipedia is a good place to start.

https://en.wikipedia.org/wiki/S.M.A.R.T.
No reference was made to SMART data, simply because for an individual user, with one or two drives, the information isn't particularly useful. This is because such data, like all statistical information, is only robust when the sample size is large. Cloud service vendors, such as Backblaze, do monitor the data and use it predict when a drive may fail, but this is for nearly 150 thousand drives at a time. Back in 2016, they reported that for 70k drives they were using in their data center, 23% of the drives that failed reported no SMART indicators above zero.

Some storage vendors, such as Crucial/Micron, recommend that one only uses their tools for analysing SMART data because "there is no industry-wide standard to tell you which numbered SMART attribute describes a specific physical property of a drive. The attribute descriptors will vary between SSD and HDD, and even between the various SSD vendors."

So rather than making the article more complex or confusing than it needed to be, I chose to focus on the aspects that any user is more likely to notice, rather than suggest that they monitor data that has a reasonable chance of not giving any indication of an impending failure or provide an incorrect prediction due to data misinterpretation.
 
I have a 1tb hdd, rather old, however it is full of music. Anyone got any brilliant ideas on how to transfer this over to. I dunno, another one?
 
Nick,
I am very disappointed that you did not include any mention of SMART data, maintained by drive firmware, as an indicator of potential drive failure.

S.M.A.R.T. is useful for confirming a failing drive is failing, but absolutely worthless for reporting what seems to be a good drive as actually failing. By the time SMART gets around to telling you that your drive is failing, your drive is usually *REALLY* failing.
 
"* Technically it's not really a low level format, which starting with IDE drives, can only be done at the factory. But LLF is what almost everyone calls this."

A Low level format refers to one which removes not only the data from the files but addressing layer as well. I never did low-level formats on non-SCSI hard drives but for many years I did low-level format on new parallel SCSI installations as well as reloads in order to remove any and all TSRs and other Resistant data from the drive.
Parallel SCSI was equally neat because you could put a SCSI card inThe main board and boot from that, Especially nice if you had a Dead "C" drive in your computer.
That trick also worked well with the PCMCIA SCSI card in an older laptop.
 
I have a 1tb hdd, rather old, however it is full of music. Anyone got any brilliant ideas on how to transfer this over to. I dunno, another one?

You want a recommendation for a program to copy large amounts of data?

Cobian Backup and Teracopy have never failed me.

S.M.A.R.T. is useful for confirming a failing drive is failing, but absolutely worthless for reporting what seems to be a good drive as actually failing. By the time SMART gets around to telling you that your drive is failing, your drive is usually *REALLY* failing.

On occasion SMART even fails to do that. In my life I've dealt with some drives that were heavily malfunctioning and about to be decomissioned (stalling a lot, very slow, corrupting data, making unusual clicks and whirrs), however according to SMART they were still perfectly healthy...

"* Technically it's not really a low level format, which starting with IDE drives, can only be done at the factory. But LLF is what almost everyone calls this."

A Low level format refers to one which removes not only the data from the files but addressing layer as well. I never did low-level formats on non-SCSI hard drives but for many years I did low-level format on new parallel SCSI installations as well as reloads in order to remove any and all TSRs and other Resistant data from the drive.
Parallel SCSI was equally neat because you could put a SCSI card inThe main board and boot from that, Especially nice if you had a Dead "C" drive in your computer.
That trick also worked well with the PCMCIA SCSI card in an older laptop.

Thanks for clarifying what low level format actually is. I assume you're talking about early SCSI-1 drives, did later SCSI standards such as SCSI-3 and Ultra-160 SCSI drives also support low level formatting? I have zero experience with SCSI drives.
 
I once had 2 hard disks, same manufacturer and same capacity. One of them was making weird noises... it was clearly dying. So I took it and slammed it into the floor just for fun. Turned out I slammed the wrong disk into the floor. The healthy one. After which it stopped being healthy. Very sad story.

P.S. Yeah, I took out those nice neodymium magnets from both of disks. They didn't need them anymore.
You sound like a bright, young fellow
 
Thanks for clarifying what low level format actually is. I assume you're talking about early SCSI-1 drives, did later SCSI standards such as SCSI-3 and Ultra-160 SCSI drives also support low level formatting? I have zero experience with SCSI drives.

Parallel SCSI (Small Computer Systems Interface) was used primarily in higher performance PCs both for a constant, higher platter rotation speed than an IDE hard drive and thus faster data transfer operations, as well as superior hard drive longevity.There is also a Serial SCSI which I have not worked with. Adaptec was the primary parallel SCS I vendor; I used Seagate SCSI hard drives exclusively.

SCSI, at the time, also offered several other technical advantages:

1- OS agnostic
2- Much lower spindle latency (4-6 ms) compared to IDE drives (15-20 ms)
3- Ability to easily create hard drive and RAID arrays

Unfortunately, very few consumer-level or gamer-eccentric computer technicians ever got to use SCSI very much before it was taken over by SATA for most applications and so, never learned its advantages. One of my first jobs in IT was as the service manager for a computer store offering both commercial and consumer divisions. Great early lessons!

Later, I used the vastly-faster spin up time of SCSI hard drives to sell a lot of Nuance direct speech recognition (DSR) PCs to my own clients. Quality components, SCSI hard drives (and lots of RAM) finally allowed that amazing software to work as it should. I use it every day on my own systems.
 
Last edited:
This is perhaps the greatest advice in the computer world because there has never been a time when it wasn't relevant. Whether the time period in question has hard drives that are >1TB in size or <10MB, it has always been a critical thing to back up your data.
 
@Wizwill I do appreciate and thank you for the effort on the history lesson, but I was already aware of all the basic technical, historical and market facts about SCSI and SCSI drives. I just never worked with them directly, since by the time I started working with servers in IT, serial attached SCSI (SAS) drives had already become standard and these are very similar to SATA drives. So I'm unaware of finer details such as which SCSI drives and SCSI generations supported low level formatting (which you didn't answer :confused: ).
 
My very first attempt at RAID in 2001 went horribly wrong when after working all night setting up my two BSN 80G WD IDE drives I had gotten off Tigerdirect (I was upgrading my 40G IDE that was failing). Got the BSOD on the first boot-up, man was I pissed! (It ran less than 5 minutes) Wound up cutting my hands on the sharp edges of my very first aftermarket case as I was literally tearing my machine to pieces after all the work I just put into it. Turned out those drives failed from being dropped when they were packaging them up to ship. I got new drives but never did another RAID setup until 2017, was convinced that the article I had followed at setup was itself a "setup" for many years. (LOL)
 
Back