Real life experience on SSD life spans

vbimport

#1

Hi all,

A few years back (3-4 years) I gradually upgraded the HDDs in all the PCs in our Dept with SSDs (most Samsung 840, EVOs or just plain, some OCZ Agility ones). I had the possibility of increased disk failures in my mind, so all important data were on a server where each client PC had access to. Everything worked fine and I relaxed after a year or so.

The last 3 months I've had 8 SSDs out of a total of 14 going bad, as if they reached their life span (for the amount of daily work done in the corresponding PCs), I must admit I had higher expectations. Now I am just waiting for the rest to "explode" on me on a daily basis.

Would you consider 3-4 years of a life cycle as normal for these (early) SSDs? Any experience from your end?


#2

Bad luck maybe?
I still have the Kingston V200 running without any issues, and also the Crucial M4 is working perfectly on a friends laptop.
The only SSD that failed on me was an OCZ Vertex2, and i remember that there was a bug on the sandforce controller that there was never fixed.


#3

Well it might depend on the amount of work done on each PC, the ones in the Dept are working 24/7 with only a few restarts per week (working with a lot of CT scan loads, so some amount of data transfer is also involved).


#4

I do not think it is down to just bad luck. I have changed several 3-4 years SSDs that has gone bad. Just as Hemispasm mentions, it is typically graphics workstations and other heavily used equipment in a production environment that simply will not last more than 3-4 years. The brands have been everything from Intel 320 series, Kingston (Hyper-V? 128GB), Samsung (do not recall at all).

Personally, I do change the SSD in my main computers every three years as they are all on 24/7/365. No power saving is activated on any of them… The SSDs then retires as external work-storage for the remainder of their lifespan :flower:


#5

AGILITY and VERTEX2 I can see happening
Firmware issues
Do you know which ones failed
My 2 840 EVO
840 EVO total host writes 107.14 TB
70.82 TB
Both at SMART 100%


#6

I had a Crucial M4 die after only light use. Too small a sample size from me however, as I’ve only owned four SSD’s.

Though the Crucial was under warranty, they replaced it with a refurbished M500, which was a not a great drive in the 240gb size (slow write speed). Can’t say that I’m interested in ever buying another Crucial for my main computer. Kingston is also off the list, but that’s another story.


#7

I didn’t have a great experience with OCZ SSDs in the past - two OCZ Vertex 2 120GB’s and an OCZ Agility 60GB totally failed, but were replaced under warranty. OCZ replaced one Vertex 2 with an OCZ Agility 3 which I still have in use to this day as a second boot drive. I no longer have the other two OCZ’s.

We have a Crucial M4 240GB and a Crucial M500 240GB between two laptops in the house. The M4 went into a semi bricked state once. Based on a quick read online, these can often be recovered by leaving them in an external enclosure powered up without the USB cable connected. I tried this for a few hours, but didn’t work, so tried this again for a longer period (I think 5 hours) and it sprung back to life. That was about a year ago and the SSD still works to this day as if nothing happened.

My laptop has a Crucial M4 for about two years now. Roughly once every few weeks, it stutters where the laptop’s HDD light lights up solid and Windows does not respond for 30 second to a minute. After the wait, Windows springs back to life as if nothing happened.

I had a SanDisk Extreme 240GB in my main desktop for almost 4 years without an issue. I recently sold it off and got a larger SanDisk Ultra II 480GB. Despite this SSD’s long life, my desktop PC was actually the worst affected with SSD failures including the OCZ Agility 60GB, OCZ Vertex 2 120GB and a Crucial Adrenaline 64GB cache SSD.

At my workplace, an OCZ Vertex 2 120GB went into a semi-failure state after about a year of usage, i.e. the BIOS detected it, but the drive would freeze when trying to access it. After a secure erase, it worked fine since until the laptop was replaced about 2 years later.

For anyone that uses an SSD, it is vital to make sure anything important is backed up. In every SSD failure I’ve seen to date where it had to be replaced, it fails like an incandescent light bulb - One moment it’s working and the next thing it’s totally dead. It’s not like traditional magnetic based hard disks which generally start developing bad sectors before completely failing, often giving a clue that it’s time to copy off anything that hasn’t been backed up right now. It’s similar also when deleting files (emptying the recycle bin), they’re permanently gone like using a paper shredder due to the TRIM command.


#8

[QUOTE=Seán;2772594]For anyone that uses an SSD, it is vital to make sure anything important is backed up. In every SSD failure I’ve seen to date where it had to be replaced, it fails like an incandescent light bulb - One moment it’s working and the next thing it’s totally dead. It’s not like traditional magnetic based hard disks which generally start developing bad sectors before completely failing, often giving a clue that it’s time to copy off anything that hasn’t been backed up right now. It’s similar also when deleting files (emptying the recycle bin), they’re permanently gone like using a paper shredder due to the TRIM command.[/QUOTE]
Exactly what will happen, so far a catastrophic situation for the few who still fail to back up :iagree: :clap:


#9

[QUOTE=Seán;2772594]For anyone that uses an SSD, it is vital to make sure anything important is backed up. In every SSD failure I’ve seen to date where it had to be replaced, it fails like an incandescent light bulb - One moment it’s working and the next thing it’s totally dead. It’s not like traditional magnetic based hard disks which generally start developing bad sectors before completely failing, often giving a clue that it’s time to copy off anything that hasn’t been backed up right now. [/QUOTE] A lack of backups can make grown men cry like little girls! :iagree: :stuck_out_tongue:

The only SSD I’ve had a problem with, a Kingston SSDNow V-series 64 GB, developed bad sectors just like a traditional harddrive. I’ve only had a handful of SSD drives, however.


#10

It’s true that most SSDs dont give any indication on when they are about to fail, and there for backup of any important data should always be the number one priority.

But I do believe that most new drives are very reliable, looking only on the warranty that they offer 3years, 5 years and some of them up to 10years, you see that some manufacturers have a lot of confidence in their product.
Probably some of us will have new drives before the end of the warranty of those 10 years, but its nice to have that and it also offers extra value if you want to resell your SSD.

Now if anyone wants to have machine that is on 24/7 and do a lot of read/writes everyday, then its probably a much better idea to look for an enterprise SATA SSD and not a consumer one.
I do understand that not every one wants to pay that extra premium, so some sacrifices will have to be made and probably learn to live with them.


#11

^ Yep, at the core of it, the real problem, the one it is hard to convince apart from servers. In other words, three years it is. Three years is a leap in technology, and we move on :bigsmile: :flower:


#12

I still have my Vertex 2 120gb SSDs running for what seems like ages and no issues.


#13

Crucial M4 256 GB for five years. So far, so good. Boot drive only… I have two 2 Gig WD blacks for data and A/V. 1 gig WD black with Linux installed.


#14

Seems like most of us have had similar experiences then. I am just wondering whether this 3-4 year life span with hard use is “expected” only from earlier SSDs and not i.e. the popular 850 EVOs. Maybe Dee will drop in at some point and offer her valuable experience/advice :slight_smile:


#15

This SSD which is the last OCZ model I had left has finally started giving issues. When I replaced my main desktop PC a year ago, I put this Agility 3 SSD in the former PC, which since has had very light use, e.g. streaming video while walking on a treadmill. It had nothing more than a basic Windows 10 installation, so no data to lose.

When I powered it up this morning, it hung at the blue Windows 10 logo. When I rebooted, it then hung just after finishing the POST. When I tried booting a Windows 10 USB stick, it also could not boot, again hanging at the blue Windows logo. At first I thought it may be a RAM or Graphics card issue, until I tried booting a live Linux USB stick.

One thing I miss back in the MS-DOS / Windows NT days which Linux still has is all the info text that appears during boot. Sure enough, Linux displayed many errors about /dev/sdb during boot and made it to the GUI. Indeed GParted also could not access the drive:

Linux%20dev%20sdb%20error

As with any other hard drive failure, I generally disconnect and reconnect the cables. In this case I also took the SSD out and turned on the computer. This time it booted straight into Windows 10. To be on the safe side and make it easier to reload Windows 10 back on if the drive totally fails, I thought I’ll try making a backup. However, by the time I attached a USB HDD, the OS crashed out and again kept getting stuck after POST.

For curiosity, I decided to disconnect the SSD, tap it a few times and hook it back up. Well, this time the SSD did not show up in the BIOS and then asked to insert a boot disk. I tried tapping the SSD again and it showed up in the BIOS, but hung again after the POST. Finally, I tried tapping the SSD a little harder and the PC fully booted into Windows 10. :grin:

I made another attempt at backing up to an external HDD and it went successful. The PC is still running at the moment, going through Windows updates, although who knows how long for.

Just a pity a never thought of tapping any of my previous SSDs that failed before RMAing them. This possibly means they may have failed from dried solder joints, especially since the SSDs were working one day and dead the following day. Indeed, bad solder joints mainly present an issue after a device has cooled down.


#16

Just over 4 years here using a Samsung EVO 840 as the system drive. I believe I’ve encountered the bug known for this model where the speeds reduce over time for older written data, though haven’t tried the firmware patch.

Still keep daily backups as I know with SSDs they pretty much have a binary ‘it’s working’/‘now it isn’t’ state, unlike HDDs in some scenarios where you can detect warning signs with system behavior or drive itself.


#17

Well I am currently using the Crucial M500 480GB SSD as my main drive on my dell inspiron laptop, the M500 is a five year drive and most it life was spent on my laptops (early on the Toshiba and now on the Dell).
The drive had a very heavy use over the last six months, it stays powered on 24/7 and it handles all the read and writes, from all the downloads and stream that I do.

The last time that I ran CDI was back in 21-07-2017 and back then it had 7618GB written, today that numbers has gone up, not that much.

Most of those writes are done over that last six months.
A quick test with AS SSD gives results that are good, for a drive that is on a laptop, and since its installation it never had a secure erase or a format. Also this install of windows was a clone and not a clean install.

Now on drives that died, well it was an OCZ Vertex 2 64GB, that drive showed no signs that it will go bad, benefits of spinning drives, one day it simply refused to be detected on any system, and that drive back then it had a very light use.

Now lets see how much the Crucial M500 can take.


#18

WHAT?!

How can an SSD die off randomly?

How can one make sure this does not happen?


#19

In my case it just happened, I am sure that most newer drives should give you a warning, plus all major manufacturers have some tool to monitor thr health of the drive.


#20

Warning as in a message (I’d be interested to know which makes do this) or just signs of failure? As I’ve known two devs’ SSD drives to just die spontaneously without warning and seems to be inherent to the tech (the lack of noticeable warning signs that is, not necessarily the early death of the drives themselves).