Yet another SSD "failure out of nowhere"

[Diggs]

One of the reasons I watch the, "What SSD do you prefer?," topics so closely is to see if any clear favorite emerges. So far, none has.

I know stickies aren't big here but it would be nice to have some common place to keep track of who is having problems with what SSD to see if any patterns or trends emerge.

I have my second 1TB SanDisk Ultra on the bench being weird. Transfer rates are a fraction of what they should be (took FABS 14 hours to pull 246GB of customer data, took 28 hours to image). No matter USB or SATA interface the problem persists. SMART and WD Dashboard testing say all is good(???). Came from an old customer machine with 17,000+ hours. He wanted it back for backup on his new Ryzen 9 system. Ummm - no. I'll give him a spinner if it's for backup.

 

[Rigo]

Not looking for any kind of solution here, this is just a rant.

I am coming to hate SSDs with a burning passion because they just crap out whenever they want to without the slightest warning. Every one I've had fail, from the ones that did so straight out of the box (almost) to ones that have lasted several months, and kept showing "entirely healthy" results in their own monitoring software, just up and died without any warning. Even if you have backups, that's just incredibly frustrating.

I never had that issue with HDDs, not that they don't fail, but almost always I had warning sign after warning sign that something was going amiss long before they actually died in a way that would require data recovery in order to get anything off them.

In this case, it's on one of my own machines, and my last backup is several weeks old, and I've been doing some work in MS-Office that I really wish I had saved somewhere other than the SSD, but I hadn't.

Nothing like having your machine freeze up, and when you have to do the hard power button shutdown and fire it up again, getting the dreaded "No OS found" message during boot!

Back blaze I think they are called that do data storage and release analyses/stats of drive failure, there was an article about the preliminary SSD failure stats and it doesn't look pretty so far. I didn't keep a link to the analysis but it appears failure rates are similar to HDDs if not worryingly worse.

 

[alexsmith2709]

Oddly enough i had my first SSD issue last week. It was on this computer and was a TeamGroup GX1 SSD. It didnt fail completely but the write speeds slowed down to about 30MB/s. I tested using the manufacturer tool and CrystalDiskMark with the same results.

One of the downsides of SSD's is generally what you described, they just go without any warning.

 

[britechguy]

One of the downsides of SSD's is generally what you described, they just go without any warning.

Yup. And they're ungodly expensive for data recovery and, from what I've been hearing so far, have a lower likelihood of success with it.

The need to do full system image backups, as well as separate user data backups, as often as is indicated by circumstances is even more important now, in the age of the SSD, than it has been in the past. And it was plenty important in the past.

 

[britechguy]

Back blaze

It is Backblaze. But even they're saying that the data that they have is nowhere near to robust enough to make any real head to head comparisons between HDDs and SSDs as they age.

But, to me, "failure without warning" is a far worse thing than "failure after signs and symptoms." And, so far, all of my SSD failure experiences have been "they quit working like someone flipped an off switch." None were old (the oldest being this 5 month old Adata SU 720). I've actually never had a single HDD "die very young" and very few die, period. It's not zero, but it is a small minority of my service calls that end up being about a dead HDD.

But, I've also not had many service calls involving a dead SSD, either. Those have happened to me, on my own machines. It's not that they die, it's that they've died early and with zero warning.

 

[Sky-Knight]

@sapphirescales Yeah, I haven't bought from them in ages, WDs are cheaper, almost as fast, and the RMA process is better. Samsung is only picked up here when the WDs aren't available and I'm in a time crunch.

Honestly, I don't much care about what Samsung does to themselves, as long as they honor their warranties. Which remain gold standard.

The rest is just internal retooling to deal with shortages and streamline operations. All of the electronics manufacturers are going through this process at this time. And yes, they should be announcing new part numbers when the spec changes. Change is fine, not using a new number to indicate that change is problematic.

 

[Diggs]

deal with shortages

We should be happy (and grateful?) we can get almost any SSD we want in a day or two. Prices bumped for a bit but other than Samsung they dropped right back down. I thought we'd see more issues with the SSD supply.

 

[britechguy]

We should be happy and grateful that we can get almost anything we routinely want in a day or two.

It will probably be several years, at least, until we go back to the "I wanted it yesterday, and I should have it today," delivery that became so common, and trained people not to plan (at all) but to think that everything should be available on demand.

Those of us "of a certain age" who never accepted that notion are a lot less stressed by waiting a short while to receive something we need, and we anticipate having to wait.

 

[NviGate Systems]

I recall in the early days of SMART and hard disks many drives failed without a SMART warning or when SMART was tripped it was game over.

The issue with SSD's is that they are based on chips. I'm sure they are capable of monitoring the chips but essentially, if the controller of the SSD dies, it's game over. I suspect in the case of random failures where suddenly it dies it's the controller that dies.

If the controller is the one that handles SMART, well, that's not exactly "Smart".

 

[Sky-Knight]

I've never had SMART successfully predict a failing drive. The only time I've seen SMART trip was false positives, I swapped the drives just in case of course but used them elsewhere for years before they quit.

 

[nlinecomputers]

I am coming to hate SSDs with a burning passion because they just crap out whenever they want to without the slightest warning. Every one I've had fail, from the ones that did so straight out of the box (almost) to ones that have lasted several months, and kept showing "entirely healthy" results in their own monitoring software, just up and died without any warning. Even if you have backups, that's just incredibly frustrating.

With respect, it’s unrealistic to expect less. Flash memory is volatile and by nature it is much more likely to have a total failure. That’s simply unavoidable as it is the nature of the technology. Trading the higher risk of catastrophic failure vs the speed benefits is the point.

 

[Sky-Knight]

With respect, it’s unrealistic to expect less. Flash memory is volatile and by nature it is much more likely to have a total failure. That’s simply unavoidable as it is the nature of the technology. Trading the higher risk of catastrophic failure vs the speed benefits is the point.

My experience indicates SSDs have a higher risk of catastrophic failure, for a lower total chance of failure. When they go... they go hard. But they just don't go all that often.

I've got SSDs out there that are working on their first decade still going strong. In numbers that exceed every HDD I've ever sold.

 

[nlinecomputers]

I use the analogy of light bulbs. They tend to work or they pop and go out. While a hard drive is a candle. They burn down and get weaker until they fade out.

 

[GTP]

I use the analogy of light bulbs. They tend to work or they pop and go out. While a hard drive is a candle. They burn down and get weaker until they fade out.

This is perfect and an apt description of exactly what happens. May I use this analogy with my clients?

 

[fincoder]

When they go... they go hard. But they just don't go all that often.

Exactly. In my experience the sudden failure of SSDs is about as common as sudden failure of HDDs, i.e. very rare.

I don't know why anyone would touch ADATA after so many reported failures over the years. And generic brands like Mushkin are a big unknown. I've used the almost-generic-brand Team SSDs and had a few failures, stopped using them now. Earlier on I used quite a few Intel and almost every one has failed, the worst brand for SSDs in my experience.

I've had very few or zero failures of Samsung, Kingston, WD, Sandisk (any well-known brand apart from Intel). The only reason to avoid the well-known brands is availability, because prices are quite low.

 

[Sky-Knight]

I use the analogy of light bulbs. They tend to work or they pop and go out. While a hard drive is a candle. They burn down and get weaker until they fade out.

You know I was going to take issue with this... but the analog only doesn't work if you're thinking about incandescent light bulbs. It's a bit of a stretch with the CFLs because those too die simply due to age. But LED bulbs? They burn with use... very slowly and eventually simply stop. Yet, while all that's true, they can also suffer from catastrophic controller failure and simply stop. They are direct analogs, and very easily understood.

So yeah, I hope you don't mind but I'm going to have to steal that! The only real problem is my clients haven't all gone LED yet. But soon enough it'll be a useful explanation.

 

[Rigo]

I don't know why anyone would touch ADATA

I second that.
RAM/USB sticks/SSDs I've seen a lot of failures from the brand.
Patriot is another basket case

 

[britechguy]

FYI, Adata seems to be "on the stick" as far as a replacement for a failed drive goes. I sent the dead drive from Virginia to California on Friday not using priority mail. I have just received an email message that a package is on its way to me from Adata. Let's see how "round two" goes.

I am so peeved with myself that I missed one of my regularly scheduled backups by just a day or two because other things in life got in the way. I still have my backup from the end of September, though, so that's going to be restored on to the new drive promptly once it arrives and gets popped in.

 

[DRPCNZ]

I recall in the early days of SMART and hard disks many drives failed without a SMART warning or when SMART was tripped it was game over.
If the controller is the one that handles SMART, well, that's not exactly "Smart".

5.25 inch Bigfoot drives SPRING to memory

 

[Diggs]

5.25 inch Bigfoot drives SPRING to memory

Haaa! I remember those. Noisy too if I remember right.