SSDs are less reliable than HDDs despite having no moving parts. When will they be more reliable?

>works on my machine

>HDD fails, potentially leaving data unrecoverable
>SSD write fails, marks blocks as read-only

UNRELIABLE

When the stop using flash memory. Really, flash memory is a stupid idea for long-term, reliable data storage considering how it works.

Is thay really so? I've heard ssds just dying completely. A HDD has reserved space in case any sectors are nof working properly.

I bet you can keep files on a flash drive for decades if you aren't constantly using it. I got a flash drive from 2006 still working. It's only 1 GB though and have used it several times a year. I'm curious to know what happens when it does stop working.

I have a 2001 model 32MB(!) UFD. Still works. Alas user, the plural of anecdote is not "data" - for our still-working-after-two-decades sticks, there are billions that are dead and in landfill.

The failure mode of flash tends to be "bullet to the brain": the device simply stops working, and resists any and all attempts to access it - whether a 32MB stick, or your 2TB Optane.

>SSD write fails, marks blocks as read-only
This never happens in home use since the write endurance of any half-decent SSD is enough for decades of use. The problem for both HDDs and SSDs is typically random failure, not wear.

Both SSDs and HDDs can fail unexpectedly and irrecoverably. So back up your important data.
>A HDD has reserved space
A SSD actually has orders of magnitude more reserved space than a HDD because it needs to constantly rotate used sectors to ensure even wear, not just replace bad ones.

More like
>SSD trips over itself once, whole drive locks up forever. Have to send it in and the contracted service company morons will steal all your data and send you back a refurb.

>SSDs are less reliable than HDDs despite having no moving parts.

They aren't unless you're buying /csg/ tier SSDs.

Pretty much. Since SSDs are so ridiculous on cost/GB basis, people are getting wise to it, and only buying cheap/small ones and using them for OS and swapfile.

Controller failure in SSDs is literally no different from controller failure in HDDs. Also if the SSD controller is completely dead, it can be difficult to impossible to steal your data, unlike with a HDD (as no HDDs have internal encryption).

>no HDDs have internal encryption

Attached: 1523310286212.png (1600x1200, 254K)

And then in 5 more years they became even more unreliable than before.

People who can buy ones that do won't be in this thread.

>When will they be more reliable

When you stop using wd shit

>2010
>SSDs are so ridiculous on cost/GB, I'll consider one when it's below $1/GB

>2013
>SSDs are so ridiculous on cost/GB, I'll consider one when it's below $0.5/GB

>2016
>SSDs are so ridiculous on cost/GB, I'll consider one when it's below $0.2/GB

>2019
>SSDs are so ridiculous on cost/GB, I'll consider one when it's below $0.1/GB

never change Jow Forums

Attached: raid0.jpg (2000x1047, 844K)

It's actually (and I'm worried about your intellect that you don't understand this):
>20xx
>SSDs are so ridiculous on cost/GB, I'll consider one when it's below five times what a hard disk costs

More like
>SSD controller dies because cheap electronics/bad PSU/whatever and you now have to pay a professional to recover your data

*tips fedora

Used some 40 gb ssds in raid0. Data didn't fail however it was unreliable as it was giving intermittent failures.
Managed to copy everything from raid 0, even it wasnt need, and retired the disks.
It was w7 and trim didn't work on raid

>HDD controller dies because cheap electronics/bad PSU/whatever and you now have to pay a professional to recover your data

I don't see any difference

t. once proud owner of two dozen IBM AVER drives

>SSDs are less reliable than HDDs despite having no moving parts.
Nah-ah.

I've never understood why you wouldn't boot Windows PE (or something that won't recognise/mount it as a RAID) and TRIM the drives individually.

>I don't see any difference
Fair enough.
Though SSDs were advertised as safe even when they fail.
And when they fail you lose everything.
And unlike HDDs, they don't give you signs before failing

>small ancient SSDs
>in raid fucking zero

Attached: cnpkXC7.jpg (480x320, 28K)

Regular disk2vhd of his boot SSD to one of his hard disk is how my mate handles it.

More like
> [Storage dies] because reasons and now you have to restore your data from the last backup, losing maybe a days worth of stuff

>unlike HDDs, they don't give you signs before failing

Except they rarely do. I've had a couple dozen HDDs fail over the years and only like 1 in 4 had shown SMART warnings, the rest died or started showing read errors with SMART reading A-OK.

I didn't cared that much, it was 8 year ago. Last year I opened one and tried to use as external HDD. Thing was working but showed failures repeatedly despite trim. It was sandforce Corsair first model. I still don't know why it failed so bad, from raid set 2/3 had same hardware failures.

Now I'm using NVM as primary driver and HDD raid as storage. No matter what people say I will keep HDD raid for a long time, it's cheap and fast. I had one raid 0 set on Seagate for 8 years, when I replaced the set there was only minor block failure on one HDD. Sold the disks

If those were 40GB drives (X25-V?), chances are they didn't have TRIM support at all.

>last backup
>a day
This is nu-Jow Forums - backups are boring and time-consuming, so they aren't done by your average desktop ricer.

Attached: 1523519207984.gif (267x219, 717K)

>You have only so many terabytes of writes to them
Say what?

Attached: DiskInfo64_2018-09-01_17-09-35.png (674x839, 77K)

HDDs have only so many terabyte writes to them, too.

Yep, I've RAIDed hard disks for decades too, only ever lost one array - which was no problem, because it was the only RAID1 I ever set up (most were level 0, two were level 5).

>Trusting SMART
Idiot, you have to listen to the noise your HDDs make.
It's like a fucking car. Would you trust your ears or a software to warn you about engine failure?

>sandforce Corsair first model

Early Sandforce controllers had really shitty reliability, they're a big part of why SSDs got a bad rap at first.
Although I still have a working 100GB Vertex 2, a model that had like 30% failure rate in the first year.

For how long data will stay on a SSD if it's completely unplugged for months/years at a time?

>12 times more writes than reads
>power cycle every 7 seconds on average

what

Less than a year before corruption starts.

Yes, many trillions. It'll suffer a head crash from you kicking it around your desk, or you'll go "fuck, I have no use for this tiny 12TB drive" long before that happens though.

I kek every time I read this. Every generation of SSDs has been a "quantum leap in reliability" over the previous, and I remember everybody wetting themselves over SandForce for being so fast and reliable.

I don't think you can get reliable stats on that m8.

Why would you use a SSD for offline storage anyway?

I do lots and lots of db conversions..

Electromigration will see the data "evaporate" over time. Some say as short as a few months. Some say decades. Physics suggests it'll be somewhere between the two.

I've had a few HDD failures. My SSDs haven't even passed the 95% wear mark.

>My SSDs haven't even passed the 95% wear mark
Now ask yourself what is the remaining lifespan of your controller

Bet you a fiver that your HDD failures had absolutely nothing to do with them hitting the can't-really-be-reached-in-a-human-lifetime write limit of spinning rust, though.

>everybody wetting themselves over SandForce for being so fast and reliable.

That never happened though. Fast yes, but I don't remember anyone ever praising SandForce or OCZ for reliability (people who wanted reliability bought X25-Ms)

>2018

>Bargain basements prices for soild-state media

What the fuck is with power cycles though

Sad attempt at retcon, but since this happens every single generation, I'll let it slide.

>assuming that CDI is perfect
>assuming SMART is perfect
>assuming the skynet-like firmware running your average SSD is perfect

That depends if you consider e.g. dirt particles accumulating on the heads as "write limit of spinning rust". And 99% of SSD failures aren't wear-related either.

This. Look at the raw values - mask off the high bits, get c2e - which is what, 3,100ish? Much more realistic number.

Same here. I've had 3 HDD's fail in the last 7 years, but not a single SSD including the first ones I purchased around 2012/13.

I had one SSD in an old desktop that was unplugged for over a year. Powered it on recently and nothing had happened to the data. "Evaporation" is an absolute meme

>And 99% of SSD failures aren't wear-related either
wtf, they said they couldn't break because "no moving parts"

If you have dirt accumulating on your heads, you have a whole lot bigger problems - your pressure-equalising air filter is broken, and you now have a ticking time-bomb pretending to be a hard disk.

>wanking over write limit rather than overall failure rates
No one on Jow Forums or any personal consumer has ever hit the write limits.

>Sad attempt at retcon

Look dude, Vertex 2/Agility 2 had stupid failure rates right in the first few months, the writing was on the wall for anyone except fanboys. There's a reason why SF-12xx was phased out within a year.
SF-22xx was indeed a big step up in reliability, but mostly because the first generation was so bad.

>"Laws of physics" is an absolute meme
This way of thinking will ultimately bite you on the ass, user.

>leaving encryption to your hdd's [[[trusted firmware]]]
Do you happen to shine amidst obscurity?

I think you're wrong about that. The new drives that I have seem as reliable as the old ones, and it's a mixture of Seagate, Toshiba, HGST and WD.

>had stupid failure rates right in the first few months
Yes. Nobody is refuting that. But when they were released, nobody could actually see into the future and learn that half of them would be dead in a year. So they trumpeted about "reliability" - because the previous gen had been such crap. And this happens EVERY generation - "oh, this generation of SSDs are so much more reliable and performant that the last generation!" - that's what I was pointing out.

They "can't break" in the sense of "you can drop a laptop with a SSD on a concrete floor and not lose all your data"

Isn't Neo-Jow Forums filled with /v/ tards on Win 10?
They probably store all their shit in OneDrive because they can't figure out how to stop their OS from doing that.

Will an SSD last longer if you never power down your computer?

Good thing Windows [Rnd(10)] doesn't store anything on OneDrive unless you tell it to. But I digress.

Since everyone's been memed into disabling UAC and Action Center at install, it never even crosses the mind of your average computer user to do backups - despite everyone having the equipment to do so for (at the very least) their My Documents folder.

You sure showed those strawmen.

>My Documents folder
Who the fuck uses that?

I'm sure it can happen, but the timescale is probably several decades, not "a few months" like some idiots believe.

No. SSDs don't really care about power cycles.

Since normies don't actually understand/like folders, they'll click Save, and then they'll press the Save button without changing the folder.

So, like, 90% of Windows users, give or take?

I wouldn't say decadeS, but yeah, guys saying a few months are hosers.

*except Sandforce

I hope samsung micron hynix won't find a dubious reason to not make ssds as cheap as they should be next year

>normies don't actually understand/like folders
I've got a total normie colleague who routinely runs into MAX_PATH limit with her folder names, lol.

>MAX_PATH
How quaint. But seriously, she's an exception, not a rule. Most My Documents' folders are a toxic waste dump of thousands of files.

Only shitskins don't use SSDs. Nonwhites are behind on every conceivable aspect.

>67407522
Quiet, child. The grown-ups are talking. And no (You) for you.

Based and redpilled

pure aryans use tape drives

Well, there goes that thread.

Attached: 1524257197398.gif (278x200, 446K)

>crucial announces new 3D QLC drives BX500
>120GB variant has only 40TBW
the end of SSDs is near my friends, these new drives will only survive 6 years of ~18GB of write per day, and it's going to be even worse in the future

>won't find a dubious reason
tweaktown.com/news/62984/intel-micron-qlc-flash-yields-less-50/index.html
they already did

yeah but I'm naive enough to still have hope

Samsung 960 pro are good, quit buying cheap ssds and take care of them. It will last for a long time, useless fucking op and thread

Attached: E0EDAEC0-0A12-4642-A099-116E45967BF7.jpg (1242x1294, 850K)

Try moving your laptop with a HDD and a SSD and say which one is more reliable with moving parts.

The whole concept of moving a computer full of sensitive parts was retarded from the start

I would absolutely want to move around with a computer full of sensitive parts

What about an onboard computer for a car? Or an onboard computer for your robowaifu?

>onboard computer for a car
>Computer fails
>You're now stranded while your car is mechanically perfectly functional
A great concept

How often are they subjected to a few hundred Gs, though (feels about right for dropping a laptop onto a concrete floor)?

>onboard computer for a *
>Computer fails
>You're now stranded/incommunicado/gasping for breath/dead while your * is mechanically perfectly functional