Slashdot Mirror
Why I'm Usually Unnerved When Modern SSDs Die on Us (utoronto.ca)
Chris Siebenmann, a Unix Systems Administrator at University of Toronto, writes about the inability to figure out the bottleneck when an SSD dies: What unnerves me about these sorts of abrupt SSD failures is how inscrutable they are and how I can't construct a story in my head of what went wrong. With spinning HDs, drives might die abruptly but you could at least construct narratives about what could have happened to do that; perhaps the spindle motor drive seized or the drive had some other gross mechanical failure that brought everything to a crashing halt (perhaps literally). SSDs are both solid state and opaque, so I'm left with no story for what went wrong, especially when a drive is young and isn't supposed to have come anywhere near wearing out its flash cells (as this SSD was).
(When a HD died early, you could also imagine undetected manufacturing flaws that finally gave way. With SSDs, at least in theory that shouldn't happen, so early death feels especially alarming. Probably there are potential undetected manufacturing flaws in the flash cells and so on, though.) When I have no story, my thoughts turn to unnerving possibilities, like that the drive was lying to us about how healthy it was in SMART data and that it was actually running through spare flash capacity and then just ran out, or that it had a firmware flaw that we triggered that bricked it in some way.
(When a HD died early, you could also imagine undetected manufacturing flaws that finally gave way. With SSDs, at least in theory that shouldn't happen, so early death feels especially alarming. Probably there are potential undetected manufacturing flaws in the flash cells and so on, though.) When I have no story, my thoughts turn to unnerving possibilities, like that the drive was lying to us about how healthy it was in SMART data and that it was actually running through spare flash capacity and then just ran out, or that it had a firmware flaw that we triggered that bricked it in some way.
Seriously, you do not. You may know the end-result sometimes (head-crash), but the root-cause is usually not clear.
So get over it. It is a new black-box replacing an older black-box.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
Happens
Hey Chris from Department of Computer Science has a problem. Let's hear about it, Chris.
Uh, if a disk dies in 2 months you need to get a replacement, not a repair.
*shrug* ?
I mean, manufacturing defects, environment, and just old plain bad luck? SSDs have come a long way, but if I have anything of importance, I'm RAID'ing it and backing up. I feel anyone with an understanding of technology knows the importance of this.
Waterboarding?
I've had two SSDs die utterly. It wasn't because there was a failure of any part of the actual storage pathways: it was irreparable failure of the embedded controller circuits. The Flash itself was still fine and safely storing all my data, but there was no means to access it. At least with a platter drive if the PCB fails, you can unscrew and detach it and replace it with a matching PCB from another drive; no way to do that with an SSD. Early on when manufacturers were spending all their time hyping the comparative robustness of the Flash medium, they conveniently forgot to mention how fragile and not-so-robust the embedded third-party controller circuits could be.
Infant failures are common in electronics ( https://www.weibull.com/hotwir... ) From a simple standpoint, imagine a poorly soldered junction on the PCB - soldered well enough to pass QC and work initially, but after a couple of heating cycles the solder joint fractures. The same kinds of problems occur inside chips - wire bonds between the package and die may be defective but initially conductive, and fracture due to thermal cycling.
Similar problems can occur on the die. The gate oxide for a particular transistor might be too thin due to process issues. If it's way too thin, it'll fail immediately and the die will get sorted out at test. If it's just a bit thicker, it might pass all production tests but fail after an hour or two of operation, or 100 power cycles. If it's just a bit thicker (where it should be), it might last for 20 years and a million power cycles.
Everyone in the semiconductor industry would love to figure out how to eliminate these early failures. No one has found a way to do it.
And the worms ate into his brain.
The spinning parts of an hdd are not the only parts that can go bad. Just as the NAND flash memory are not the only parts of an ssd that can go bad. There are other components: controllers for the computer interface and the NAND chips, and the power to everything. One bad electronic component can take down either. One dead capacitor can stop a whole motherboard from running.
In my experience with HDDs you'll usually get some warning that your drive has issues before it completely calls it quits. Whether it's bad sectors turning up or noises from the drive itself. If you pay attention to that (and you're a little lucky), you can manage to salvage most of the drive's contents before it dies completely.
With an SSD one minute it's working completely fine and the next it's completely gone. While most of the data itself is probably still perfectly intact on the flash memory, getting at it is completely impossible (afaik) without going to a professional recovery service.
With a spinning disk, you'll usually get an indication of a problem with a plethora of S.M.A.R.T errors.
It's been my experience that when an SSD dies... you just suddenly appear to have an empty drive cage. It's a really ugly binary failure.
I've taken to building my boxes with mirrored SSD's combined with taking and validating my backups.
Yes Francis, the world has gone crazy.
Electronics wear out slowly. In fact most will long exceed their usefulness before they die.
Mor often electronics will die early due to manufacturing defects. It's why if your device lasts the first month it will probably keep working until you upgrade it. SSD's are a different beast though. thus they have excess capacity to handle wear leveling. Still a young drive that dies is usually, again, a sign of a manufacturing defect.
It's bad firmware. Some of the drives can supposedly be resuscitated by the factory or people who have reversed the private ATA commands.
I mean, at a minimum unless it's a PHY failure (and there's no reason to suspect those) the firmware could at least report missing storage (I've actually seen a 0MB drive failure once or twice) but their usual failure mode is to halt and catch fire, as the author notes as their usual behavior.
With the recent reports about the inexcusable security problems on Samsung and Crucial drives this is starting to feel like the old BIOS problems with Taiwanese mobo companies outsourcing to the lowest bidder and shipping bug-laden BIOS with reckless abandon. It's OK, all the world's servers only depend on this technology.
To be fair, I have batch of 20GB Intel SLC SSD's that have never done this, but those are notable exceptions. At this point only low-end laptops like Chromebooks don't get at least a mirror drive here.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
I'm a retired IT guy and there's no kind of something that didn't fucking break. I'm not a goddam engineer. My job was to locate the problem at a black-box level and get the shit running again. Contemplating the "why" of a hardware failure is wheel-spinning instead of pulling the stuff out of the ditch.
For new purchases under warranty, I exchanged them and sent the dead one back to the vendor. Let them hook it up and do diagnostics over a cup of coffee.
I had work to do.
It little behooves the best of us to comment on the rest of us.
Doesn't know how SSD's work.
No offense to CS majors, but this EE major tends to understand "How a computer works" at a lower level than most of you programmer types. While not universally true, in my experience a Computer Science major generally get's outside their comfort zone with hardware once you get past "Plug it in and turn it on." I don't blame them, there is a lot of stuff happening at lower levels than a CS major needs to know to do their job.
That some CS major is concerned about how SSD's fail because he doesn't understand their failure modes is fine. We tend to fear what we don't understand and let's face it, there is a LOT of stuff going on inside a computer that high level users simply don't need to know. Heck, even I don't need to know some of that stuff and I've designed computing systems in the past. Fear not, if it works, it works, if it doesn't you just replace it anyway.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Yes, you can listen for mechanical issues, yes you can (sometimes) read bad block and other SMART data. But, ultimately, without millions in equipment and skills, you just do not know. It is a cheap data storage brick. Choose one appropriate for your capacity and I/O needs, have a good backup plan in place, and quit whining.
Silence is a state of mime.
..and the more complex a machine is, the more that can go wrong with it.
The controller PCB on a brand-new modern HDD can fail, rendering the entire device useless; any piece of silicon on a modern SSD can fail also, rendering the entire device useless. The only difference here is that with a HDD, if you happen to have another working drive of the exact same model and revision level, you could theoretically swap the controller PCB and be able to access the data on the platters again (I've done this). With an SSD it's all one PCB and short of actually diagnosing the failure and replacing failed component(s), the chances of accessing the contents of the flash memory is a snowballs' chance in hell.
There's no point in worry about it, though. Back up your important data and forget about it. If the system in question is mission-critical and up-time is essential, then use two SSDs in a mirror set, and don't worry about it. If someone is going to get their head lopped off if there's any chance of the system in question failing due to SSD failure, then mirror your mirror-set to another mirror-set (i.e. use 4 SSDs) and back the whole mess up to an off-site location regularly. Sitting around biting your nails down to the quick isn't going to help anything.
Maybe it helps the author to develop a narrative, but the long and short of it is, the author's non-volatile storage unit died, he needs to replace it to get the system back and he can send it back to where he bought it from because it died under warranty. Or, he might want to have it destroyed locally if it contains proprietary information.
If you're in IT, I'm sure you'll see everything eventually break (including things like cases which don't make any sense at all) so why sweat it?
Mimetics Inc. Twitter
Waterboarding?
Well... Funny, but water mixed with electronics tends to produce situations where little communication takes place....
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Despite what others have said, this comes down to the brick wall nature of error correction codes. Every time you erase and rewrite a flash cell, you as wear to the transistors that make up the memory cell. Eventually (and probably immediately too) some of the bits won't read correctly. To compensate for this, the controller runs a mathematical function on your data, allowing it to recover from a certain percentage of bar bits. This is good, as that combined with wear leveling allows it to run a long time. However, one it hits that percentage, it's like hitting a wall and it can't recover.
...si hoc legere nimium eruditionis habes...
Then read it as "Samsung would not ship the replacement until it received the returned unit." This still implies a week's downtime.
Uh, for the massive performance boost you get from an SSD, they are totally worth setting up a backup job. Image the disk, set periodic backups to a server or even iDrive/Crashplan/Dropbox/etc and carry on with life. Hell, even leave the spinning disk in place and backup to that. For $60 you can extend the life of an old PC by several years simply by swapping in an SSD.
You should have backups anyway.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
I doubt that most home PC users have both the case space and the cash for a RAID. A user of a mainstream laptop sure doesn't.
Reports from the ISS are that 9 out of 24 SSD drives failed in an HP supercomputer they'd brought up there. Quite scary how fragile those things are from radiation.
Design for Use, not Construction!
That happened to me three or four times already. They die without warning. No SMART indication, nothing. It really pisses us off. Someone needs to technically give us some kind of anticipation. Maybe SMART is not supposed to work well with SSD after all.
One thing I like about spinning disks is that a lot of times the failure is gradual. Bad sectors and such and you have the opportunity to grab data off the drive (noting, you really should have backups).
With SSD, whatever the issue, it's more like losing a controller board on the drive, everything dies and ceases to operate.
So... I'll go along and say SSD is "better" and more "reliable", but when it dies, it dies hard. Just the way it is. (not talking about performance degradation... speaking about failure)
Correction: PROPERLY DESIGNED electronics wear out slowly. Improperly designed electronics may not even last past the warranty period. Since there is a huge demand for SSDs in increasing capacity, I can't help but think that manufacturers are pushing the bounds of reliability in favor of capacity. The manufacturers may just be relying on the SSD's built-in correction capability to correct for the decrease in reliability, but that will only get you so far.
had a 2gb memory card once. A Day One fault of one 512 mb block dead. Windows could not recognise this fault nor fix it. Instead writing to the card had corruption (obviously) when the faulty block was engaged.
Then Microsoft messed up by not offering a "try writing to all unallocated clusters" mode in the surface scan in chkdsk.
Improper handling of ungrounded components really can mess them up. They work but are defective. Take a look at some micrographs of ESD damage sometime.. ESD does not always kill a part it maims -- sometimes only slightly. Anti-static mats and wrist straps are no laughing matter, Okay. They are. But use them anyway.
"No fear. No envy. No meanness." Liam Clancy
Most of the time heat kills electronics. Either they get too hot and something fries, or they suffer thermal fatigue.
One of our competitors trademarked the term "hypothesis". From now on, we will call them "boneheaded ideas".
I had a Sandisk USB stick recently go read only. I had been using it as a hypervisor boot drive and the boot was crashing. When I inspected it, it was read only and any attempts to format it, diskpart it, fdisk it failed with some kind of error. I looked it up and apparently this is the designed failure route for these USB drives. When the controller detects an inconsistency or uncorrectable error, the drive is locked from writing so you can get data off of it.
SSDs really are unpredictable timebombs, so act appropriately - take frequent backups and use RAID if the downtime from a sudden SSD failure with zero warning is unacceptable. Any IT department that hasn't been prepared for the nature of SSD failures since long before they were available off the shelf was doing it wrong anyway.
I'm most worried about what SSDs mean for the Average Joe, whose data is largely protected by the predictability and recoverability of most hard drive failures. SSDs throw all of that out the window and lure them in with the warm glow of performance like moths to a flame. Average Joes need a real wake-up call on the importance of backups with the switch to SSDs.
"When information is power, privacy is freedom" - Jah-Wren Ryel
You can almost always pay for advanced replacement. You get your money back when they receive the drive. Then it is usually just a day you're done provided you don't have a spare around. If things are really that critical though then you've already failed and should have setup RAID for your SSD and had backups.
Not using TRIM doesn't have a huge effect on SSD life. Just performance. Write amplification adds some wear, but not enough to be drastic. And it won't cause sudden failure either - just normal wear on the wear-levelling curve. Sudden failure is by definition going to be something that's not related to routine depletion of a fixed lifespan.
Who the hell cares? Replace it and restore your data.
The data on a failing drive might be a newer version than the most recent weekly backup. I see value in backing up the newer version elsewhere as the first part of replacing the drive. But SSD failure modes allegedly make this newer version inaccessible sooner than HDD failure modes.
You should demand cross shipping for that. Any professional would.
So you can have peace of mind:
If it dies suddenly, without warning, it's 1) buggy firmware (I think this is by far the biggest culprit), or 2) bad components/soldering/cleaning on the PCB board, or 3) a really dumb controller that isn't doing wear leveling on every single thing (think the master index), so when a critical flash cell dies the entire thing is dead even though there's plenty of good flash left (this was common with crappy little 'SSDs' that were just Compact Flash), or 4) a badly designed controller that leaves the drive in bad state when power suddenly goes out and can't recover
If it sloooows down and starts getting more and more sluggish you've lost enough flash cells that the wear leveling is losing its capacity to cope. Take some stuff off the drive to give it some breathing room and prepare for its demise. I had this happen with one of the original Intel SSDs (the X-25M). It took ten years of continuous use, though - yes, just this year.
SSDs have a bunch of tiny wires. When you push electricity through wires they heat up, they're not perfect super-conductors. If you heat it up too much, it will of course burn, but they avoid that. Still, heating up a wire over and over will have some wear and tear. For big thick power-lines in houses, this doesn't have too much effect, but for tiny precision electronics, it builds up. And SSD's have a LOT of those wires with a little bit of manufacturing variance which makes some parts fail sooner.
They burn out the same way lightbulbs burn out. They don't have moving parts, right?
What, you don't have at least 32GB of RAM?
I see your point about prefetching most of your environment to disk cache. That's why Microsoft added the "SuperFetch" feature to Windows over a decade ago and Canonical added "ureadahead" to Ubuntu. But there are three problems:
First, many tablet computers and compact laptops lack slots for 32 GB of RAM.
Second, even on those machines that can take 32 GB, loading 32 GB when booting or when waking from hibernation takes a while before the prefetch stops being a source of read latency.
Third, when a file is written and flushed, the application that you are using still needs to wait for the data to be written to spinning rust in case the power fails or the kernel panics. That adds several milliseconds of latency.
Don't blame the OS. Blame "no backups." Failure should be expected and accounted for with a backup plan.
Should have skipped Intel and OCZ and just waited for the Samsung EVO line. I've installed dozens over the last few years and not a single failure yet.
Metal migration limits the lifetime of the interconnect in ICs. Absolutely a wear mechanism.
I'm going to disagree with the people saying that spinning disks don't give you a warning of imminent death. A bad spindle will start whirring, and steadily get louder, and my experience has been that most drives go that way. Hence, the old trick of sticking the drive in a freezer to get a few minutes more life out of it (because, you didn't keep your backups updated....again. :-(
This is a phenomena that should always be kept in mind when switching from mechanical to electronic systems. The electronic are usually MORE reliable, in the sense that they are less likely to go belly up, but WHEN they do, they won't give you any warning. I could arguably make my home-built airplane MORE reliable and feature rich by replacing the flight controls with a fly by wire system. But, one day a gate in one of the processors will fry itself, and the whole system will quit working at once. Woe unto me if I'm at altitude at that point. The mechanical system will require more maintenance, but it will slowly wear out over time, controls will get sloppy, and exhibit more play. That is the system telling me, "I'm getting kinda tired here. I'm getting old, y'all. Replace me. Screw it. I quit." It gives warnings to the operator that knows what to listen for.
So, the article does have a point. . . sort of.
Aah, change is good. -- Rafiki
Yeah, but it ain't easy. -- Simba
Doesn't know how SSD's work.
No offense to CS majors, but this EE major tends to understand "How a computer works" at a lower level than most of you programmer types. While not universally true, in my experience a Computer Science major generally get's outside their comfort zone with hardware once you get past "Plug it in and turn it on." I don't blame them, there is a lot of stuff happening at lower levels than a CS major needs to know to do their job.
That some CS major is concerned about how SSD's fail because he doesn't understand their failure modes is fine. We tend to fear what we don't understand and let's face it, there is a LOT of stuff going on inside a computer that high level users simply don't need to know. Heck, even I don't need to know some of that stuff and I've designed computing systems in the past. Fear not, if it works, it works, if it doesn't you just replace it anyway.
This ^^^. I had a brilliant CS college roommate. But when he built his first computer himself, the motherboard was held to the case with one screw. He couldn’t figure out why it was crashing all the time. Everything in the machine was barely in their slots/socket. This is back in the Pentium days. Days of VLB and very early AGP. And sometimes IRQ switches.
" in my experience a Computer Science major generally get's outside"
Yup, I can believe you're an EE. While you go on and on congratulating yourself almost as hard as a doctor, you can't even tell the difference between GET IS and GETS.
New SSDs, failure could be a die bond failure, a sometimes defect that allows it to pass inspection then fail. Or a ball bond to PC failure that can be intermittent as the package, solder ball, and PC change dimensions due to different thermal expansion coefficients. The tiny contacts on the PC versus relatively huge contacts on the mechanical hard drive make these happen more often on SSDs.
On older SSDs there could be degradation of the ability to hold or modify the stored charge that represents bit. Not likely unless you are a heavy duty user. Or metal migration from the mask layer, or metal migration at the bonding level wire physical aluminum or gold wires are bonded to the actual chip. Less likely are bonding failure to the underlying substrate as the wire material used is chosen for high compatibility.
Now Chris, feel better knowing just a few ways you can envision the failures?
- Tjp
I am in wallow with my inner money grubbing capitalistic pig. ... Oink!
Backup your data frequently. Stop worrying. Is that so hard?
Just imagine the unicorn in the drive died.
It's about as accurate as what you imagine happened to the spinning disk.
I had a 4 tb spinning drive fail, after only 2 years. It was 75% full. That is what is scary to me. The only narrative I came up with to explain it was that it was in my system, but powered on, 24x7. Now my backup drives are external and I power them on when I need them.
As drives get bigger, that is when I get nervous. I know, there's options to mitigate that, but I'm on a budget. I just migrated my OS to an SSD a couple of months ago, and still have spinning drives holding everything else.
My beliefs do not require that you agree with them.
Realistically speaking, he almost certainly got a replacement, but return policy he had required it to be returned.
However for electronics of this class, the manufacturer in all likelihood *could* repair it. The neat thing is if they do repair such a disk, it could come back with the data intact. In practice, I don't think any manufacturer would offer such a service or even try.
XML is like violence. If it doesn't solve the problem, use more.
I'm guessing the author never lived through the era when there were a lot more companies in existence for mechanical HDs than there are now. HD's can spontaneously die from a failed motor, electronics failure or catastrophic crash. Some small companies went completely under and were swallowed up by larger manufacturers due to massive defects. SSDs have gone through the same era as well with buggy firmware. Generally speaking thou if you stick to the big manufacturers like Samsung and Intel the chances of fatal issues goes down a lot. That said an SSD is not a guarantee of safe data. They're far more reliable but circuit failure or static electricity can kill SSDs. Besides, SSDs won't save you from an accidental erase all.
Donald? Is that you?
I've fallen off your lawn, and I can't get up.
Just chiming in My Crucial M4 128GB (Micron) drive also died on me 2 months ago after very mild use since February 2013. It was my OS drive in a Windows 7-10 desktop which O mostly used for 3-5 multiplayer games through the years, or the odd media consumption. It was a machine that was on about 1/20 of the entire 5 years and 8 months.
There's another problem I've found with SSDs in addition to their failures occurring with no previous warning signs. That is that the process of obtaining warranty replacements can be terrible.
Perhaps because hard drives were expected to fail, manufacturers put procedures in place (such as "Advance" RMA) to ship a replacement very quickly. This is important when, for example, you have a single-drive failure in a RAID configuration that can only tolerate losing one drive.
My experience with obtaining two warranty replacements on Intel M.2 SSDs has been really poor. In each case the replacement drive took so long to arrive I had to purchase a replacement drive in the meantime.
Yeah, that's the answer...turning SSDs into WinModems.
I object to power without constructive purpose. --Spock
You can get the best of both worlds by setting up a RAID of both an SSD and a platter drive! :-P
Quite so. As it happens I was just fixing a dentist's office computer yesterday, and used the dental air blower to get the dust-bison out of the heat sinks since I didn't have any compressed air on hand. Let me tell you I was *really* careful not to touch the water jet button. Clearly whoever designed the "two small identical buttons side by side" interface never intended it to be used in a setting where a stray jet of water could be a major problem.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
Most likely reason is a firmware bug cause enough corruption that it can't even low-level format. If it were a prototype that a developer could diagnose, it would be easy for them to patch it and get it going again. But without that specialized environment you SSD and the data on it are trash.
In many ways I think I would have preferred the raw NAND systems like SmartMedia (now obsolete), where the host had the real brains and the media was as primitive as possible. SmartMedia formatting was about conforming to a software standard on the host side and was managed by a driver. A real driver that a could be debugged with ordinary tools, not some obscure firmware embedded in a device.
“Common sense is not so common.” — Voltaire
We have experienced from mechanical, SSD, and NVMe drives that there are points of failure that we can detect, and there are points of failure we can't. Most cases where an unpredictable failure occurs is almost always at the power source, and is mostly indicative of voltage irregularity in our tests with bad drives from these 3 types. While we'd like to think that new hardware will hold up to a degree of it's certified life span; voltage as a whole to power said hardware will almost certainly add the anomalous layer for a margin of error from minimal to catastrophic.
The chips store data in a capacitor.
The capacitor is connected to (or is the) the gate of a mosfet so the state can be read.
To charge or discharge the capacitor, electrons must be forced over the insulation later that stops the capacitor discharging on its own.
Every time that happens the insulation breaks down a little. Once it's all gone, the cell can no longer store data.
It's a gradual process that happens every time a cell is written to or erased. SSD's wear out as they're used, it's how they work. You should treat them as a consumable.
Or something randomly broken. like a solder joint from thermal cycling or something.
Well,to be fair you *didn't* fix it - you just worked around it. Almost as good in many settings. I've "fixed" several hard drives in a similar manner - one section of the drive is clearly bad, and spreading when used? Fine, re-partition it so that that section, and a generous buffer zone, are never used. They typically work fine for years after that.
Certainly not something I'd generally recommend given the nature of such HDD failures, but perhaps justifiable if you just want to buy some more time before an upgrade, or until a kid destroys the thing more permanently.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
That's the long and short of it.
See SSD Failures in Datacenters: What? When? and Why?.
Failures include retention errors caused due to leakage current, which worsens with time when not acted upon. Second, they also suffer from phenomenon such as read disturb and program disturb errors, where read or program of a row or block of cells affects the threshold voltage of untouched cells in its vicinity. data retention, program disturb, read disturb, endurance, and power faults.
Flash controllers have proactive and reactive mechanisms in place, to prevent the flash error propagation to higher levels in the system stack. Consequently, not all of the above-mentioned failures propagate to upper layers. But, ones that do propagate can result in fail-stop failures.
I spent 3 years on a "deep dive" into EE basics, analog circuit design, then microcontrollers, and it really improved my software development a lot.
I don't think this is a natural blind spot in CS, I think it is just manufactured ignorance by dividing the fields in an unrealistic way. Which seems to have happened during the rush to train workers during the .com boom, so maybe it wasn't even thought out at all.
Does anyone really know why a spinning disk dies? Sure - maybe if the last operation was "dropped laptop down stairwell"
A narrative over what went wrong?! Whenever a HDD failed a light came on the RAID array - and I'd find a package from FedEx on my desk at 9AM with a replacement disk in it. As for personal computers - the drive stops working and you lose data.
What is there to think about?
I do agree about the "timebomb" thought. I know that SSD just give up the ghost. On a HDD many times "check disk" starts reporting a high number of failures and you can be prepared...except when the head falls off the arm. That's a rather rapid failure.
SSD have a write-lifetime that I can't predict. HDD goes until it doesn't work anymore. In both cases you break out the backup tapes.
If you begin to notice vibration from the SSDs then you know they are near the end of their life.
I'll see your senator, and I'll raise you two judges.
We set them up in either a RAID or EC configuration or other redundant configuration , so that the operations department can swap them out when they fail without downtime.
Unless we start to see an unusual high number of failures, we don't care.
L'Idiot
Just out of interest, how much time in your day is spent stalking me on slashdot after your anti-science drivel got exposed in that one argument?
Well, I do think it's natural for CS majors to be a bit farther away from hardware. Let's face it, much of their work these days doesn't really care what operating system they run on much less the hardware it's actually running on. I don't blame them, really the state of programming has evolved away from hardware dependence, and that's a good thing..
Where I understand hardware details of what's happening behind the programing model seen by the CS guys and gals, and I believe that I have a different perspective when doing software development, I'm not sure they would benefit all that much. Programming Java is pretty hardware agnostic anyway, C/C++ a bit more specific (assuming you have the libs and compiler), but still largely portable unless you are handling actual hardware or kernel level stuff. My hardware knowledge really only serves to make me more aware of performance implications of my choices perhaps, but the CS folks do just fine with most higher level languages.
So I don't agree, CS folks really don't need to know all the same stuff I do to program. It used to be true, it used to be valuable to understand what the hardware had to go though, both to be able to optimize your code for performance and size and get it to do what you wanted. However, with the advent of the higher level languages, most CS folks don't interact with the hardware anyway, but abstract programming models like the JREs which for all the world look identical regardless of the hardware being used.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Don't blame the OS. Blame "no backups." Failure should be expected and accounted for with a backup plan.
While I am the first to agree to this at the enterprise server level, it's far more difficult for the consumer or typical desktop user, especially for laptop users. RAID isn't always an option for laptops (frequently it's impossible) so you're left with some sort of external (USB or Thunderbolt) backup device or cloud storage. The former is difficult for road warriors and is nearly impossible to schedule since it's manually attached. The latter depends on an always-on Internet connection to have current backups.
My strategy was for laptop/desktop users to have their My Documents (and any other crucial directories) mirrored using OneDrive (included with Office365). It worked most of the time but nothing could be done if a drive failed while someone wasn't connected to the Internet. Any changes made since the last sync were irretrievably lost.
In the end they will lay their freedom at our feet and say to us, Make us your slaves, but feed us. - Fyodor Dostoyevsky
I blame the autocorrect software.
Yea, with water, I see a reduction in resistance too... :)
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Disclaimer: I've known Chris since we were CS undergraduates together in the 1980s, and we currently work together in the CS Department in Toronto. It may seem a bit odd to some that a hard disk failure isn't unnerving but an SSD failure is. That's because one of a good sysadmin's skills is properly focused anxiety, used to motivate a mental model of how things can fail, and what to do about it. Data storage is a key part of this mental model, since data access loss, or even worse, data loss, is a major risk. That's why it's helpful to know how disks work, how they behave when they fail, and how likely it is for such things to happen. Chris has a few decades of experience in dealing with disks. SSDs take the place of disks, and they store stuff just like disks do, but they work differently, and they behave very differently when they fail. In particular, SSDs often don't seem to give any indication that things may be wrong: one moment all is well, the next moment, all is dead. So instincts honed over a few decades of experience with hard drives don't apply. Of course Chris (and we all) will develop new instincts as we get more experience with SSDs. But in the meanwhile, it's indeed unnerving. And no, this isn't some sort of profound insight. It's merely an observation. Many experienced sysadmins, I think, will "get" this. People newer to the field might not. That's OK.
That's the crux of the article. I should care why? He's a technical guy, he knows about memory. He just refuses to apply his knowledge to get rid of his paranoia. This guy's nothing but a low-level conspiracy theorist.
As I wrote in one of my books: "They're all alike. Conspiracy theorists. They'd rather live in a terrifying fantasy world than the real one."
Do you get this anxious when a RAM module fails? There really is no difference between a RAM module failing and a SSD failing...
Uhh...people don't usually store critical files in volatile RAM. Kind of a huge difference there. Further, RAM failures may crash the computer but it rarely destroys anything else in the process. A mass storage failure -- be it HDD or SSD -- virtually guarantees you'll lose whatever data you had on it. Your only recourse is RAID (which isn't an option on most laptops) or some sort of backup (which is difficult to enforce on mobile users).
Yes, you can blame users all day long for not backing up their data. It doesn't help when you're still responsible for IT as a whole. The problem lands on your desk whether you want it or not.
In the end they will lay their freedom at our feet and say to us, Make us your slaves, but feed us. - Fyodor Dostoyevsky
Sandforce controllers self-brick at the first sign of trouble to prevent competitors from reverse engineering their controllers. Or at least that is the reason stated for their crappy design. IIRC, Intel developed a customized version that has better failure modes.
This is the uncanny valley in which the world of REAL slowly sinks, sinking...sunk into the technological relative world of NOW.
There is no bridge between. You stand stranded on the shores of reason while the world in which you live sinks away, out of sight and out of mind.
Millennials know the futility of questioning the NOW, its irrelevant to wonder ' why?' Just BE now!
Enlightenment as to why, what went wrong - much less how to prevent bad things is not among possibles. Shit happens!
Exactly this. I bought a shit SSD, it lasted three years. Not too bad, I suppose. When it died, which it did last week, I was back up and running in an afternoon - including the time taken to drive to the store and buy a new one.
It's a really odd article in any case, why be so paranoid about the precise failure modes? Hardware is hardware, and it can break. Plan for it, and you won't have any problems.
I'm not sure what they call Computer Science these days, but my bachelors had a required digital design component. We started by wiring together transistors to build a gate. When you'd demonstrated that you could use 74HC00s, and you had to build an adder. When your adder worked, you were allowed to use an ALU chip. You had to set the thing up with supporting logic and DIP switches and invent a machine code to demonstrate instruction processing and register transfers.
In the compiler class we started out by writing a simulator for that hardware, then an assembler, then a compiler.
Because yes I am your God, man.
"No fear. No envy. No meanness." Liam Clancy
I'm not sure if you're a troll and are trying to evoke annoyance, or if you suffer from severe reading comprehension difficulties and are trying to evoke pity. In me, you evoke both.
I was actually thinking that if they had more understanding of the hardware, they'd have a better idea what the layers actually are, and they'd end up with more portable code not less portable code as you seem to imply. Knowing about how hardware works helps to be more hardware agnostic, because if you're using intermediate layers with no idea of the hardware and OS coupling that it creates then you'll do it more often.
I was actually thinking that if they had more understanding of the hardware, they'd have a better idea what the layers actually are, and they'd end up with more portable code not less portable code as you seem to imply. Knowing about how hardware works helps to be more hardware agnostic, because if you're using intermediate layers with no idea of the hardware and OS coupling that it creates then you'll do it more often.
Yea, I see what you are saying, but remember they are stamping out CS degrees with little more than Java and Database Skills. The whole point of Java was to let you ignore all that hardware stuff though abstraction layers any way. Most of them don't need to know how to dig though all those layers to do what they need and with Object Oriented concepts, hardware is becoming trivia to them.
But I agree, a bit of understanding of hardware is a good thing, especially when you start talking recursion and how pointers/references are actually working. I've always been amused at the BSCS holders who didn't understand what the call stack was or how they where killing performance with all the objects going in and out of scope, or why the math was being in done using integers when they wanted floating point (or vice versa). I just don't know if they have the scope in an undergraduate CS curriculum to throw that stuff in. Many won't need it, use it or remember it anyway.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101