What's up with SSDs dying for no reason with plenty of write capacity? I thought they were supposed to more reliable than HDDs.
>>109051520HDDs die for no reasonSSDs die for no reasonif you absolutely need absolute reliability just set up a raid config
>>109051520>>109051532This. I had a lot of vids of my ex on an SSD that I would jerk off to from time to time. I popped it into an enclosure and tried to mount it only for it to just not workGparted just says it's formatted in an unknown format, but I'm sure it's just fucked. Should have set up a raid for this shit
>>109051532Yeah, but at least HDDs had an excuse with moving parts. SDDs just seem like shoddy engineering. SMART doesn't even tell me when they're about to die.
>>109051520I've noticed SSDs "deteriorate" either due to high temps and unusual use or both. Silicon lottery probably factors into it too, bad chips don't last long.And what I mean by "unusual use" is the normalization of Windows hibernation mode that kicks on if you leave your PC inactive for 30 minutes or something like that which is lame.Quick boot and hibernation will fuck up any HDD/SSD over time thanks to Microshaft and it's the first thing I turn off if I set up Windows because it fucking sucks.
>>109051565maybe make a back up? you fucking retard!
>>109051565SSDs honestly make horrible if not terrifying cold storage medium
>>109051605thisbrowsers and windows both rape your SSD by constantly reading and writing to it. I remember on my old pc I had a slow HDD with windows 10 installed, task manager would consistently show 100% disk usage by "System", this is the reason you now "need" an SSD
>>109051520>I thought they were supposed to more reliable than HDDs.No? It's common knowledge they're less reliable.
>>109051565retarded coomer
it's very likely the nand is ok, but the controller broke.the controller is what gets hot. There might be other passive components which might also be a cause for failure.nand is very resistant to total failure, so it's possible that parts of the data might get lost due to the nand, but a total failure is unlikely to be nand related.TBW is just the time when the firmware will lock your data to read-only so that you have time to migrate your data, this happened with my 128gb msata ssd. Which is helpful since HDD's are designed to run-until-they-die, since if I was a server operator, I already have raid, I don't care if the HDD dies, I want my moneys worth.>>109051565to me this sounds like user error.I think it's possible to accidentally configure a disk so that it requires the original HDD to be installed on the mobo for it to detect what format it is in. I don't know exactly what's wrong, but I think the EFI partition on a separate disk from the actual data.This would be the case if you only detected this failure after taking it out of your original PC.you should be able to recover the data, I think I used testdisk to recover some data after accidentally formatting the disk to a different format. I don't think the EFI partition actually contains the super critical info, like the filesystem, so a brute force recovery should work.
>>109051793ya, I think my controller broke. too bad I don't have a thermal cam or else I'd try to go rossman on it. RIP Sabrent Rocket 4
>>109051786>source: my ass
>>109051786yup everyone knows when you reduce moving parts things get less reliable
>>109051796if you are >>109051565you can test if you SSD failed by re-formating it. If it doesn't fail during the format, it's probably still good.You will lose some data, but you will be able to recover something hopefully after this destructive test (but you could also just use testdisk, I'm sure AI can walk you through it)
>>109051822not him, and my bios doesn't even recognize the disk anymore. it's toast
>>109051584>SMART doesn't even tell me when they're about to die.Also doesn't tell you when your HDD is about to suddenly die for no reason.And the moving parts of an HDD aren't the only things that die.
>>109051786>It's common knowledge they're less reliable.A few vocal faggots with no data while all the data is against them.
Maybe buy an SSD that says Samsung on it next time.
>>109051839then I guess they invented SMART for no reason and it does nothing?
>>109051520>>109051532/thread
>>109051851SMART is monitoring other things.It can't monitor if a controller IC is going to die. Or if a power regulator is going to die.
>>109051870what exactly is it monitoring then?
>>109051875read/write performance, read/write errors, seek times, and shit.It's essentially monitoring the media, if it can be accessed reliably, and if that starts degrading (bad sectors, rewrites, slowing access) it can warn that the drive is DYING.Sudden death causes, like heads crashing or ICs dying, don't give you warnings.
>>109051930yeah ok it's useless. thanks drive defender
>>109051520>I thought they were supposed to more reliable than HDDs.where did you ever hear that? everyone knows SSDs have limited cycles and can randomly die after a couple years, where HDDs can get decades
>>109051951is that how they were advertised? if no, how did you know unless after experiencing a failure?
>>109051951>knows SSDs have limited cyclesalso notice i pointed out: "with plenty of write capacity". I believe my drive died at 2% or 3% of write capacity used, ssd shill
>>109051520>What's up with SSDs dying for no reason with plenty of write capacity? I thought they were supposed to more reliable than HDDs.false advertising or meaningless data that they generated in perfect lab conditions. This reminds me of the mechanical keyboard meme that switches are supposed to last at least 100m actuations and if you do the math that's basically 10-20 years of typing a full novel every single day. It's obvious bullshit.
>>109051870>if a controller IC is going to diesilicon life management IPs are realyou can just expose register valuescompanies decides not to
>>109051565I do the same but that's why I have backups.>>109051793Photorec and testdisk work very well but they don't recover 100%. The first step is always to generate an image of the whole drive
>>109051769a ramdisk can solve this
>>109051951>everyone knows SSDs have limited cycles and can randomly die after a couple years, where HDDs can get decadesA few vocal faggots with no data while all the data is against them.Yeah, HDDs can get decades, they can also die randomly after a couple of years.SSDs can also get decades. I have still had zero SSD failures.
CIA/MOSSAD gets to it
>>109051520>>109051532At least HDDs warn you with clicking sounds before dying.
>>109052480>At least HDDs warn you with clicking sounds before dying.That is one of the failure modes.
>>109051520Overheating, remember to use proper cooling fans and heatsinks on them
>>109051520Not all SSDs have DRAM.
>>109051769>browsers and windows both rape your SSD by constantly reading and writing to it.No they don't. Jesus christ is this board retarded. Any none shite SSD comes with a TBW that is so fucking huge, you could rewrite a third of it ever single day for 5y. We have TBW in cheap consumer drives, that we had for server drives a decade ago. Just buy a none shite SSD. It is easy to spot them, they come with a 5y warranty.
>>109051520Samsung SSDs have a fuse on them which sometimes blow, leaving the entire drive not getting any power whatsoever, despite every component still flawlessly working. I once repaired a 860 Evo by removing the eFuse IC and bridging its input/output. The drive is still working since (52k power on hours as of now).
>>109051565The enclosure probably forced a formatting. So, user error.
>>109053803>my converter cable box thing formatted it instead lf just connecting pinoutsShit product if that was true.
>>109053793wait whatelaborate?first time hearing that
>>109051866dd if=/dev/random of=/dev/ntpd0p2 bs=4M
>>109054615
>>109051520user errornever happened to me
I boughted an XPG ssd and only looked at the reviews after boughtening it died a few days after I plugged it into the slot, the controller on those are ganges tier
>>109054667Damn. I thought XPG SSDs were supposed to be good. The more I read about SSDs dying, the more I get the impression that Samsung SSDs are the only ones that the controller doesn't shit the bed after some time.All my Samsung SSDs both SATA and M2, some already more than 10 years old, are still working fine.>>109053793I've heard about that before, but seems to be pretty rare
>>109051520Anything can die randomly. Your Mobo, GPU, power supply, etc.Nothing has really came out to convince me SSDs in general have a higher failure rate than anything else.>>109054667>>109055141I have a sneaking suspicion that when an SSD dies, it's probably NOT the controller or NAND.All SSDs have support components like the PMIC and other passive components. There is also general construction quality like the PCB and soldering.You can have an SSD with a good controller and NAND but everything else is subpar. Non-first party SSDs kinda have to skimp where they can.
>>109051520>What's up with SSDs dying for no reason with plenty of write capacity?controllers can die too. nand are usually very reliable.>I thought they were supposed to more reliable than HDDs.depends on use case, they survive better in extreme environments, they draw less power and take less space but have less capacity than hdd, it's all about tradeoffs.>>109051584>SDDs just seem like shoddy engineering.taht's the thing with everything digital, you rarely have "artifacts", it either works or it does not>SMART doesn't even tell me when they're about to die.because how the hell are they supposed to know when a capacitor/resistor will blow?>>109051769>browsers and windows both rape your SSD by constantly reading and writing to itit's not really an issue in 2026, ssd endurance is>>109051786wrong.>>109051843I agree for pre-980 pro samsung, anything newer than that is of poor quality and I see a lot of defective drives>>109051951> everyone knows SSDs have limited cycleshdd have the same limits btw, you can't infinitely write on them, most enterprise ssd have better DWPD than enterprise-grade hdd btw>>109053695>Any none shite SSD comes with a TBW that is so fucking huge, you could rewrite a third of it ever single day for 5y very high.also reminder than these TBW and DWPD are NOT hard cap, most ssd can write well beyond this limit>The more I read about SSDs dying, the more I get the impression that Samsung SSDs are the only ones that the controller doesn't shit the bed after some time.sadly not true anymore and it's on purpose, I think samsung want to escape the consumer ssd market.
>>109055452>anything newer than that is of poor quality and I see a lot of defective drivesMaybe it's just an issue with the faster drives that run hot? I've had zero issues with my "slow" evo drives.
>>109051520A few years ago I bought several SSDs. Mostly chink ones, but also some WD as well. I've had 2 of each fail. Thankfully, WD had a warranty and Sandisk replaced it, but it will probably fail again. I've had HDDs fail, but usually after much, much longer use and likely due to dropping them or something mechanical.
>>109051769>be storage device designed to read and write>NOOOOOO NOT LIKE THATSSD excusers need to be executed. It's literally a storage device designed for massive reads and writes. It should be able to hibernate and browser cache just fine without hurting the drive.
>>109055479>Maybe it's just an issue with the faster drives that run hot?you can't really trust samsung consumer ssd anymore as they have "regular" firmware issues (like killing 990 pro within days), they swap nand for lower quality ones under the same sku and probably more things we don't know about, basically the western digital of the ssd world, you can only ever trust their datacenter lineup, the rest is hit or miss, with the misses rate increasing with time.
>>109051786retard>>109051842^this
>>109051520NAND is probably fine. Controller shit itself.
>>109055227$3k+ down the toiletPC Master Race of faggots
>>109052480>>109052504My HDD is occasionally making a chirping sound. Is that the same as clicking?
>>109051520I've been using SSD's for about 15 years now and have yet to have one die.I should probably stop being retarded and actually set up a backup though, I currently have none.
>>109056718Solder shit itself due to repeated thermal stress and lack of lead.
>>109051532what if all the hdds or ssds in the raid die for no reason?
>>109061782then it's the fate god assigned to youconsult your local data recovery services
>>109051520dogshit controllers are the bane of ssds. had three dead drives and all due to bad controller. the data is fine and intact but the controller cannot read or write properly.
what about data recovery?I know on an hdd you can recover a lot from the physical drives but how do ssd compare?
>>109055664Servers have cooling, anons here don't know about it
>>109061824the data lives on the nand chips and they harvest them and put them on a donor ssd. if the nand chips are dead then so is your data.
>>109061885so basically the same deal as shattered disks
DRAM cache was a mistake
>>109060481you can check the health of your HDD, sadly I don't think they have a % life remaining like NVME's, since HDD's don't have a TBW metric, but you can see if your hdd is experiencing end of life errors.
>>109061868...and?ssds are pretty much unaffected by temperatures, nand lifespan actually INCREASES at higher temperatures, HDDs on the other hand suffer from more degradation the higher the tempso what you're saying makese no sense, if anything this means that on regular computers hdds will last even less.btw you also have to account for the fact that backblaze uses enterprise grade hard drives, but their ssds are mostly regular crucial mx500 or shit like that which regular consumers also have access to, that also doesn't help.
>>109062475dram cache is awesome. dramless ssds are big memes. if your ram is ever slightly unstable so is your ssd.
>>109064013The controllers get killed by thermal stress.
>>109064786what if we moved the controller to somewhere integrated like the PCH or CPU? that would solve it right?
>>109051520>I thought they were supposed to more reliable than HDDs.yeah that was a lie they die without warning or maybe that's true but they still die without warning
>>109051565back to instagram, sex-haver
Long term storage? use HDD, external, and only use it when you NEED toshort term? SSD, and they do tend to die more
>>109064786I have never heard of a cpu killed by thermal stress....and an ssd's controller is literally just a cpu.do you have any data to back this bullshit up? of course not.
