Thread 106459760

I did a small research (pics included) and I've noticed there's something strange happening with recent WD hard drives.

For instance let's take a look at this 4TB WD Purple (WD43PURZ), the datasheet says it's CMR but if you look at the graph it behaves like SMR and TRIM is supported.
https://files.catbox.moe/xdudsc.jpg

Same applies to 4TB WD Blue/Red Plus models, all these have 256MB cache and 181 buffer zones (like WD Red EFAX) instead of typical 61 like normal CMR WD drive.
https://files.catbox.moe/htn2k3.jpg

Look at these 4TB drives - they all look like this SMR WD Red EFAX drive.
https://files.catbox.moe/cz46vl.jpg

And 6TB WD drives are also identical to this EFAX drive.
https://files.catbox.moe/nd55ho.jpg

So WD is lying again like they did when they released their first SMR Red. Avoid all 4-6TB models, unless it's enterprise class (Gold, Black, Ultrastar).

2) Seagate is also guilty of this. Compare this 6Tb Ironwolf model to a SMR Barracuda,
https://files.catbox.moe/xod0q4.jpg

They both belong to the same disk family (V11) and the 256MB cache is another indication.

3) Most recent hard drives except for enterprise drives have hard locked head parking timers (from 8 sec to 5 min) and it's impossible to change it.

4) Some modern WD drives have that little piece of red tape glued on the chip. I have no idea what it does but the drives that have these red things run much hotter that the ones without it. Some recommend to remove it.
https://files.catbox.moe/qhtr5u.png

I'll list the WD drives that I'd recommend and avoid (continued in the next post).

Here's the list.

1) Best:
WD85PURZ (Purple) - genuine CMR based on HGST, good read/write graphs, no head parking.
https://files.catbox.moe/8owsjz.png
Acceptable:
WD80EFPX(Red) - same as WD85PURZ but the head parking timer is locked at 5 min
A bit worse:
WD80EAAZ (Blue) - same as the previous two 8TB models but the head parking timer is locked at 8 sec
Good:
WD20EARZ (Blue), WD20EFPX (Red), WD23PURZ (Purple) are all physically the same drives with different TLER and head parking settings that can be turned off.
https://files.catbox.moe/qzh574.jpg

These are genuine CMR with good read/write graphs and 64MB cache, 61 buffer zones.
https://files.catbox.moe/gcseg2.jpg

If you have the money then choose either larger 8TB+ drives or Ultrastar/ Exos/MG but they run hotter and require extra cooling.

2) Rather avoid unless proven that they're CMR: *All 4-6TB WD Blue/Red Plus/Purple are pseudo-CMR drives based on previous EFAX Reds (look at the pics in the previous post)

WD40EFPX, WD60EFPX (Reds)- pseudo-CMR with 256MB cache and 181 logical zones (no TRIM), head parking timer is locked at 5 min.

WD40EZAX, WD60EZAX (Blues) - same as the previous drives but the head parking timer is locked at 8 sec.

WD43PURZ, WD64PURZ (Purples) - same as the previous drives but also has TRIM and a weird write/read graph, disabled head parking.

So here's the conclusion - WD simply doen't give a fuck and they still sneak SMR trash into NAS and Survelliance models and they need to be called out for their bullshit, otherwise I'm afraid that the worst might happen - CMR 3,5" drives will soon cease to exist except for maybe enterprise drives just like they stopped making 2,5" CMR drives, all hdd manufacturers are deliberately lowering the drive quality and lifespan by adding unturnable head parking timers, PWL, adding red tapes on the chip which increases the temperature etc.

>>106459760 (OP)
>Western Digital is still lying about their "CMR" drives
Might you be the same OP who posted in April?
https://desuarchive.org/g/thread/105032250
>WARNING! Don't buy any Western Digital hard drives

>>106459863

Yes, it was me, but this time I want to show concrete examples that expose misinformation coming from WD and tell people which drives to choose or avoid. I just did another long research (finding out info about drive vendor, family model, read/write graphs, actual cache, if they're actual cmr or smr etc). I think it's necessary to call these greedy lying fucks out.

I wanted to make such post back then in that thread but 4chan got shut down for two weeks at this exact moment and I totally forgot about that. Sorry about that, just wanna be helpful.

>>106459954
>I just did another long research
Do you test the drives yourself?
>Sorry about that
That's quite all right.

I've moved to seagate 2Tb green for storage.

Thank you for your service.

I was actually deciding between buying a large WD Red (22/24 TB) or just going with 8TB Blues.

My main concern with the reds is that many people report an annoying as fuck recurring sound which I believe is PWL whereas the blues wouldn't have that issue.

Are head parking timers really not configurable? I remember people messing with their timers back when greens existed.

>>106459778
Shit, I had bought a WD40EFPX drive shortly before that original post. Is it safe for data storage? Any recommendations to extend its lifespan?

OP is a seagate shill. Never had a WD drive die on me. Had multiple maxtor/seagates/toshibas off themselves for no reason

>>106460065
lol you sound like a wd shill. what the fuck does that have to do with wd scamming people with smr drives?

>>106459969

Yes I did test three 2TB Red/Blue/Purple drives personally (they belong to the same disk family CarmelR Plus 2 Lite) and I'm happy with them because they're turned out to be real CMR drives with 2TB per platter density and I could succesfully disable head parking via Victoria's APM settings (on purple it's automatically disabled). Although people would prefer bigger drives.

I also did buy a pair of WD80EFPX (which are based on HGST drives and run at 5640 RPM despite that 5min head parking which can be bypassed via KeepHDAlive), one 8TB HGST, a couple of 2TB HGST as system drives and managed to avoid 4-6TB Reds/Blues/Purples once I found out that they're not real CMR. Unfortunately 4-6TB drives are taking the most hit right now.

>>106460020

It's technically impossible to disable head parking on 4, 6 and 8TB Reds/Blues. Only via software such as KeepHDAlive if it's at least 5min. 8 sec (which 4-8TB WD Blues have) is too fast to react and will accumulate thousands of load/unload cycles very fast. Yes PWL is another deliberate act of sabotage because older drives didn't have it and despite that many of them are capable of working 10+ years with no issues.

>>106460093
I thought it wasn't recommended to buy Purples because TLER is locked to 0 on them, because for CCTV it's more important that it keeps recording than to retry if there's a write error.

>>106460093
>Victoria
Do you know if it's possible to accurately test an OS drive with Victoria? Meaning, is it unnecessary to wipe a drive, connect it as a secondary, and test it that way? For Victoria, in Test & Repair, if one were to wanna test a drive the way you do, should Verify or Read be selected? Is a quick scan enough? Or does it need to be a full one?

>>106460049

Yes absolutely. It's just not a genuine CMR drive, but it's not a typical SMR drive either (but it's not gonna behave like a typical SMR (with catastrophically low speeds when out of cache) due to more zones and better zone configuration.

Here's the graph, it actually looks pretty good, like an actual CMR.

>Any recommendations to extend its lifespan?

It has a 5 min head parking timer, long enough to keep it awake via KeepHDAlive if you're worried about load/unload cycles. But 5 min is not that bad, it's not gonna accumulate hundreds of thousands of load/unload cycle under a few years like some WD Greens/Blues.

>>106460065

I did point out the fact that Seagate also does sneak SMRs into NAS drives, like this Ironwolf ST6000VN001 which is based on SMR Barracuda.

>2) Seagate is also guilty of this. Compare this 6Tb Ironwolf model to a SMR Barracuda,
>https://files.catbox.moe/xod0q4.jpg

All three manufacturers are doing this currenly. They all act the same, especially when you notice that all three of them introduce "pro" versions of each drive category (wd red pro, ironwolf pro, n300 pro) while previous manufacturers such as Fujitsu, Hitachi, Samsung, Maxtor had their own unique hdd making technologies.

>>106460093
What's the big deal if drive is smr/cmr, most are used for data storage with infrequent access these days. I just got 8tb drive for nas, didn't even bother to check what kind it is.
Got wd red because previous drive in nas was also wd.
As long as it doesn't die on me I don't care if it's smr/cmr and I have pretty good luck with drives, never had 3.5 drive totally die on me, 2.5 is another story.

>>106460118

No, it's only applicable if the drive is connected to DVR. In a desktop it's gonna behave like a regular drive. As I said before it belongs to the same disk family as WD20EARZ, WD20EFPX.

>>106460153

The drive needs to be uninitialized, unallocated so the OS won't interfere. Download Victoria 5.37. Look at this pic (choose read/ignore and then SCAN). Read test will reveal any bad sectors, the scan will take several hours depending on the drive size (from 3 to 16 hours). Write test will wipe out all of the data so do it before copying anything to it, but the read test alone is enough to test the drive's health.

Also write and read graph must be identical speed wise. If the write speed is lower then it's SMR drive

>>106460234
>The drive needs to be uninitialized, unallocated
So just connect the drive (to be tested) to the mobo after already having the OS drive (separate one) connected and test it with Victoria? The drive to be tested doesn't need to be wiped via nwipe (or a similar program)?

>>106460227

>8tb drive for nas

If it's Red Plus then you're good. It's actually a HGST drive (should make it more reliable) with reduced RPM but has WD firmware and locked head parking.

>What's the big deal if drive is smr/cmr, most are used for data storage

For write once, read many it's not that crucial as long as you don't fill them completely and not modify the files too often. It's just SMR drives are inherently worse and less reliable in every aspect even when it comes to cold storage. SMR heads work in a much more complicated method, especially when working with logical zones, they are much harder to defragment, they also seem to fragment files by themselves even when nothing happens (I can show the video, where one data recovery specialist tests smr drives with PC3000 but it's in russian). They have more complex built and more fragile heads with a second-level translator module 190 that has to work several times as hard as a CMR drive head for the same kind of operation.

>>106460290

Yes, if it's a secondary hard drive (new from the store) then just plug it in, don't do anything with it (like initializing etc, it must not appear in "My PC", only in disk manager as an "unitialized drive"). This is how the test should look like. If it passes the test with no warnings then initialize it, allocate, assign a letter, quick format it etc.

>The drive to be tested doesn't need to be wiped via nwipe

If it's a new drive then no.

>>106460162
So it wasn't a waste of a purchase then, good. Thank you very much!
Hopefully something is done about WD's false advertising. Maybe some techtuber could do an exposé on it to get the ball rolling?

>>106460296
ok man, thanks for reasonable explanation.

>>106460335
>Victoria
Do you know of any programs like Victoria, but can operate via command line? That way, one could, hypothetically, test one drive without needing a second OS drive for the testing program. Example: on a single drive system, use SystemRescue to boot into nwipe, wipe the drive, and then use a command line equivalent of Victoria to do a proper test of the drive.

I'm too poor for cmr

SEAGATE CHADS WIN AGAIN

>>106460296
>If it's Red Plus
Just checked invoice, it's "red plus nas" so should be cmr.
Drives are so damn expensive tho, had to shell 200€ for it, feels like such a waste.
Old nas drive was WD40EZRX so actually green, though it was red. It's good drive tho. 45k hours and absolutely raped by swapping in shitty synology nas with severe lack or ram.

>>106459760 (OP)
I've been wondering about all this new SMR, HAMR shit coming out. Thanks for helping with this. These HDD companies are indeed lying fucks. What irks me is the larger drives they kind of bury the speed now. It's such a basic fucking question. Is it 5900 or 7200? You have to go on a scavenger hunt to find that now. Ridiculous bullshit.

>>106460401

No, I've never tried anything except Victoria, it's good enough for me since I ever discovered it, does the job very well.

>>106460460
I don't like WD really, just personal preference but their promotional materials are honest about tech specs.

Just buy Toshiba MG drives like a normal person?

And who the fuck buys 4tb drives anyway, you can just buy SSDs at that size.

>>106460227
>>106460296
8tb wd is garbage unless you bought an old used helium drive. they transitioned from that to an air drive for their 8tb and 10tb lineup and it's one of the worst drives ever, extremely loud, and gets very hot. like 50c idle 70c load without active cooling.

you want to get WD Red Plus for 1-6tb super silent drives, or toshiba MG if you want large drives where noise is not a concern. anything else is garbage.

>>106460745
Damn I fucked it up then but it's for 4U rackmount server in basement so I don't care about noise.
High power draw would suck when I did everything to keep overall consumption down, itx celeron mbo, pico psu and so on.

>>106460765
hdds guzzle power, there's no way around that, you can alleviate the issue by enabling spindown but if you do zfs it'll spin the drives up to do integrity checks whenever it wants (which takes over a day on 20tb CMR drives).

my simple file server with 6 helium drives uses like 80W of power and of that the mobo/network is only maybe 30w or less. if the drives are not spinning, it's only 40w. i plan on getting an odroid instead, it can idle at 5w without disks.

>>106460745
>WD Red Plus drives are engineered to use less power (versus previous models) and run cooler, which reduces operating costs and helps reduce heat in thermally challenged NAS boxes.
Tho man you're saying stuff that is direct opposite of what prom materials says?
>>106460792
>hdds guzzle power
Will post results for that 8tb wd if I get it today. I don't expect more than 4w idle spun up to be honest.
My proxmox "server" which is really box of ewaste idles at 12w - 1tb 2.5hdd some slim "eco" series, j1900 mbo, 1x120, 2x80, 1x60 fans, aliexpress picopsu clone, small 19v power brick from LG monitor. We'll see how much 8tb will add to it.

fuck seagate. Wd just werks

>>106460792
>my simple file server with 6 helium drives
What are you people storing to need that much space? I'm genuinely curious? My whole life fits in 1TB or so.

>>106460886
I love Seagate as brand but their drives do feel sketchy lately.
Have two identical drives in office machine, both developed bad sectors but continued working.
They're over 100k hours, disregard counter in crystal, seagate rolls over after high 70k's.

>>106459760 (OP)
Still better than shitgate

>>106459760 (OP)
Honestly OP IDK why people still fuck with sub-10TB HDDs much less consumer HDDs.
It's even more confusing because you are aware of the fuckyness in that sector of the market but yet you still participate.

I shouldn't worry about smr if I use a copy-on-write filesystem like btrfs, right?

I only buy the WD Red Pro and nothing below 16TB

Why are they even selling 1TB drives anymore?
It's not 2015.
We should have 100TB HDDs at very affordable price right now. What's going on with technology?

>>106460336

I'm afraid that people are too ignorant and indiferent, plus all these big youtubers are scammers who get paid by such companies for advertising shitty products and make money from ads. I'm shocked that this topic isn't even brought up yet apart from some obscure russian tech forums where I found this info.

>>106460337

I also forgot to tell that it's much harder to recover data from SMR drives compared to CMRs. And SMR drives technologically work exactly like SSDs. It's basically a SSD with magnetic platters minus the speed/multithreading.

>>106460460

Yes WD also lies about the rpm class, many "5400" rpm WD drives actually spin at 7200. Absolute scum move.

>>106460731

>SSDs

Not suitable for long term storage, plus SSD need refreshing at least once per year to keep your data safe. Some years ago it wasn't a problem to find a good 4-6TB non-enterprise hdd, now it's nearly impossible.

>>106461842

Because such tendency will also affect larger enterprise drives. For instance SMR technology gives 25-30% extra storage per platter which is a significant increase for 20TB drives. It all starts with small and seeming insignificant things. SMR plague will completely displace CMR which will become a forgotten lost technology.

>>106461979
>Because such tendency will also affect larger enterprise drives.
Not in the way you think
There are zero enterprise drives where if you just put them in a system and use them that you'll suffer from SMR
SMR in enterprise is limited to host managed SMR which requires the OS to explicitly zone the HDD for it to record data with SMR.

>>106459760 (OP)
>buying anything but Blacks
See, now that's just asking for trouble.

>>106461955
Single platter hard drives, lighter and fewer points of failure.

>>106459760 (OP)
I made one thread abot SMR vs CMR and the dick riding took off

>>106459778
What abot these types?

>>106460899
Install game to SSD.
Finish it but game is bready good
Offload it to spinning rust so I don’t ever need to download it again.

Have my own cloud backup. Everything hits my NAS. Use 20TB Toshiba spinning rust to backup my important files from the NAS.

I honestly can’t understand people that run a single SSD in their systems and it’s shockingly common.

>>106459760 (OP)
OP, is TRIM a bad thing. If it is, why?

>>106459778
So OP is it safe to get 8TB WD reds? Or do they have to be plus/pro?

If you're buying WD, you buy Red Plus or better.
If you're buying Seagate, you buy Ironwolf Pro or better
It's that simple.
>>106460943
>barracuda
found your issue.
>>106462288
pic related

>>106462869
So, I have six of those types of HDDs from WD. ALL 12TB that I Tyrone and will continue to Tyrone as long as I'm and to Tyrone them, but, they seem to be a line of HDDs made, but no definitive naming. Just a white and black disk that's actually CMR and they work pretty well.

>>106463270
it will have a model number you can search to find out what type it is

>>106463270
able*

>>106459760 (OP)
You seem to be knowledgeable about hard drives. I want to ask you something. Which HDDs above 1TB are the best bank for your buck? I'm asking because I'm considering buying a hard drive but im not sure about which one to actually get. A short list of the best hard drives for the money would be greatly appreciated. Thanks in advance!

>>106459760 (OP)
Tell me one (1) (四) good reason to care about how my harddisk physically saves my data to its disk.

>russian
hide

>>106463277
they seem to be
>WD120EDBZ - Western Digital 12TB 5400RPM SATA 6Gb/s 3.5-in Hard Drive
>Western Digital WD120EDBZ 12TB 5400RPM SATA 6Gb/s 256M Cache 3.5-Inch Hard Drive

>Brand: Western Digital

>Part Number: WD120EDBZ

These work very, very well

>>106462288

Some large helium drive, most likely it's CMR. I personally don't trust helium drives.

>>106462541

It's sure indicator of a SMR drive, it lets the drive know which blocks are free to erase and reuse thus speeding up the process, logically (not physically since it has heads) SMR drives work just like SSDs in terms of reading and writing but it lacks the speed and cannot perform several processes simultaneously, that's why it's slow as hell.

CMR doesn't benefit from TRIM at all,

TRIM also might accidentally erase the existing data a will during some canning/recovering operations (such as chkdsk)

>>106462708

Yes, the only drawback is that you can't turn off the 5 min head parking timer (unless using software that accesses the drive before reaching the 5 min cycle).

>>106463326

WD10PURZ, WD10EZEX, or WD10EFRX or WD1003FZEX (if the last two are still available). These are known to last at least 10 years if used properly.

Make sure the drive has no more than 64MB cache size (some enterprise drive might have 128MB cache), any of the enterprise drives new or used (HGST, WD Gold, Toshiba MG, Exos 1TB models shouldn't run hot since they have only 1 platter, fewer heads and should be more reliable based on their 5 year warranty).

>>106463332

CMR is the most bulletproof and reliable way and can easily last for 10-15 years. SMR is a cost cutting technological clusterfuck which is worse at every aspect, I already explained above,

>>106463326

>Which HDDs above 1TB are the best bank for your buck?

Sorry I missed this part. I'll copy the model names I already listed in the second. WD20EARZ, WD20EFPX, WD23PURZ, skip the 4-6TB variants unless they're enterprise class, then WD85PURZ, WD80EFPX.

I'm not knowledgable about Seagate and Toshiba models. But the smaller ones with 64MB cache size or 8TB models (except for Barracuda) should work just as fine.

>>106463778
>CMR is the most bulletproof and reliable way and can easily last for 10-15 years. SMR is a cost cutting technological clusterfuck which is worse at every aspect, I already explained above,
SMR is just overlapping by 10-30% the data tracks like tiles of a roof because the magnetic medium can hold data more densely than magnetic heads are able to resolute during writing. It only have a order/block problem that forces you to write more, similar to flash. Other than that it changes nothing.
And the future will be energy assisted writing that removes the writing problem of SMR and increase density beyond the limit of the magnetic head.

>>106459778
>mfw my wd43purz which i ordered last sunday just arrived
fuck sake OP you shouldve made this thread a week earlier

this stuff is a headache I basically only buy 20TB+ Ultrastars or Toshiba MG also 20TB+ sizes

I've only had 2 drives die on me 1 4TB Seagate and 1 4TB WD. SMR CMR I dunno but I stopped buying 4TB after that from anyone.

>>106463913
haha, damn, a fellow nerd was telling me about his research in roofing hd's at a party like 15 years ago

>>106464163
>I've only had 2 drives die on me 1 4TB Seagate and 1 4TB WD.
what models were they and how long did they last?

>>106463913

That's only the tip of the iceberg when it comes to explaining how a SMR HDD works.

>Other than that it changes nothing.

No, that's the point, SMR tech works drastically differently from CMR in terms of writing, not to mention that it's a zoned device with a second-level translator (not found in any CMR which makes them simple and reliable for many decades).

Unfortunately there are no lectures and tests made in english because western manufacturers and even tech youtubers are driven by greed, profit and rely on viral marketing, false advertising. Here are three videos from two professional ukrainian/russian data recovery specialists who test SMR drives with professional hardware such as PC3000 and give exhaustive explanation about why a SMR technology is the worst thing that was ever invented in terms of data storage. It's a deliberate cost cutting measure and nothing else.

https://www.youtube.com/watch?v=67arwgtTyxA&ab_channel=Vitaliy%27Rozik%27Roziznany

https://www.youtube.com/watch?v=2iZaIivWHkI&ab_channel=R.LAB

https://www.youtube.com/watch?v=-oohWiuO4Jw&ab_channel=R.LAB

I watched all of them and since then I know why I should completely avoid SMR and spread the word about its harmfulness.

>>106464116

It should work fine as data storage. It's not a pure SMR, rather pseudo-CMR with 3 times more buffer zones and TRIM, unlike real SMR it should handle small block writes much better without horrible speed dips (to >1MB/s).

>>106464457
>it's a zoned device with a second-level translator
doesn't explain what any of this means.

>>106464457
do we have a confirmed list of years when wd started to shift these drives to be smr instead of cmr?
>It's not a pure SMR, rather pseudo-CMR with 3 times more buffer zones and TRIM
regarding this, my wd43purz has trim like you mentioned and the manufacturing date on the label says 2023. although on second thought it might be the case that the entire production line under the same serial regardless of age is already on SMR tech as opposed to the claimed CMR. how did this fly for so long under peoples noses?

>>106464498

Ok, I'll try to explain but you probably won't understand a word.

1) Logical zones in a SMR drive have nothing to do with physical zones on the platter itself, they're always dynamic even at idle unlike on a CMR drive (CMR is completely static and predictable in this regard).

SMR automatically fragments files even when you just read them (yes that's how ridiculous they are). That's why a SMR is complete clusterfuck and it's nearly impossible to degragment them, any extra reading and adding files actually worsens fragmentation of existing files. And that happens to a healthy SMR drive, if the drive has problems then it's 10x times as hard to recover data from it.

2) CMR drive have only one translation system (physical addressing sector-track-head to logical addressing LBA), SMR disks have two translation systems (further complication). These are the classic "sector-track-head in LBA" translator and the new "sector-track-head in track" translator, and both of these translators are interconnected. Losing any of them will result in complete data loss (by the way, this is what the "fast erase" technologies of SMR disks are based on - we reset one of the translators and that's it, there is no data). Recovery will be possible only if the lost translator can be restored. This is already a task for data recovery companies, and at the moment it is quite complex and expensive.

>>106464546

I don't know people haven't noticed it. This WD managed to mask SMR much better than when they secretly released their first SMR Reds when people complained
about Reds shitting themselves in RAID.

can't you just disassemble one and look how it's made? surely SMR and CMR will look different

>>106460464
>CMR drives with 2TB per platter density
Do you know if it's safe to assume any drive 2TB or smaller is always gonna be CMR?

>>106464828
>>>SMR automatically fragments files even when you just read them
what the fuck i was unaware of this

>R*ssian schizoid thread
OP, it's time to join the rest of your friends. Let's see that tripcode.

https://www.youtube.com/watch?v=HWkz0wvop9w

oh god what now

>>106464886

No. I'm not a data recovery specialist and any DIY may make things even worse, disassembly requires a sterile room with zero dust particles, moisture or any foreign object which is impossible to achieve at home.

>>106466159

No, quite many 2TB drives and less are SMR. Especially 2,5 inch (all of them are SMR). 1-4TB CMRs should have no more than 64MB cache (apart from some enterprise models). CMR doesn't even need cache, it will work just as fine with 8MB buffer. Also they shouldn't have a TRIM command.

>>106466365

Yes, that's what I'm talking about. Any attempts to work with a SMR drive fragments the files even more, regardless if you just read the files or write new ones. Already posted the video from one ukrainian data recovery specialist who shows how does this a WD10SPZX SMR behave during the read tests.

https://www.youtube.com/watch?v=67arwgtTyxA

This is why SMRs fail in RAID, the files become so fragmented that it's becoming impossible to "assemble" them while accessing, the drive will either drop its speed to >1MB/s or even completely freeze (sometimes corrupting the affected tracks) which will never happen to a CMR drive.

The funniest thing is that defragmentation doesn't even work with SMRs, SMRs require to be treated like SSDs (backing up the data, refreshing the sectors and then copying back the data which will make them work faster). SMR is basically a flash storage in terms of reading/writing but is more suitable for cold storage due to being magnetic unlike SSD which can lose charge and data when not accessed for too long.

>>106467200

It's up to you to trust me or not. I'm simply pointing out and providing evidence about how WD and other hard drive manufacturers are scamming you. I'm surprised that I'm the only one who's bothering to speak about it. But it seems that people are totally ok with getting fucked over. So be it.

This will be my last post because according to people here telling the truth means that you're a schizo.

>>106469799
>SMR is basically a flash storage in terms of reading/writing but is more suitable for cold storage due to being magnetic unlike SSD which can lose charge and data when not accessed for too long.
aren't translation layers on SMR drives stored in flash?

>>106459760 (OP)
Thanks OP, I've had WD drives fail on me all the time.

They're an aging company that probably earns most of their money from legacy contracts. Even if their new products tank or succeed, it won't affect their bottom line. Many hardware companies are like this too, like Intel.

>>106469873

Yes but the actual data is stored on the magnetic platter. The problem is that the logical adress (processed by the 2nd lvl translator) is not aligned with the physical adress (platter), so the physical writing process is not done linearly despite what the manufacturers are telling you hence the further fragmentation and other complications of using a SMR drive.

While CMR is completely static in this regard (written tracks are simply set in stone unless you interfere) and has no several background processes made by the 2nd layer translator that simply cannot calm down and always does something.

SMR should have never becoming a thing, fuck Seagate for ever inventing it. If SMRs would offer double the storage (in reality it's only 30% max which is pathetic and not even cost effective) for the half of price (along with a sticker that says "FOR COLD STORAGE ONLY" with big red letters), then this wouldn't be a problem.

>>106469972

No, problem. Yes, the age of reliable hard drives is over. All hdd manufacturers are similarly devious because they know that they can get away with it. This happened right after smaller hdd brands such as Samsung, Hitachi, Fujitsu, Maxtor got sabotaged and aquired by bigger companies. Even current Samsung SSDs which used to be bulletproof are getting sabotaged by WD, because american companies hate fair competition.

>>106470035
if the translation layer is stored in flash then all the cold storage benefits from magnetic recording are gone, they would need to be powered at least once a year just like ssd's

>>106462099
>buying anything but Blacks
>See, now that's just asking for trouble.

>>106470068

Well, not exactly, because magnetic storage doesn't care about passing time even if it's shingled and it still has physical heads, magnetic platters, the translator has the info about media cache map, logical zones, used/unused tracks, trimming function which doesn't has to do anything with the written data on the platter. So it should be suitable for cold storage.

Here's the article.
https://www.radianmemory.com/about/news/smr-meet-flash/

But it behaves just like SSD despite being mechanical and being limited to execute only one operation at time hence all the slowdowns, freezes, fragmentations etc.

Also, just turning on SSD regularly is not enough, SSD requires refreshing every single of its block (preferably after making a backup) to prevent non-rewritten cells lose the charge and thus become slow or even become empty. It applies to SSDs, flash sticks, SDs as well.

https://www.puransoftware.com/DiskFresh.html

just bought and shucked a 24tb WD HDD, what is the best tool to fully test it before starting to use it for archival process?

>>106470158
im not sure personally but ive been using a pirated copy of hdd sentinel on my new drive, first i did an extended smart self-test now im waiting for the full write + read surface test to finish.

>>106470419
thanks i will give a try

>>106463778
>>106463853
Thanks guys! I appreciate all the info I can get. If anybody else has any recommendations, I'd love to hear them all!

>>106460065
fucking retard.

Is there a way to parse the model name?
I bought two external drives and shucked them to put inside my pc, both 6TB WD60EDAZ-11U

>>106469799
>This will be my last post because according to people here telling the truth means that you're a schizo.
These are well-poisoning shills, don't listen to them.

>>106459760 (OP)
So just buy toshiba N300 (nas), X300 (performance), MG11 (enterprise) and im fine? These lines were always CMR.. hope they did not sneaked some SMR or semi SMR shit into them


>Toshiba does not maintain an exhaustive public CMR list, but models not listed under SMR (e.g., N300 series) are confirmed CMR. Enterprise lines like MG series are typically CMR unless specified otherwise.

>>106470419
all my hdds that died it always happened the heads started clicking and it was over. I think no smart test or surface test would detect the heads will off themselves in next few months

>>106473368
stress testing a drive before copying all my shit is good enough for me. bathtub curve and all that

Just in case let me show you a difference between a pseudo-CMR (modern Purple) and a real CMR (older Red). See these dips in the upper graph? This is a sign of a 2nd layer translator activity which is used in SMR drives, it's static, but still exposes itself during small block writes. The lower graph looks much smoother.

>>106473323

Not sure about toshibas and the fact that the NAS/survelliance category which is designed for performance contains both SMR/CMR (even if they point that out unline WD) is rather concerning. But the MG ones are rock solid, if I'm not mistaken some of them are based on native Hitachi drives (not HGST). Also it's funny to see 512mb cache in some MG drives because in real life it's no different from 8mb cache, it doesn't improve access time or small block write speed.

Another thing that bothers me is that It's just becoming impossible to find a 4-6TB 5400rpm CMR HDD, but oddly there are still decent 1-2TB and 8TB variants and nothing good in between. Back then there were great choices such as WD30-60EFRX, WD30-60PURZ, but for some reason they decided to stop producing such drives and intruduced a new disk family called Venice (that infamous first SMR Red was based on it),

Instead of learning their lesson after getting caught they decided to double down and introduce a new type of advanced SMR hybrid which has special buffer zones that emulate a CMR behavior, a hidden 2nd lvl translator with a static module instead of just making a classic CMR drive. That's why these new 4-6TB WD Red Plus/Purple drives have 256mb cache (Purple even has TRIM), these are not CMR despite what the datasheet says.

>same retard from april is now trying again
>his only evidence continues to be screenshots of a russian forum that don't even prove the point he's claiming

>>106474198
where is your counterproof?
he does explain the graphs and where to look, so you should be able to easily disprove them if it bullshit, no?

>>106474198

Because I learn from other's mistakes instead of making mistakes (like stumbling upon SMR Reds or current pseudo-SMR Reds/Purples) and I want to help other people and pay attention on how WD is still deceiving its customers to this day.

I did test my own 2TB CMR Red/Blue/Purple (after watching how a data recovery specialist tests one and comfirms that it's an actual CMR).

Here's the pic of my old pc with 3 of these 2TB installed (2 active clones and another clone for cold storage). I also bough a pair of 8TB Red Pluses (although 8TT Purple'd be better due to disabled head parking), one 8TB Ultrastar which are installed in another PC (I can show them later if necessary).

tl;dr

>>106469799
>This will be my last post because according to people here telling the truth means that you're a schizo.
Pls don't get discouraged because of retards like that guy. You have provided me and some other people on this thread a lot of important information! Thank you, anon! Pls continue doing what you are doing.

>>106473323
>>106474196
I'm also interested in your recommendations for HDDs that aren't WD. What about Toshiba and Seagate? I'm sure they have some actual CMR drives themselves. What are your recommendations there, OP? Maybe they have 4-6TB models?

Can someone explain this like im 5? what does crm and smr and trim mean?

>>106477661
>cmr
good fast reliable
>smr
bad slow unreliable
>trim
red herring

Is there any actual reliable way of getting CMR drives, or is it still just searching forums and praying?

>>106459778
What is head parking and why is it a problem?

>>106477843
Buy enterprise drives

>>106478271
Already do. Just more wanted to know if this was creeping in to enterprise drives (yet).

>>106478234
head parking is when the needle is "parked" onto a safe location off the platters
if you're using an external drive that's prone to movement then it's a safety feature, in the case of it falling the needle will be cushioned and it won't scratch the platters
but if you're a high-load consumer then the hdd is needlessly and constantly moving the needle off the platter and then on again, increasing wear and degrading it much faster

>>106478370
And only these purple drives don't do it? So I need surveillance camera drives for my nas to store movies and shows?

A bit off topic but I've noticed that all these nas systems have their drives set up vertically. But all of my cases since the dawn of time have drive bays horizontal. Wouldn't having your drives vertical make them have to fight against gravity to move the needle and fall apart quicker? If I get one of these should I turn it on it's side?

>>106478408
you're not under constant load just storing shit, and even if you are streaming from it you won't be slighted that much, and even if you do constantly stream shit I still think it's better to have a caching disk than to buy security camera drives

>>106478432
servers often have them oriented this way, as long as they are perfectly level it doesn't matter if they are horizontal or vertical, but if you have them like dangling in a pc case at a weird angle they will absolutely degrade faster. at least that's my understanding.

>>106478432
All the HDDs I had that died were in horizontal position.

>>106460899
I use my computer for work, plus I wanted to build a backup box for my stepdads stuff (something like 15 hdds going back to 2008).

technically if I didn't want to keep an archive of my stepdads works and I didn't need to keep my movies drive, then I wouldn't need any of the helium drives.

>>106478793
There's no such thing as "work" on a computer

people are still buying single digit capacity drives?
in the current year?

>>106461955
>We should have 100TB HDDs at very affordable price right now. What's going on with technology?

Perpendicular magnetic recording has hit a density wall and we had to figure out how to use more magnetically robust platters, so we can make them denser. A side effect of this is we needed completely new write heads (that can heat up sectors to drop their magnetic resistance while writing, then the cooled down material will retain data better). Existing write heads ended up getting deformed from the heat/power required to do this.

This took nearly a decade to figure all out. But it's done now, and already allowed us to jump from ~20TB to 30TB in the last few years, with Seagate planning to hit 50TB "soon".

>>106461979
>Not suitable for long term storage, plus SSD need refreshing at least once per year to keep your data safe.

Horse shit, all my SSD drives lasted 5+ years, been using them since the first 40GB Intel drives with the abysmal write speeds.

>>106478854
What is going to happen with the prices?
>will a 20TB be 2x more expensive as a 10TB one?
>will a 30TB be 3x more expensive as a 10TB one?
>will a 40TB be 4x more expensive as a 10TB one?
>will a 50TB be 5x more expensive as a 10TB one?
The price/TB has to go down by at least 2x-3x, so much data and barely any significant read/write speed improvements become a liability.

>>106473323
Just use Enterprise drive, period. They all have datasheets that list exactly what they are, since they'd get fucked hard if they sent different drives to datacenters. Karen buying a $30 2tb external backup drive won't give a fuck about CMR/SMR period, but a datacenter who buys tens of thousands of $500 drives will have the funds to fuck you up in court if you mislabel CMR/SMR/4kn/512e/whatever.

see picrel for Toshiba MG series.
pretty easy to figure out which drives are SMR and which ones are not.

>>106478909
I could buy 20TB drives last year for the same cost as 12TB drives in 2020. Prices are going down, slowly.

Keep in mind the newest largest drives are enterprise only first, then they slowly trickle down to consumer product lines as they slowly shift production. So it will take a while, but I guarantee you that a 20TB drive will cost the same after some time, as a 10TB one costs today. I just can't tell you how much time it will take.

>only hdds I've had just up and fail ever in my life without me causing it being a gorilla were some 2tb smr (literally fucking why you cheap ass fucks) wd reds from a couple years ago from their 1st false advertising bait and switch scheme
>and 2 hgst 7k1000s sometime a decade ago
>Still have 10-15 year old wds seagates and toshibas both 2.5 and 3.5 still working perfectly to this literal very moment

>>106476598

Ok let me take a look. When it comes to Seagate the tendency looks similar. For instance picrelated 6TB Ironwolf is visually reminiscent of SMR Barracuda and has 256mb cache (while the previous Ironwolves had 64mb). Highly likely they're introducing the same pseudo-CMR/Hybrid SMR technology

But I found a few 4TB 5900rpm Ironwolf and Skyhawk HDDs with 64mb cache, they should be safe choices:
ST4000VN008
ST4000VX007

No luck with 6TBs though.

Toshibas seem to be more interesting. I took a look and besides these SMR S300s other choices such as N300/X300 seem to be solid but I'm not sure. Also all Toshibas spin at 7200 so they will be hot regardless. So I'd recommend their best offer - MG series due to 5 year warranty.

Just bear in mind that WD, Seagate and Toshiba work in unison and imitate each other and cut costs similarly.

>>106477843

Nowadays - no. Only by reading forum topics, reviews, watching tests. Manufacturers simply like to torment their customers and treat them like guinea pigs. The choice of consumer level HDDs is shrinking rapidly. I think ordinary air filled 3,5 CMRs will no longer be produced in 5 years or so (just like 2,5 CMRs are no longer a thing). The only remaining choice would be expensive 10TB+ helium drives.

>all my SSD drives lasted 5+ years

Because older flash storage (including flash sticks that could last for years) was more durable and reliable. Old SSDs were SLC and MLC. Modern SSDs are TLC at best, even Samsungs and Crucials. The quality of flash storage has dropped drastically in recent years.

So yeah even SSDs are becoming shit. I guess they're also trying doing it to go after piracy and datahoarders by deliberaly lowering the reliability of storage devices apart from being greedy lazy cunts.

>>106479051
>I guess they're also trying doing it to go after piracy and datahoarders
They are a drop in the bucket compared to enterprise clients and they most likely buy used enterprise drives at half price too.
HDD manufacturers would unironically make more money with lower prices aimed at datahoarders.

>>106479051
>Just bear in mind that WD, Seagate and Toshiba work in unison and imitate each other and cut costs similarly.

That's probably why Seagate is far ahead in size and cost, and Toshiba is ahead of cost compared to WD.

>>106479051
>Nowadays - no. Only by reading forum topics, reviews, watching tests.

Or you can not be a retard and read the enterprise drive datasheets which tell what a drive uses. Hell even consumer data sheets list them. If you are buying slav tier cheap junk that doesn't have CMR/SMR listed in the datasheet, that's your fucking problem, stop buying cheap junk and expecting not cheap junk.

>Because older flash storage (including flash sticks that could last for years) was more durable and reliable. Old SSDs were SLC and MLC. Modern SSDs are TLC at best, even Samsungs and Crucials. The quality of flash storage has dropped drastically in recent years.

I've been using TLC drives for 10+ years and they are all fine. Again, stop buying cheap crap and buy high quality drives, and you won't end up with junk that dies after 200TB written like the Kioxia drives.

>>106476560

Thank you. I had a feeling in my gut that there's something wrong with modern HDDs, but couldn't put my finger on it.

These obscure russian tech forums such as ixbt, 4pda and ru-board along with amazing articles from russian data recovery companies (R.LAB and ACELAB) and some lesser known russian/ukrainian data recovery specialists (Vasiliy Dorin, Vitali Rozik) on youtube helped me a lot to understand about hard drives in general. They post free information, show their hdd tests, data recovery process and give overrall recomendations.

Some of them are legendary (for instance one guy from RLAB company managed to reverse engineer an ancient fujitsu hard drive and create his own, turn a SMR into CMR and did many other crazy experiments).

Picrel is my favorite one. He has 25 years of experience but he doesn't care about promoting himself and only has 2k subsribers apart from his telegram channel.

Unfortunately I couldn't find anything even remotely close in english speaking web resources. Probably because it's a crime to investigate such things in the West. Russians are very curious and inventive in this regard. I wanna see that faggot Linus or other so-called tech youtubers show something similar apart who can only make clickbait videos, shill products and make money from ads.

>>106479156

>If you are buying slav tier cheap junk that doesn't have CMR/SMR listed in the datasheet, that's your fucking problem, stop buying cheap junk and expecting not cheap junk.

No, datasheets don't even matter at this point. I already posted an example of 4-6TB WD Red Plus/Purple drives which are listed as CMR but in reality they are Hybrid SMR. These aren't cheap. And WD already fucked their reputation after that SMR Red fiasco.

>I've been using TLC drives for 10+ years and they are all fine.

10 years is a long time and so much has changed. Even reputable companies such as Samsung and Crucial are no longer what they used to be. I wanted to buy a 870 EVO but found out about their mass failures so no, Crucial SSDs are also failing a lot lately - I will never ever look at SSDs, I use a pair of cloned CMR HDDs as system drives.

And SSDs do indeed require refreshing along with incremental backups, they're no intended to keep the static data for too long.

>>106479193
>I wanted to buy a 870 EVO but found out about their mass failures so no
oh, I remember that, i was looking at buying my first nvme drive and found that out, digging further i found similar issues from other manufacturers, the real problem is that this issues are discovered years later by enthusiasts and barely addressed by manufacturers (iirc it took 3 years for the issues to be confirmed and a firmware update to address the issue), and this issues are usually popularized by techtubers with well known conflicts of interest like ltt
i ended up with a drive with a innogrit controller that a few weeks later i found out was known to fail with ymtc nand, fortunately my drive has bics nand, hopefully no one will find down the line this drives fail too

>>106479452
I went down a rabbit hole today reading about how AMD chipsets have such serious flaws with their USB controllers that you can't even reliably write data to USB storage drives on them. We're in the year of our lord 2025 and AMD can't even get USB working. God help us all. All technology is is retards shipping broken products, ignoring flaws, and telling consooooomers to just buy the newer version and maybe it's fixed this time.

I've only been buying WD golds for 10 years now, my last non gold drive is a 4tb red, forgot which model but it had about 9 years of power on hours when I got no use for it anymore.
Hopefully those won't become cheap garbage any time soon.

>>106479533
>AMD chipsets have such serious flaws with their USB controllers that you can't even reliably write data to USB storage drives on them.
I have never read any article about that, but from my own experience had already realized that this is true. Whenever I write data to USB drives on AMD platforms I have to triple check their integrity and sometimes still must rewrite the data. It's a pain in the ass.
One of the reasons I stay an Intelchad for my main daily drive PCs. AMD CPUs might be better but Intel platforms are simply more reliable and stable.

What about the red pro hdd's are they okay?

>>106460093
How do you check if they are real cmr drives?

>>106480179

By looking at these specs: on ordinary 3,5 inch 1-6TB drives cache should be no more than 64mb (rarely 128mb but not 256mb) (on enterprise drives like on Ultrastars/Golds cache might be larger like 128mb-512mb), also it should have no TRIM.

Look at the read/write graphs of CMR WD85PURZ and compare it to Hybrid SMR WD43PURZ (I posted the pics it in the first two posts ).

But the best way to find out is to test it (or look at how someone's testing it before buying it).

Download the program called r.tester

https://rlab.ru/tools/rtester.html

See picrel for guide. Click on the disk you want to test. Also r.tester will automatically show if it's SMR by detecting its family. Click on Toolbox - CERT Tool Lite - Load Script - choose "SMR_detection_3w - START.

It will write huge amount of 1mb and >1mb files which will cause a SMR drive to drop its speed drastically and then freeze, while a real CMR will easily swallow every file and pass the test. Make sure it has no file or back it up before doing it.

>>106480164

They're no different from wd ultrastars/golds and have that shitty wd firmware and 5 min head parking that cannot be turned off. Better buy ultrastar or gold. Just take care of cooling because they run hotter than 5400rpm drives and can easily reach 40 or even 50C+

>>106479452

Yes manufactorers don't care about quality control or optimization at all, they just want to sell more junk that barely works. They would rather sweep it under the rug and memoryhole it because the average buyer is dumb enough to step on the same rake over and over again instead of protecting his consumer rights and sue these greedy lazy fucks who can't even do their job right.

What's the estimated lifespan/wear and tear on the HDDs I use strictly for my Jellyfin server?

>>106479452
I have a 870 Evo that has the efuse blown, making the drive unable to power up. I just desoldered the efuse and bridge the traces, it's still working ever since.

>>106479533
>>106480072
Are you sure you weren't just using garbage cables? The last time I had problems with USB3 was on sandy bridge. On my B550M I could have 4 hard drives hooked up for a 3 day stress test via USB and they all worked fine. The only issue was that I had to hook up two to the back and two to the front panel, because all the ports on the back must've shared bandwidth, I could only do ~700MB/s instead of 8-900MB/s if all four drives were on the back of the board.

>>106482406
>>106480072
Here's a post detailing the woes of USB on AMD https://lemmy.ca/post/18336941
>Even though resilver finished successfully, there were silent USB resets in the logs with the OWC connected to CPU-provided ports.
Basically the USB ports constantly silently fail and reset themselves and it's only ZFS's insane durability that kept the data intact on this guy's system. I've read other posts saying they can't keep drives plugged in via USB longer than a couple of days before AMD drops them.

>>106478432
Yeah but it means it's not fighting gravity when the needle is on the way down. This is why they recommend turning your NAS upside down every other year to spread out the wear.

>>106482814
That's retarded. We just ship our servers to the southern hemisphere.

>>106482628
>Here's a post detailing the woes of USB on AMD https://lemmy.ca/post/18336941
yeah that guy is doing basically something you shouldn't do. like running a ZFS array over USB and expect it to be stable (he mentions right at the beginning that he had to put heatsinks on the USB-SATA bridge chip to prevent overheating).
He's also using a 5950X on a B350 board, that thing may not even have proper traces or enough VRMs to handle that chip, plus it's 3 generations later and wasn't even meant to be backwards compatible.

It's not an AMD problem. I've ran 4x external drives simultaneously over USB for a 2-3 day continuous read/write test with a B550 board and a 3600, and it had zero issues (other than speed).
The guy is just a retard who expects 50$ garbage setups to have enterprise resiliency. It's the slav bullshit as what OP does, expecting lowest end garbage drives to not be garbage.

>>106479533
not sure if its truly an AMD thing but ive experienced that a lot with my PC. i got an nvme usb enclosure (i didnt know about pcie adapter boards at that point) for an ssd i had laying around and it kept disconnecting itself during writes, nevermind how speeds didnt even reach 100MB/s. all this time ive just blamed it on the cheapshit B450 assrock board i got.

>>106483347
>He's also using a 5950X on a B350 board, that thing may not even have proper traces or enough VRMs to handle that chip
Nigga what?
You speak like some pcbg autist
The chip shouldn't matter for USB stability, the CPU is also going to be power limited by the capabilities of the board but should not be unstable.
Unless this is another one of those magical AMD moments where AMD allows board vendors to let whatever chip be inserted no matter how broken the experience may be because muh am4 longevity promise.

>>106479157
Louis Rossman, from the handful of videos I've seen, has mentioned how much of a scam the data recovery services are in the US. I believe he has done some surface level at least data work.

There's obviously people out there with the know how, but I'm not sure if there's enough interest from the public at large for anyone to do content on it.

I suppose it's a cultural thing. I've always heard (on average across every person) Russian computer scientists were more willing to get down to the nuts and bolts tinkering on hardware and explorting theory in depth more than the westerners due to the lack of available compute time at universities and the lack of processing power that goes along with that.

You had to make the most of what you had and be very knowledgeable. I assume that translated over to the younger generation that followed in the 80s and 90s. Then the Yeltsin economy once again forced people to get by with what they had; more simply said, using your shitty PC and whatever's laying around.

>>106484451
>The chip shouldn't matter for USB stability
Signal integrity follows from power integrity

>>106483347
Ok but it just werks on intel

>>106483766
I keep hearing about this a lot but I have never experienced this issue myself, my two machines have 5600x 7600x and each have two external usb drives attached to them that I use a lot.

>>106479157
>Unfortunately I couldn't find anything even remotely close in english speaking web resources.
You should check this guy out:
https://youtu.be/hAijmZsTd2A
You can tell me if it's something similar to the Russian content that you were talking about.

>>106479193
>I use a pair of cloned CMR HDDs as system drives
Which models? Do you recall why you chose them?

>>106486311

Are you kidding? This is the most shallow explanation I've ever watched.

While these russian/ukrainian dudes are literally explaining how technologically (both logically and physically a SMR actually works (one of the is using PC3000 which costs like 8000$). They even list disk families (no not sticker colors and datasheets that mean nothing) that use SMR. But you won't understand a word even with a translation.

https://www.youtube.com/watch?v=67arwgtTyxA
https://www.youtube.com/watch?v=-oohWiuO4Jw
https://www.youtube.com/watch?v=2iZaIivWHkI
https://www.youtube.com/watch?v=CZSZ8I5pg3U

They tell about media cache, 2nd layer translator, usage of two partly overlapped read heads, mismatch of logical and physical sectors, switching between the buffer zones, in-built defragmentator that fragments files for no reason, the dangers of file destruction during track reconfiguration. They literally prooved that SMR is just a SSD program code shoehorned into HDD (in the most shitty way possible that barely works), why SMR is an evil deviced created by profiteers and marketeers instead of enginers and programmers and people should totally not buy any SMR. They also tell that CMRs will be nearly totally replaced by SMRs to cut costs (it's already happening how).

>>106487610

HUS722T2TALA604. They're bulletproof oldschool hitachis and are known to last like 100k hours and work non-stop easily. But they run a bit hotter so you need a well cooled PC.

>>106459760 (OP)
Going to buy a replacement 2.5 inch drive for one of my older laptops. Are there any you'd recommend OP, or are they all bad? Should I just go with Toshiba or something? I don't trust Seagate.

>>106484456

Thanks I'll read about him. Yes russians (and many other slavic folks) are forced manage available resources and optimize them, make sure that exploitable things last as long as possible, repair it whenever possible even if it's a cheap junk (especially applies to old soviet/early russian cars and trucks that can be fixed easily and passed to your grandchildren), they approarch things in a militaristic/utilitarian kind of way (like these sturdy commieblocks with central heating that are easily repairable and can even withstand shelling), treat things with care, stitch holes in old clothes, they never throw away old outdated stuff and utilize it beyond its limits (sometimes even worship it if it's some 90s Mercedes or some old japanese electronic device), they're kinda similar to japanese people who still use fax machines and web 1.0, unlike americans who are obsessed with consumerism, latest cool trends, reckless spending, debts and leaving nothing behind. They lived through bloody revolutions and collapses which definitely affected their mentality. Russian/soviet scientists are very underappreaciated (soviets did a lot for space tech) because being a sientist in USSR was very highly praised and rewarded (which was even reflected in soviet movies where a role models were either working class men with families or scientists who got hot girls and even the best soviet sports champions earned less), many of them moved to USA during the collapse and even worked in Microsoft, Silicon Valley.

Russians are still working with legacy stuff instead of throwing it away and moving on (for instance currenly they're making Windows 95-98 work with the modern internet and software).

Based, fuck WD

>>106477661
CMR (conventional magnetic recording) is the original way old school spinning hard drives work. SMR (shingled magnetic recording) is a new cost cutting way that a lot of smaller (and increasingly bigger, it seems) hard drives work.

SMR has roughly or equal to the read speads of CMR, but its write speeds are atrocious. This is because SMR drives literally overlap information, like how shingles on a house work. I'm not fully certain of how it works, but as far as I understand that means a CMR drive needs to overwrite good information during its regular write process, leading to complications and slowdowns.

Trimming is the process SSDs use to clean up unused sectors. I guess it's also used for CMR drives, too.

>>106488816

Unfortunately there are no more CMR 2,5 drives. All 2,5 inchers are SMR only (even 500gb wd blacks) CMRs no longer being produced for laptops, only SSDs.

I managed to buy one Toshiba HDWJ110 for a ridiculous price (100$) and I can't find anything else. I also saw a WD Red 2,5 for 200$ but it was too expensive.

But look for used ones if you (but they'll be expensive due to not being produced anymore). Make sure the cache is 8 or 16mb (not 128mb because these are SMRs). Here are the models.

WD10JFCX
WD10JPVX
WD10JUCT
HDWJ110
Hitachi Travelstars
Samsungs that start with HN* (not ST because these are shitty seagates)

>>106459760 (OP)
>>106459778
what 8-10-12tb hdd would you recommend for a pc, simply used as media storage
I was looking at a 10tb wd gold WD103KRYZ since it was cheaper than anything else with good storage but idk if its ok for my use case since it says datacenter

>>106488886

8TB HGST Ultrastar HUS728T8TALE6L4VY is the most reliable largest model so far (I wouldn't trust anything over 8TB or with helium which is another point of failure). Proven to last until 80k hours at least. It's air filled and needs cooling though because there are lot of platters and heads inside which can heat up to 60C+ easily and die earlier than expected.

>10-12tb

Any Ultrastars/Golds (they are the same Hitachis with different firmware).

>>106488881
Damn, that sucks. Which of the SMR ones do you think is the least worst? I suppose I could go for an SSD too, though the laptop is very old and I feel like a decent SSD would just be overkill since every other component would bottleneck it.

>>106460162
>(but it's not gonna behave like a typical SMR (with catastrophically low speeds when out of cache) due to more zones and better zone configuration.
does this still negatively affect zfs?

>>106488985
or just raid in general

>>106488985
Yes, SMR is bad for ZFS

>>106489004
even the ones that op says are close to CMR drives in performance?

>>106488966

NO NO NO! Don't even think about using a SMR as a OS drive, NO! Trust me I had one laptop with a SMR inside and it was a total nightmare with dozens of freezes per days, catastrophic slowdowns, data losses/system file corruptions. SMR sucks even as a file dumpster because if automatically fragments files each time you read them (despite being write once read many).

I'm sorry but such is the reality nowadays. Either search for the used 2,5 models I listed above, buy a desktop and use drives 3,5 1-2TB CMR drives listed in OP (Like WD20EARZ, WD10PURZ or even 1-2TB Ultrastar), or buy a SSD (not the cheapest one, at least something like Samsung although the latest models such as EVO 870 are known for failures). Or even search for some older laptop with a HDD inside (use it or take out the drive).

>>106488985

I don't know, I haven't used them, only real small block write tests will show. They're not as shitty as a real SMR but they're not CMR either, they will definitely perform worse than a genuine CMR.

>>106488933
so the wd gold I was looking for is fine?
again, i just want to save movies and music in there and have it inside my pc without much hassle

>>106460899
Some of us want to have more than like 5 movies to watch our entire lives, I guess.

>>106489073

Yes absolutely, these are designed to endure heavy constant workloads 24/7 for 5 years straight (which I doubt that you will ever do) so in your case they should last even longer. But still, make backups, no matter how good the drive is it might fail suddenly so buy a second one, copy everything to it and put it on a shelf just in case.

>>106460899
Every thread a Sayanim comes in and says: "No, don't use local storage. Let us give you new facts and history on a regular basis."

>>106463853
What are the best 12-24 TB drives and are they better/more reliable than the 8 TB Purple?

>>106489198

They should be, it's just that 8TB Purple is also based on HGST and spins at 5640rpm which is cooler than 7200rpm and has no head parking unlike its Red counterpart. That's why I prefer it.

But the absolute best drive in terms of size/reability ratio would be 8TB Ultrastar (HUS728T8TALE6L4VY), but it runs hot and needs cooling.

I'm not sure but any of the Ultrastars/Golds and even Toshiba MGs should do it as long as you take care of cooling because such amount of platters/heads generate a lot of heat. And testing/backing them up will take very long time.

>>106489148
absolutely ill have like 3 different backups specially for the important shit
thanks for the info anon, appreciate it

>>106489105
ok man I kinda can't watch same movie multiple times unless there's a decade and a half gap in between but I do seed rare torrent which are dying.
>>106489179
I feel you, I also use local storage and what I have on google/le cloud I do weekly backups, always have paranoia of google locking me out of account. Thing is there ain't that much data, like I would not know how to fill up 20tb drive other than downloading movies which are always available as torrents anyways. Then there are guys with 100tb arrays and what not.

>>106489007
I'm not sure

>>106489007

As I said before. They're still SMR, despite being vastly improved compared to early SMRs, they have more buffer zones, better track configuration, static (not dynamic like in the previous SMRs) 2nd layer translator module 190 that lets it behave as if it's CMR, but despite that it's not a CMR (CMR writes in a competely different method and it's 99% physical). It also fragments data at will in the background just like a SSD which doesn't care because unlike a HDD SSD is fast enough.

It still has this dreaded 256mb media cache that can destroy data when collecting it and not being able to store it in the physical adress if something in the firmware goes wrong (even when a hdd is 100% healthy), CMR doesn't rely on cache, it hops data from physical track to another physical track instantly plus it has several emergency modes.

In other words it's much easier to recover data from a failing CMR than a healthy SMR with in-built firmware SSD-esque bugs (all SMRs are full of them). I would never buy a SMR in under any conditions it's technologically flawed and unreliable.

>>106460296
>I can show the video, where one data recovery specialist tests smr drives with PC3000 but it's in russian
please do!

>>106490500

I already did. But here are two links for you personally.

He explains in russian but the english subtitles should let you understand at least 70% due to russian phraseologisms,pronounciations and dialects that cannot be translated normally.

1. How does a SMR fragment its files during reading just like a SSD and becomes even slower.
https://www.youtube.com/watch?v=67arwgtTyxA

2. Here he explains that a SMR behaves exactly like a SMR (works faster after refreshing).
https://www.youtube.com/watch?v=CZSZ8I5pg3U

So in short - SMR is shit that everyone should avoid.

i only buy WD blacked drives

>>106490994
oт дyши бpaт

>>106488716
>HUS722T2TALA604
Did the RPM and cache size factor into your choice?

>>106488716
Well, I was mostly referring to the channel, not the specific video. You can check the guy out and see more than just a video. But at least he's not sponsored and actually does proper testing, unlike most other channels out there.