Jump to content

The three major HDD manufacturers are selling slower drives, without telling us

hitardo
31 minutes ago, Lady Fitzgerald said:

🍿🍿

🍹

 

Need a drink with popcorn, makes ya thirsty.

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, leadeater said:

🍹

 

Need a drink with popcorn, makes ya thirsty.

No worries, I drink like a fish (keep in mind what fish drink). I treat my water bottle like an American Express card; I don't leave home without it. Right now, I could be a two fisted drinker; I have a bottle of ice water (I freeze the water in the bottle) and a can of Caffeine Free Diet Coke on my desk.

Jeannie

 

As long as anyone is oppressed, no one will be safe and free.

One has to be proactive, not reactive, to ensure the safety of one's data so backup your data! And RAID is NOT a backup!

 

Link to comment
Share on other sites

Link to post
Share on other sites

@mr moose

Yeah, I'm sure a different fucking oxide layer halves the fucking write performance. Are you really grasping at this straw? :rolleyes:

Link to comment
Share on other sites

Link to post
Share on other sites

48 minutes ago, RejZoR said:

@mr moose

Yeah, I'm sure a different fucking oxide layer halves the fucking write performance. Are you really grasping at this straw? :rolleyes:

 

Again, you really have to stop thinking like you understanding the problem and consider only what we can establish as a fact.  The point with talking about oxide layers is that we have no idea how much effect they have on performance, just like YOU have no idea how much SMR in modern drives effects performance. 

 

The fact you tried to refute it by claiming you understand the performance gap just goes to show you are not thinking about this from a critical position, you have failed to grasp the fact that we have ZERO data to indicate that SMR is a problem here.  Hence why I pointed out that anyone claiming SMR is the problem here can make the same argument against any of the other components I listed.   It is not grasping at straws to point out what should be plainly obvious, but it seems everyone is too wrapped up in their desire to be a victim consumer to see that the only thing we know is a few WD reds failed in a few raid setups and they were SMR drives. That's it we know nothing more so stop pretending you do. 

 

Don't bother posting again until you have some sort of evidence there is a problem with SMR.

 

EDIT: and just to clarify (though I shouldn't need to) I never made any mention or insinuation that there was a known performance gap in anything (that's what you are doing),  my claims have ALWAYS been that we don't know, and without knowing it is pointless and erroneous to claim you do know.

 

EDIT2: I see after Leadeater  clearly pointed out what my examples meant,  that you ignored it completely and quoted me to continue arguing.  Please read what he has said, read the links he provided (especially the first one as it pertains to DM-SMR).  Although they are scientific articles and you may not understand them, at least try.  if you aren't going to listen to people on here who are clearly well educated on the matter you may as well try reading from other experts.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

6 hours ago, leadeater said:

Did I read this right?

Quote

Random I/O throughput (with the volatile cache en-abled or with high queue depth) is high (Test 7)—15×that of the equivalent CMR drive. This is a generalproperty of any SMR drive using a persistent cache.

From the first source paper. - If a drive is using persistent cache then it's random read and writes will be faster than a CMR drive?  If that is so then I can see why they would use it in consumer grade hardware,  you'd likely never know it was SMR and yet get the bigger storage metric as well.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

8 hours ago, mr moose said:

Did you know it was SMR when you bought it?

Yes as I said multiple times.

8 hours ago, mr moose said:

and now you are arguing had you known that they were SMR before hand it would have made a difference to your purchase

I never said that. I knew about it and decided it was OK. Which is the whole point, if I did not know I'd have expected something different, and I would have been pissed when I found out they were. And you keep advocating that we don't need to know.

8 hours ago, mr moose said:

now I want to know why you are using laptop and external storage designated drives in a raid array.

Becasue nobody makes HDDs that are actually designed for the specific purpose I had, so that's what I had to go with. They're in RAID becasue that way I can at least write to them at 130MB/s or so instead of 70...

 

F@H
Desktop: i9-13900K, ASUS Z790-E, 64GB DDR5-6000 CL36, RTX3080, 2TB MP600 Pro XT, 2TB SX8200Pro, 2x16TB Ironwolf RAID0, Corsair HX1200, Antec Vortex 360 AIO, Thermaltake Versa H25 TG, Samsung 4K curved 49" TV, 23" secondary, Mountain Everest Max

Mobile SFF rig: i9-9900K, Noctua NH-L9i, Asrock Z390 Phantom ITX-AC, 32GB, GTX1070, 2x1TB SX8200Pro RAID0, 2x5TB 2.5" HDD RAID0, Athena 500W Flex (Noctua fan), Custom 4.7l 3D printed case

 

Asus Zenbook UM325UA, Ryzen 7 5700u, 16GB, 1TB, OLED

 

GPD Win 2

Link to comment
Share on other sites

Link to post
Share on other sites

13 minutes ago, Kilrah said:

 And you keep advocating that we don't need to know.

 

I'm actually advocating (arguing) that people find out whether it's an issue or not before demanding they be told.    Remember you can look up many hdd's performance ratings on both passmark and userbenchmark, so if you have a HDD in mind you can get an idea how it's going to perform compared to other models SMR or not.

 

EDIT: to be more precise, if you look over both the current threads on this topic, I have been arguing that the reasons people are using to argue it should be labeled are not rational.  Because majority of the comments are that it will halve performance or fail in a raid.  We know neither of those to be factually correct on current drives. All we know is some did get kicked out of a raid and that some consumer drives are SMR. 

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

SMR has in impact on performance and use case, it can no longer be trusted that a drive in a line that is marketed for a particular use case is appropriate for it, and manufacturers are specifying these things in other circumstances e.g flash type used on SSDs so there's no reason they don't do this for HDDs as well now that they have started introducing technologies that have similar impacts on them.

F@H
Desktop: i9-13900K, ASUS Z790-E, 64GB DDR5-6000 CL36, RTX3080, 2TB MP600 Pro XT, 2TB SX8200Pro, 2x16TB Ironwolf RAID0, Corsair HX1200, Antec Vortex 360 AIO, Thermaltake Versa H25 TG, Samsung 4K curved 49" TV, 23" secondary, Mountain Everest Max

Mobile SFF rig: i9-9900K, Noctua NH-L9i, Asrock Z390 Phantom ITX-AC, 32GB, GTX1070, 2x1TB SX8200Pro RAID0, 2x5TB 2.5" HDD RAID0, Athena 500W Flex (Noctua fan), Custom 4.7l 3D printed case

 

Asus Zenbook UM325UA, Ryzen 7 5700u, 16GB, 1TB, OLED

 

GPD Win 2

Link to comment
Share on other sites

Link to post
Share on other sites

43 minutes ago, Kilrah said:

SMR has in impact on performance and use case, it can no longer be trusted that a drive in a line that is marketed for a particular use case is appropriate for it, and manufacturers are specifying these things in other circumstances e.g flash type used on SSDs so there's no reason they don't do this for HDDs as well now that they have started introducing technologies that have similar impacts on them.

 

SMR has an impact?  Pick any current market drive and show what that impact is for the use it is marketed for.

 

 

 

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, mr moose said:

 

SMR has an impact?  Pick any current market drive and show what that impact is for the use it is marketed for.

 

 

 

Dude, SMR literally halves the write performance across the board on all drives no matter where or what they are marketed for. Particularly when there is mixed operation, scattered write or with near full drive. SMR isn't some novelty that we're just discovering how it even performs. It's well know and documented. If there was no difference (like there isn't coz of oxide layer type), no one would give a F about it.

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, RejZoR said:

Dude, SMR literally halves the write performance across the board on all drives no matter where or what they are marketed for. Particularly when there is mixed operation, scattered write or with near full drive. SMR isn't some novelty that we're just discovering how it even performs. It's well know and documented. If there was no difference (like there isn't coz of oxide layer type), no one would give a F about it.

 

 

3 hours ago, mr moose said:

 

Don't bother posting again until you have some sort of evidence there is a problem with SMR.

 

I was serious, if you can't back up what you are claiming with evidence then I don't want to talk to you. I am not going to waste time debating things that aren't evidenced.   Everything you just posted has been addressed very clearly if not by me then definitely by leadeater.  

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, mr moose said:

 

 

I was serious, if you can't back up what you are claiming with evidence then I don't want to talk to you. I am not going to waste time debating things that aren't evidenced.   Everything you just posted has been addressed very clearly if not by me then definitely by leadeater.  

Holy fucking shit, calling you a vendor shill or a fanboy at this point would be an understatement. Literally whole world know show SMR affects performance and you want me to post "evidence". What the F dude.

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, RejZoR said:

Holy fucking shit, calling you a vendor shill or a fanboy at this point would be an understatement. Literally whole world know show SMR affects performance and you want me to post "evidence". What the F dude.

 

So it shouldn't be hard to link to some independent testing of these drives should it?  

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

Yeah, good luck finding anything now that entire web is flooded with this fiasco shit if you just mention SMR anywhere in the search string... It's a fucking common knowledge, like water is wet. But if you want to know exactly how wet water is, go find it yourself. ffs

Link to comment
Share on other sites

Link to post
Share on other sites

For anyone who is interested in where the technology is so far:

 

Quote

 RAID can be used with DM SMR drives and HA SMR drives that support drive-managed mode. SMR drives in RAID arrays have the same limits as SMR drives used as individual drives. Despite this, RAID can help aggregate the performance of multiple SMR drives in the same way that it does with CMR drives, achieving higher overall performance levels and data availability in workloads.

 

From:

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=6&ved=2ahUKEwjPrcKU3_voAhWkxDgGHaPaAYcQFjAFegQIBxAB&url=https%3A%2F%2Fwww.microsemi.com%2Fdocument-portal%2Fdoc_download%2F137506-optimizing-smr-drives-in-raid-configurations&usg=AOvVaw1o9QbWD6DwVjewLBT3Ke5q

 

Roughly what that means is these devices failing is not because they are SMR as DM SMR in raid is a thing.

 

So far the only real metric that I would expect to suffer would be random writes.  everything else is the same or better in newer SMR drives. 

 

 

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, mr moose said:

From the first source paper. - If a drive is using persistent cache then it's random read and writes will be faster than a CMR drive?  If that is so then I can see why they would use it in consumer grade hardware,  you'd likely never know it was SMR and yet get the bigger storage metric as well.

Yes but I don't think any are currently using a persistent cache setup like they were, flash/SSD. You still get 10x on current drives from the large caches and I suspect the larger cache is to negate some of the problems outlined to do with the map update. You can already see this much better random I/O performance on the old Seagate Archive v2 disk, which is being used in both papers I linked. This disk only has 128MB cache where current models have either 256MB or 512MB.

 

You can see the improved random I/O here:

seagate_archive_8tb_sata_4k_randomtransf

 

seagate_archive_8tb_sata_4k_randomtransf

 

seagate_archive_8tb_sata_4k_write_latenc

 

However if you do sustained long term writes you'll exhaust the the Media Cache and I/O will take a backseat to cleanup

seagate_archive_8tb_sata_main_4kwrite_th

So there is a known limitation to such disks but I would expect after 5 years since this one came out and ones made for NAS usage rather than archival would be better, but will still have the same problem if you write to it for long enough.

 

For actual usage of this drive in a NAS performance is acceptable:

seagate_archive_8tb_cifs_main_4kwrite_th

 

seagate_archive_8tb_cifs_main_4kwrite_av

 

seagate_archive_8tb_cifs_main_4kwrite_ma

 

Nothing makes it standout bad however the disk is designed for archive usage and to be in systems with a lot more than 8 disks. These are for massive arrays with hundreds of disks and controller caches and deployed with workload in mind. The price of these were never really low enough to be worth it in low qualities over standard NAS/enterprise disks.

 

So basically yes there is a downside to DM-SMR but it's questionable if people will actually encounter it and they won't have alternative options. Thing is we are talking about the cheapest NAS disks on the market, what is really the expectation, are we really surprised a budget focused cheap device went for a cost optimized design? I'm not.

 

On a more expensive cost front as the paper shows you can deploy a DM-SMR disks with flash media cache on the disk controller board. I'd be very interested in such a disk with 128GB single flash chip on it, wouldn't be cheap though.

Link to comment
Share on other sites

Link to post
Share on other sites

No no no,The whole article is misleading!,

The article of the original source investigated SMR drives sold as NAS drives in the Western Digital lineup,the problem is that SMR drives are not suitable for NAS/RAID therefore it's false marketing on Western Digital's end!.

Other manufacturers issued a statement of course - It's a huge scandal for Western Digital and for the Hard Drive industry.

I am working on a more accurate article!.

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE GTX 1660 GAMING OC @ Core 2085MHz Memory 5000MHz
Cinebench R23: 15669cb | Unigine Superposition 1080p Extreme: 3566
Link to comment
Share on other sites

Link to post
Share on other sites

10 hours ago, Vishera said:

I am working on a more accurate article!.

Or you could just read leadeaters responses, they are the best written and contain all the evidence to support his explanations.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

Did Seagate's policy on listing smr change? 

I bought a Barracuda little over year ago and knew it was smr, could have sworn it was on product page, not listed now when I rechecked.

 

Smr on barracuda is just fine for intended use, large cache makes it perform well.  Only time slower speed shows up is with 20-100gb file copies, start fast and then drop off noticeably.  Work just fine to hold media or steam library on a desktop.

 

WD Red marketed for nas is issue.  They would still work in raid 0/1/mirrors/jbod, but causes possible loss of array in raid5/6/7/z.  If it was listed, then when setting up a 4 drive array one would just be sure to use raid 10 and not raid 6.  Since it is not listed, user only finds out after data loss.

 

Red drives were recommended drives for freenas, the WD makes change to product that causes data loss in freenas with out any indication, when whole point of free nas is preventing data loss.  Wonder how long it will take for all the recommendations to change from Red to Ironwolf?

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Sophia_Borjia said:

Wonder how long it will take for all the recommendations to change from Red to Ironwolf?

I am sure the negative rep is already taking effect.

11 hours ago, mr moose said:

Or you could just read leadeaters responses, they are the best written and contain all the evidence to support his explanations.

Well i already made a thread about it in General Discussion instead (It got moved to a different forum):

I think reading the comments in the discussion can be interesting and helpful to understand it better :D

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE GTX 1660 GAMING OC @ Core 2085MHz Memory 5000MHz
Cinebench R23: 15669cb | Unigine Superposition 1080p Extreme: 3566
Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Sophia_Borjia said:

Wonder how long it will take for all the recommendations to change from Red to Ironwolf?

I'd say basically already. Couldn't really see anyone recommending it for that right now. But even then it looks like problems exist outside of FreeNAS so I'd be hesitant to recommend until a firmware update for the disks comes out or someone can show exactly how/why the fault condition comes up.

 

Problem is Ironwolf cost more and is a performance tier up from WD Red, cost is higher than Red Pro even.

 

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, leadeater said:

I'd say basically already. Couldn't really see anyone recommending it for that right now. But even then it looks like problems exist outside of FreeNAS so I'd be hesitant to recommend until a firmware update for the disks comes out or someone can show exactly how/why the fault condition comes up.

 

Problem is Ironwolf cost more and is a performance tier up from WD Red, cost is higher than Red Pro even.

 

What about the TOSHIBA N300 NAS drives?

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE GTX 1660 GAMING OC @ Core 2085MHz Memory 5000MHz
Cinebench R23: 15669cb | Unigine Superposition 1080p Extreme: 3566
Link to comment
Share on other sites

Link to post
Share on other sites

12 minutes ago, Vishera said:

What about the TOSHIBA N300 NAS drives?

Personally I have no experience with them, at least as far as I know. I've only used Seagate, HGST and Hitatchi at work and those are all the top end product line they have and at home same story as I take the used/decommissioned stuff. I've got some WD Reds at home as well but only for Plex so they do almost nothing.

 

The N300's spec wise look fine and review well enough, https://www.storagereview.com/review/toshiba-n300-nas-hdd-review-8tb, but don't they also cost quite a bit more than Reds?

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, leadeater said:

The N300's spec wise look fine and review well enough, https://www.storagereview.com/review/toshiba-n300-nas-hdd-review-8tb, but don't they also cost quite a bit more than Reds?

Yep,They are more expensive but they at least work as intended with RAID unlike the Red drives...

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE GTX 1660 GAMING OC @ Core 2085MHz Memory 5000MHz
Cinebench R23: 15669cb | Unigine Superposition 1080p Extreme: 3566
Link to comment
Share on other sites

Link to post
Share on other sites

On 4/18/2020 at 4:23 PM, hitardo said:

Toshiba told Blocks & Files yesterday that its P300 desktop disk drives use shingled magnetic recording technology (SMR), which can exhibit slow data write speeds. However, the company does not mention this in end user drive documentation.

So that's why my HDD is slower than my ancient Maxtor...

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×