- Feb 4, 2016
- 2,520
Western Digital Fesses Up: Some Red HDDs Use Slow SMR Tech Without Disclosure
- Thread starter LASER_oneXM
- Start date
scorpionv
Level 2
- Apr 20, 2020
- 87
Yep, heard it on the news. Western Digital, Seagate and Toshiba as well. Sneaky move.
The benefit for these companies is that SMR increases the drive capacity with the same hardware. Speed and especially IOPS suffer terribly though, but the average customer chooses a certain capacity instead of looking at IOPS.
The benefit for these companies is that SMR increases the drive capacity with the same hardware. Speed and especially IOPS suffer terribly though, but the average customer chooses a certain capacity instead of looking at IOPS.
- Mar 2, 2017
- 318
How is this sneaky? I see all the uproar over this as some in the IT media simply whining and trying to gain attention with sensationalized head lines.
Before SMR came about, did drive specs specify they used CMR technologies? No!
AS LONG AS the drive makers publish the essential performance and warranty specs (which they do), why does it matter?
Users have been demanding higher capacities and cheaper prices for years and SMR delivers that. The only ones with a valid complaint, IMO, are those who want the highest performance. And they just need to do their homework before buying.
When it comes to mass storage of data on hard drives, I don't really care how fast the hard drive is. And in discussing this with several colleagues, they don't either. If we want the highest performance, we buy SSDs!
scorpionv
Level 2
- Apr 20, 2020
- 87
Well, my previous link has been blocked or something, but I found more or less the same article from a more respectable source:
Western Digital, Seagate Are Shipping Slow SMR Drives Without Informing Customers: Reports [Updated] - ExtremeTech
When you downgrade a productline, with your own profit in mind, without clearly informing the customer: that's sneaky in my book. Feels like being scammed.
Western Digital, Seagate Are Shipping Slow SMR Drives Without Informing Customers: Reports [Updated] - ExtremeTech
When you downgrade a productline, with your own profit in mind, without clearly informing the customer: that's sneaky in my book. Feels like being scammed.
- Mar 2, 2017
- 318
First, Western Digital is not a non-profit charity. Of course business decisions are made with profit in mind. The company would go out of business if they didn't. This is especially true in this highly competitive market.
Second, where did it ever say WD Reds always have and always will use CMR?
Third, they did clearly publish performance specs.
Fourth, those drives that were made with CMR did NOT suddenly start using SMR on the same models. That is, they have different model numbers.
It is not like WD intentionally mislabeled the drives. They did not claim CMR then use SMR. They did not claim transfer rates of 250 MB/s when in reality, they only achieved 210 MB/s. They did not claim they came with 512MB buffers when in reality they only came with 256MB.
To be sure, marketing weenies are not my favorite people. I put them right up there with ambulance chasing shysters, big pharma, insurance companies, wannabe journalists and [so called] news outlets who think it their jobs to make or spin the news, instead of just reporting it.
Had WD and the other drive makers lied, then I would be 100% in agreement with you. In fact, I would be calling that criminal. But they didn't. Users asked for bigger, cheaper drivers. Drive makers delivered.
I see this similar to past Corsair business decisions. Years ago, if a PSU said Corsair on it, you knew you had a quality, reliable power supply. Then, Corsair looking to stay competitive and keep profits up decided to change OEM suppliers for their entry level lines. They were still VS and CS series productline PSUs, but they had different model numbers. This tarnished Corsair's reputation even though their upper tier supplies were still top notch. Did I and other consumers like it? NO! But was it sneaky? No.
Companies MUST make a profit. That's how they pay their employees, maintain overhead, pay taxes, and invest in R&D - and pay their shareholders.
Now if you want to talk about being sneaky, have you looked at toilet paper lately? The industry standard used to be where a sheet of TP measured 4.5 inches by 4.5 inches. And a "Mega" roll used to have 1000 sheets on it. Now look at this Cottonelle (hover over fine print in lower left corner of package to zoom). The sheets are just 3.82 x 4 inches and there's only 340 sheets per roll - 1 ply sheets at that. And that "Mega" roll is supposed to equal 4 regular rolls! When did a "regular" roll of toilet paper ever come with just 85 sheets of 1 ply toilet paper on it? I'm surprised the package does not say "New and Improved" on it.
Second, where did it ever say WD Reds always have and always will use CMR?
Third, they did clearly publish performance specs.
Fourth, those drives that were made with CMR did NOT suddenly start using SMR on the same models. That is, they have different model numbers.
It is not like WD intentionally mislabeled the drives. They did not claim CMR then use SMR. They did not claim transfer rates of 250 MB/s when in reality, they only achieved 210 MB/s. They did not claim they came with 512MB buffers when in reality they only came with 256MB.
To be sure, marketing weenies are not my favorite people. I put them right up there with ambulance chasing shysters, big pharma, insurance companies, wannabe journalists and [so called] news outlets who think it their jobs to make or spin the news, instead of just reporting it.
Had WD and the other drive makers lied, then I would be 100% in agreement with you. In fact, I would be calling that criminal. But they didn't. Users asked for bigger, cheaper drivers. Drive makers delivered.
I see this similar to past Corsair business decisions. Years ago, if a PSU said Corsair on it, you knew you had a quality, reliable power supply. Then, Corsair looking to stay competitive and keep profits up decided to change OEM suppliers for their entry level lines. They were still VS and CS series productline PSUs, but they had different model numbers. This tarnished Corsair's reputation even though their upper tier supplies were still top notch. Did I and other consumers like it? NO! But was it sneaky? No.
Companies MUST make a profit. That's how they pay their employees, maintain overhead, pay taxes, and invest in R&D - and pay their shareholders.
Now if you want to talk about being sneaky, have you looked at toilet paper lately? The industry standard used to be where a sheet of TP measured 4.5 inches by 4.5 inches. And a "Mega" roll used to have 1000 sheets on it. Now look at this Cottonelle (hover over fine print in lower left corner of package to zoom). The sheets are just 3.82 x 4 inches and there's only 340 sheets per roll - 1 ply sheets at that. And that "Mega" roll is supposed to equal 4 regular rolls! When did a "regular" roll of toilet paper ever come with just 85 sheets of 1 ply toilet paper on it? I'm surprised the package does not say "New and Improved" on it.
- Jul 27, 2015
- 5,458
Buyer beware—that 2TB-6TB “NAS” drive you’ve been eyeing might be SMR
Hard drives were already bad at random access I/O—but SMR disks are worse.
arstechnica.com
What is this claim that ZFS data array rebuilding doesn't work on SMR drives?
scorpionv
Level 2
- Apr 20, 2020
- 87
From the Arstechnica article:
There has been speculation that the drives got kicked out of the arrays due to long timeouts—SMR disks need to perform garbage-collection routines in the background and store incoming writes in a small CMR-encoded write-cache area of the disk, before moving them to the main SMR encoded storage.
Could be true, I did have similar issues with Samsung disks in a Synology NAS. They kept spinning down, and back up again, and so on, and ended up being thrown out of the disk array, leaving it degraded.
Last edited:
scorpionv
Level 2
- Apr 20, 2020
- 87
Well, lets agree to disagree. I think they should have labeled these drives differently, to make it absolutely clear that they are inferior to the previous WD Red drives. Call it WD Red Economy or something. That would involve cutting prices and profits however.
The change of type revision code does indeed point out change. So technically, no lies being told, but truth being omitted.
Ah got it. The dreaded timeout interactions with storage controllers.
Yeah in my opinion it's not the best business practice to keep the same marketing name but significantly change the underlying revision. Not everything that's legal is commendable. When shopping for hard drives, it seems perfectly reasonable to expect that the 4/6/8TB versions of "WD Red" purely differ in storage capacity and not in how they act.
It's a gray area, but that's been the whole NAS market IMO. I have two NASes with a mix of WD Red and HGST drives, and will readily admit: I'm using about $1000 to replicate the functionality of a $50,000+ enterprise storage array. But it works for me and I'd like to keep it working, but I totally understand how disk manufacturers don't really care about this market.
- Mar 2, 2017
- 318
Huh? You quoted me where I said WD is not a non-profit charity, then you say agree to disagree? So you are saying they are a non-profit charity?
And did you see in upnorth's link where not only did WD label the drives as SMR, the Ars Technica article even included a link to drives that clearly shows SMR on the label?
There is nothing in the description for the Red drives indicating they use SMR or CMR. And the products ARE being relabeled to show those that are SMR, are labeled as such.
What product in any industry do you know puts "New and Degraded Performance" in big flashing letters on their packaging? Why should WD be different?
I am NOT saying I'm in favor of any company using cheaper parts or technologies that degrade performance. I would much rather see products improve in quality and performance while costing less and having less detrimental impact on our environments. And some times that happens. But those are pretty much exceptions.
All I am saying is WD is being criticized for being a sneaky evil beast, purposely deceiving its customers when I see nothing suggesting that. What I see is WD being asked to provide bigger drives at cheaper prices, and they did that. There's nothing to suggest these are failing before warranties wear out or they are bursting into flames. Some are speculating things - but speculation means nothing without hard evidence.
scorpionv
Level 2
- Apr 20, 2020
- 87
I'm sorry, I should have quoted the while post, but did not so for the sake of brevity.
I hear your opinion, and I respectfully disagree. I strive to be clear and fair in what I do. That means I would never do business like WD did, and in my opinion, they deserve the heat they get about it.
- Jul 27, 2015
- 5,458
Seagate says Network Attached Storage and SMR don’t mix
Seagate’s IronWolf and IronWolf Pro NAS drives are all certified SMR-free.
arstechnica.com
- Mar 2, 2017
- 318
Thanks for that.
FTR, I had my own business too and while we did provide products (custom built PCs) it was much more service oriented than product oriented. However, the only advertising I did was "word of mouth" so our goal was to always give our clients value for their money so (1) they would come back and (2) they would recommend us to their friends. There's an old expression I learned in the military that fits when your business depends on customer satisfaction and "word of mouth" advertising, especially when it comes to getting new customers - essential for growth. And that is, "One aw sh!t wipes out a 1000 attaboys".
And sadly, that "one aw sh!t" can simply apply to "perceptions" too. If a client "perceives" he has been deceived or misled, even if he really wasn't being deceived or misled, that upset client may tell 5 friends who then tell 5 of their friends (exaggerating and sensationalizing the story as it is retold). And then those friends post unfavorable reviews on Facebook and other media outlets and suddenly, this innocent business is shamed and blacklisted out of business for no real reason. Is that fair?
So I would not want even the perception I was doing what WD is accused of doing either.
But to stop things like that from happening, I feel it is the responsibility of those in the IT media to do their "due diligence" to research and suss out the facts, cull out the rumors and falsehoods and publish the unbiased truth so we consumers have the real facts to make informed decisions. Is that what happened with the original article? I don't think so. Is that what happened when other media outlets parroted the first? I don't think so.
So that leaves me wondering, what did WD really do wrong, deceitful, or misleading that makes them "deserve the heat" they are getting?
What I am saying is, I don't see what they have been done deserves being "accused" of being deceptive or misleading. Consequently, I am not understanding what you base your opinion on. I respect your opinion and will defend your right to express it. And while we are all entitled to our opinions, we, and especially the IT media, are not entitled to our own facts.
What I see is a company introducing "NEW" models with "NEW" model numbers to their line up that offer greater capacities at cheaper prices.
I don't see where they misrepresented the product as having greater performance or better features than the "published" specs.
I don't see any WD document or marketing ad that says all "Red" drives will be CMR drives.
In other words, I don't see any false advertising, deceptive or misleading product descriptions.
The complaint seems to be "The company failed to disclose the use of SMR in advertising or spec sheets." Okay. Where did they EVER disclose the use of CMR before? Why are they not being chastised for not disclosing the country origin for the capacitors used on the logic board (since stating caps are "Japanese" seems to be a thing)? Why not chastise them for not disclosing the type bearings used in the motors?
I say if we don't like a product a company produces, don't buy it. That's how you send a message to a company.
I've seen time and time again that this means the company will try a little harder next time to be less forthcoming about differences that they perceive to affect whether or not one would buy the product.
Critical publicity has its place too and gets us talking about some of these considerations. Perhaps it does turn out that the worries around SMR drives are unfounded. If you remember, there was some initial controversy around iPhones when they switched to TLC NAND, and how that would usher in an era of poor performance reliability. Except now almost everything except enterprise caching SSDs are using TLC (or even more layered) NAND and they even more performant and reliable than SLC/MLC NAND from back in the old days.
- Mar 2, 2017
- 318
Try harder to be "less" forthcoming?
First, there is a big difference between marketing hype and published technical specs.
Second, WD did not market SMR as being superior to CMR. And their tech data sheets did state SMR.
Third, these are "NEW" models WD put out. They did not make "sneaky" changes to the technology in existing models, then market those models as the same.
Fourth, as consumers, we still need to do our homework.
Fifth, whether or not I will buy from Company A or not will depend on my own past experiences with the company, "professional" reviews from [hopefully] reliable, unbiased sources, and word-of-mouth comments from friends and colleagues I trust.
If I am looking for an inexpensive, high capacity drive, and WD or Seagate or whoever delivers, and that drive does not fail on me prematurely, do I really care if CMR or SMR? No. Will I buy from them again? Probably - at least I will give them a good, fair, look.
If I am looking for a high performance, top quality, high capacity drive, with a long warranty will I be looking at a company's budget lines? No.
This discussion reminds me of super cheap printers. People demanded cheaper and cheaper printers and the printer companies delivered. Do super budget printers use the latest and greatest technologies and offer the best performance? NO! Is it fair to expect $400 performance and reliability out of a $60 printer? NO! But do some people still expect it? Yes!
Should HP, Epson, Brother, etc. be chastised for putting out a budget printer that does not incorporate the latest, greatest and more expensive technologies? If the same judgmental reasoning is used on HP as it is WD, then yes.
I am NOT defending WD for using SMR. I am defending WD for the unreasonable, unfair criticism they are receiving for using it on some of their models - especially from the criticism from those claiming to be journalist in the IT media who should be unbiased when reporting the "facts".
Try harder to be "less" forthcoming?
I am NOT defending WD for using SMR. I am defending WD for the unreasonable, unfair criticism they are receiving for using it on some of their models - especially from the criticism from those claiming to be journalist in the IT media who should be unbiased when reporting the "facts".
The point I was really trying to make (as someone who's been working for multibillion dollar corporations for 10 years) is that even as someone feeding you the dogfood, I do not think "Don't buy this product and that will punish companies" actually works well in the real world. The way that gets spun is "if we fool you into buying it and not returning within the return window, we have profited". The value of "loyalty" is kind of a past concept. Companies that sell commodity parts don't really care who you'll buy from NEXT, just whether or not they can complete this sale today.
As far as criticism in tech journalism, I agree with you on that point. However, journalism has NEVER withstood the scrutiny of domain experts. Like we've seen a bunch of negative headlines about Norton products causing blue screens and then it turns out it's only Symantec Endpoint Protection and on specific hardened Windows Server configs. Or the latest Windows Update did something horrible, and it turns out that it's for a small fraction of customers and still under investigation. It's frustrating and worth putting into perspective, but it's kind of a fact of life too.
For NAS drives though, it's kind of frustrating because WD Red drives aren't supposed to be some budget low cost thing. It is a product line that NAS owners have been forced to use because the cheaper WD Green drives started removing configuration options that made them behave in RAID-friendly ways (Western Digital Red (with TLER) vs Green (with wdidle set to disabled) drives) and the WD Red was pitched as the RAID controller friendly drive for NASes. If it's actually true that SMR drives are having long commit times resulting in RAID controller timeouts, that is something that WD should be criticized for. Having a drive drop out of a RAID array during rebuilding can be an extremely risky and devastating process. Especially with ZFS, it can be extremely difficult to recover data once you violate the base premise that if a ZVOL loses all redundancy you will lose the entire pool.
scorpionv
Level 2
- Apr 20, 2020
- 87
That is a good point. False rumors can hurt a business, and that is exactly why WD should have made it absolutely clear that a new WD Red drive is functionally different from an old one. I do not care if they put Japanese or Turkish capacitors in drives, as long as they perform the same. But performance decrease, or incompatibility with some NAS configurations is more than a minor 'under the hood' change. Be transparent about it.
That is exactly what I am going to do. Normally I buy Crucial SSDs and HGST HDDs. HGST is WD nowadays, but WD hasn't shut down the HGST designs yet.
Absolutely true. Without bad publicity, WD would not change a thing. Now I hope they clarify their communication about SMR, and best case this publicity leads to innovation to improve SMR performance for future drives. When WD accomplishes that, they are encouraged to communicate and celebrate that of course
scorpionv
Level 2
- Apr 20, 2020
- 87
Yep, that was frustrating as well, but at least WD made it perfectly clear that you should buy WD Red drives for NAS/RAID use. New product line, new market share.
Yep and we learn each step of the way. We can't be stuck with old storage technology forever. But at the same time, the path towards the future leads to a lot of growing pain and learnings in the meantime. Maybe SMR won't end up being a big deal. Maybe it's the right tradeoff for vastly increased storage at the cost of IOPS. Maybe a firmware or software update to the drive or RAID controllers is what we need to make these drives work well in a NAS.
