Did Western Digital did it again ?

[ChrisRJ]

RMA replacement drives are typically recertified unless there is insufficient stock of repaired product.

Thanks! Shouldn't it read "aren't typically"?

The retail return period is unfortunately over. That would indeed have been my preferred way.

I am currently thinking of getting myself another spare disk, although probably a Toshiba or WD/HGST model.

 

[jgreco]

Thanks! Shouldn't it read "aren't typically"?

No, it shouldn't.

Drives you receive back from RMA are typically someone else's drive that had been returned for RMA, repaired, tested for acceptability, and then boxed up. This process may take days or weeks, and they don't want to take responsibility for your data, so you don't get your old repaired drive back, ever, as far as I know. This also allows them to provide a guaranteed turn-around time.

Some percentage of drives sent for RMA will have damaged platters, or ruined heads, or fried circuit boards. In such case, I would guess that the repair techs may attempt to stitch together parts from two or more drives to make one recovered frankendrive. In such a case, two drives came in but only one repaired drive leaves. In such a case, there would be a shortfall of repaired product, and in this case, the supply of drives to be returned to customers has to be supplemented with new devices.

When new devices are no longer available, such as with my story about the Intel 535 SSD's, manufacturers typically substitute an equivalent-or-better product, although substituting a slightly-crappier product has been seen to happen as well.

I believe my wording above correctly but tersely conveys this.

 

[ChrisRJ]

No, it shouldn't.
[..]

Thanks a lot!

 

[Constantin]

Now that they did it twice, they will end up blacklisted permanently on my side...

Unfortunately, WD has been a poster child of how to destroy customer goodwill for a while. No doubt, some day their current management practices will show up in a B-school case study, just like Tylenol vs. the Dalkon Shield. WD started with the silent Red SMR switcheroo, followed by "5900 RPM class" drives, and now finally this QLC/controller deception. Clearly, this is a firm-wide problem, they have not learned from their past transgressions, they continue to offend against their user base.

I have no doubt that OEMs struggle to get parts, much like many other firms in the IT supply chain. But switching controllers, flash type, etc. is not like swapping in a different, but similarly-spec'd voltage regulator. When significant changes are made to the product, it needs to get a new SKU and spec sheet to make these changes transparent.

The current approach at WD of replacing the inferior drives with performant ones for customers who notice and complain is particularly offensive. It allows the company to capture the profit from all the users who didn't notice or want to bother with getting their drives replaced while accommodating the few "squeaky wheels".

Given the repeat pattern, I hope that an AG will take notice and start suing not just the company. Management has to be held to account, ideally with jail time, in order to put a cost on this sort of behavior. Otherwise, we can expect every storage OEM to do this in the future (think prisoners dilemma).

 

[Forza]

What's IX System's comment on this? They are partnered with WD, at least based on the amount of WD ads on this site.

 

[winnielinnie]

followed by "5900 RPM class" drives

What are these drives exactly? Even the Red Plus and White Label (shucked from externals) spin at 7200 RPM, but are only rated for 5900? Or is "5900 class" relegated only to lower-end consumer drives, such as WD Blue?

The 8TB Red Plus drives I purchased in the past claim "7200 RPM" on WD's web site, yet smartctl reports "5400 RPM". So which is it?

Code:

# smartctl -a /dev/ada0
Model Family:     Western Digital Red
Device Model:     WDC WD80EFAX-68KNBN0
User Capacity:    8,001,563,222,016 bytes [8.00 TB]
Sector Sizes:     512 bytes logical, 4096 bytes physical
Rotation Rate:    5400 rpm
Form Factor:      3.5 inches

 

[jgreco]

What's IX System's comment on this? They are partnered with WD, at least based on the amount of WD ads on this site.

No one on these forums, except a handful with an "iXsystems" badge under their icon, work for iXsystems. iXsystems generally does not make comments about this kind of thing on these forums, with the possible exception of "we'll look into it" type of vagueness.

A review of their commentary surrounding the SMR debacle would likely support my viewpoint, and this is nowhere near that.

 

[Constantin]

What are these drives exactly? Even the Red Plus and White Label (shucked from externals) spin at 7200 RPM, but are only rated for 5900? Or is "5900 class" relegated only to lower-end consumer drives, such as WD Blue?

As I understand it, the production of 5,X00 RPM,3.5" drives has ceased at WD (or did, temporarily). Rather than stop selling these popular drives (NAS users liked them for the marginally lower wear, noise, heat, and power consumption), WD started selling "5,900 RPM class" drives that combined the worst of all worlds, i.e. the marginally higher wear, heat, power consumption, and lower life of a 7,200 RPM drive with the slower performance of a 5,x00 RPM drive.

Seeing that this started not too far after the SMR debacle, the follow-on public mea culpas and pledges for transparency from WD, this misdirection did not sit well with me. With the silent SSD downgrade, this is now the third time that WD management has been caught with its hand in the cookie jar, followed by mealy-mouthed pledges for more transparency in the future, case-by-case replacements of the hardware sold under deceptive conditions, etc.

To me, this is a pattern of deception, not a one-time mistake, and financially defrauding a large group of customers by selling them a inferior set of goods than the ones they agreed to buy should lead straight to making new acquaintances at the local state or federal penitentiary. If it doesn't then we should expect more of this, not less, as all other members of the SSD and HDD oligopoly follow in WDs footsteps (hello, Crucial, Samsung, and among others).

The 8TB Red Plus drives I purchased in the past claim "7200 RPM" on WD's web site, yet smartctl reports "5400 RPM". So which is it?

As with anything published at WD, I'd verify rather than trust. The RPM deception was uncovered using a microphone and looking at the frequencies the drives produced.

 

[winnielinnie]

Excellent clarification! I'm just beside myself how they continue this pattern of obfuscation and mistrust, yet manage to remain a household name with storage drives.

While it's tempting for me to say "I'm only going with Seagate IronWolf from now on," I don't get get to choose when a drive starts failing or completely drop dead. If that was the case, I'd only allow my drives to fail on Black Friday or Cyber Monday.

It's hard for me to "vote with my dollars" after-the-fact, since beyond initial setup, it's a matter of purchasing drives for the sake of replacing dead/failing ones. Whatever happens to be the best deal is what I'm likely going to purchase. I don't look forward to paying an extra 20% for an 8TB replacement "NAS" drive when the time comes. (I'm referring to the costs of a shuckable WD versus an IronWolf.)

 

[jgreco]

To me, this is a pattern of deception, not a one-time mistake, and financially defrauding a large group of customers by selling them a inferior set of goods than the ones they agreed to buy should lead straight to making new acquaintances at the local state or federal penitentiary. If it doesn't then we should expect more of this, not less, as all other members of the SSD and HDD oligopoly follow in WDs footsteps (hello, Crucial, Samsung, and among others).

Yeesh. A little harsh, maybe. I think you'll find that a hard sell to your local prosecutor. Clearly there are some behaviours like the CMR/SMR debacle where the actual usability of the product is adversely affected, and the SSD controller/speed shenanigans is problematic, but trying to nail WD for manufacturing 7200RPM drives and selling them as a lower tier product is going to be a hard sell, especially as the product hasn't changed specs mid-run. Switching components mid-run is sometimes necessary due to supply constraints. I would love to be guaranteed the exact same product every time, but many years in this business suggests that product revisions are relatively common. I would be fine with some transparency in the whole process.

 

[Constantin]

I completely agree that the "5,900 RPM class" baloney doesn't fall into the same category as silent downgrades of extant products. For me, the repeated patterns out of WD are the issue - i.e. deception followed by pledges for greater transparency. Now for the third time.

As a AG I'd settle on prosecuting a scummy pattern of behavior. The company can then settle quickly by pledging to be actually performant re: transparency. Set up some guide posts for the next 10 years. If the company is found to be materially breach, the entire c-suite agrees to forfeit any annual pay over $150k, 100% of bonuses, and all accumulated stock options, going back to the date of the complaint. Whistleblowers get to keep their pay, bonus, and stock options.

In other words, don't make the company pay for c-suite decisions that may benefit the individual but ultimately hurt the company, its stockholders, and employees. If you're really bloodthirsty, require the board to get AG approval for all future c-suite compensation changes to prevent end-runs. If the company stalls, sue civilly and criminally. Unfortunately, I don't see another way than to make a very public example, as the current HDD/SDD oligopoly prevents meaningful market feedback.

I don't have an issue with companies changing components but as you noted, transparency is key. If a change has a material impact on the product (such as CMR->SMR) then the least that I should be able to expect from a OEM is a new SKU, a new datasheet, and some sort of announcement noting these updates. I should not find materially inferior products being sold under old SKUs with better specs.

Given the lack of serious consequences to decision-makers at WD, it is not surprising that competitors are also starting to engage in this behavior, as it has been shown to be profitable and without meaningful risk. Unless a strong signal is sent to the market, expect more of this, not less.

 

[Etorix]

I have no doubt that OEMs struggle to get parts, much like many other firms in the IT supply chain. But switching controllers, flash type, etc. is not like swapping in a different, but similarly-spec'd voltage regulator. When significant changes are made to the product, it needs to get a new SKU and spec sheet to make these changes transparent.

Swapping a controller for another one with similar characteristics could be accounted for by adding a "A" to the product name or bumping the product number by one unit.
Replacing TLC memory by QLC, which performs significantly worse, would require a wholly different label. So WD is, once again, being caught red-handed for deceiving customers.

The issue is that there are not so many alternative vendors. For SSDs, it is still easy to just "avoid WD"—but less so to avoid bumping into similar issues with other vendors. For "spinning rust" hard drives there are only two alternatives to chose from.

 

[Constantin]

Careful. The SSD market is actually quite consolidated also. Not down to three like HDD, but a number of brands out there are partially or fully owned by larger OEMs. Just as many folk swear by HGST Reds, nowadays all those HDDs are made in the same factories as the other WD stuff. The only "premium" aspect that WD can't cheat on as easily is Helium-content - those drives tend to be made for data centers, not home use.

Switching controllers or other components is not a new thing either, as Toms Hardware pointed out in 12/2020. Rather, it's a problem that is creeping up from lower-tier brands like AData into the mainstream like WD Blue. I suppose WD Blue has some catching up to do as AData at one point was shipping three different drives (Controllers, clock, etc. changing) with the same SKU per that article.

My main issue with drive components getting swapped is that as a consumer it is hard to detect these issues consistently. The latest WD deception would only become apparent to heavy users who exhaust the SLC cache before it can be flushed to the much slower now-QLC not-TLC backend. To non-experts this would present as usually-snappy performance followed by unexplained I/O hiccups, but only under heavy use (which the user might infer to be a CPU-issue). The lower life expectancy of QLC is another factor that is not immediately apparent either.

I don't want to live in a world where purchases become a lottery between what was sent to reviewers, what was shipped at one point, vs. what was shipped to me. I don't want to have to verify firmwares, performance, or component use. I just want the purchased parts to meet the specifications of the OEM and to perform as consistently as the gear sent to reviewers or the broader market. Change something important like the controller, the flash type, etc. and a new SKU and datasheet should be coming due, no exceptions, so the marketplace can make an informed decision.

The same issue was at the heart of the WD SMR deception. WD may have changed the SKU but they didn't disclose the switch to SMR in their WD Red documentation and shipped a "NAS" product that is fundamentally unsuitable for many NAS applications based on official presentations made by their own employees. Instead, they intentionally misled their consumers until the evidence, uproar, and finally lawsuits cowed them into submission. But we also know why they did this, as an informed market values SMR drives at a discount relative to CMR drives due to their lower performance.

Given how many times this kind of deception has happened all across WD in the last two years, I see a pattern that only company leadership can stop. Hence the above. The WD board chair should have moved to reign in rogue management (per his/her/their fiduciary duty) but they have not. Thus, only the AG remains as a check on a pattern of deceptive management practices.

 

[joeschmuck]

The article wasn't a bad read. Too bad that this happened. I preference for the past several years have been Samsung SSD's, I own four of the EVO series (three 500GB and one 1TB). I'm very happy with them. I also own a few other brands which work well but those are much older and are 120GB or 256GB in capacity and used for smaller tasks. If they fail before I'm done with them then a nice new Samsung will replace it.

It's sad that this kind of stuff is happening, I depend on reviews and these test results to purchase my products, the bait-and-switch is a dirty move.

 

[Constantin]

Careful, my friend, Samsung and Crucial have recently been caught doing the same thing with their SSDs as WD, AData, among others. According to ArsTechnica, the Samsung EVO 970 Plus also had a silent controller switch-out, with significant performance impacts. The new packaging is identical to the previously-shipped 970 EVO Plus, just the product on the inside has a new OEM P/N.

Thus, buying a EVO 970 plus is no different than buying a 8TB WD EasyStore back in the day and wondering if you'll shuck a Red or a White label helium drive... except that WD made no warranties re: what would be shipped inside! I'm OK treating a shucked drive as a loot box experience, I am not OK with a OEM making material changes to an extant product that I, and the rest of the market, are willing to pay a premium for due to past reviews, test performance, user experiences, and so on.

Whereas SSD companies in the past upgraded their SSD lines as newer controllers, faster flash, etc. became available, I fear the common trend will become to ship a great product to reviewers and the greater market for a limited time, get the good PR and user testimony from early adopters, followed by "value-engineering" a bait-and-switch later on once the model has established a good reputation.

Should OEMs get caught, they perform a mea culpa, WD style, pledge greater transparency in the future, and offer to replace drives for customers who complain. The rest of the consumer base lives with their bait-and-switch. The company, and management in particular, hence profits in the short term, though long-term reputation damage is likely. But management typically cares less about that since tenures are getting shorter and shorter, quarter-to-quarter performance drives the annual bonus, etc.

 

[Etorix]

And the last paragraph explains why the company board will NOT weight on the issue. Better take the benefits while looking ostensibly elsewhere. If an inordinate amount of manure hits an unexpectedly large legal blower, the board will pretend to be totally unaware and blame the chief executives—who will leave with the blame but a nice parachute and all the bonuses they cashed while the scam was working.
Rince, lathe and repeat.

 

[jgreco]

I understand what you folks are upset about, but I still think that this is a bit heavy on the side of assuming deliberate malfeasance. The fact of the matter is, we're in an era of silicon shortages, and if WD's options were to choose between switching to slower flash or to stop producing a product entirely, their decision is reasonable in that light. The WD Blue is a value-line SSD, and switching to a less-great silicon is not ideal, but it may have been their best option to keep selling the product.

Samsung's weird move on the 970 Evo Plus is probably similar, though there they had a number of surrounding products that they could have focused on instead (980, 980 Pro, 970 Evo), and it wasn't a secret that they were having problems with flash controller availability.

In both cases, some transparency or a part number change would have been a better strategy, of course, but racing to assume that this is some deliberate ploy to screw the end user is probably unreasonable.

 

[Diff]

Related to this topic, I have few naive question to experience ZFS/TrueNAS community members:

1. Anybody have experience with WD Gold Enterprise Class SATA Hard Drive ? Not sure I fully understand if this would be something recommended for TrueNAS? It states being Energy Assisted Magnetic Recording (EAMR).

2. On Seagate side, seems like Exos 18 (18TB) drives are very reasonably priced (under $400), but another questions to experienced ZFS users - from experience, do people see a difference in reliability between Exos series on SATA vs SAS? Would be SAS potentially more reliable just with guess this is created more for Enterprise world?

3. Any feedback on datacenter level drives, like Ultrastar DC HC550 (seems like also available in SAS or SATA)?

 

[Spearfoot]

Related to this topic, I have few naive question to experience ZFS/TrueNAS community members:

1. Anybody have experience with WD Gold Enterprise Class SATA Hard Drive ? Not sure I fully understand if this would be something recommended for TrueNAS? It states being Energy Assisted Magnetic Recording (EAMR).

2. On Seagate side, seems like Exos 18 (18TB) drives are very reasonably priced (under $400), but another questions to experienced ZFS users - from experience, do people see a difference in reliability between Exos series on SATA vs SAS? Would be SAS potentially more reliable just with guess this is created more for Enterprise world?

3. Any feedback on datacenter level drives, like Ultrastar DC HC550 (seems like also available in SAS or SATA)?

We used 8 or 16 of the WD Gold 4TB drives at work before moving to larger capacity disks, and I have UltraStar HC520s (SATA) and HC530s (SAS) here at home; all performed well.

I don't have any experience with the Seagate EXOS disks. Many users here on the forum have had good service from them, but many users (myself included) shy away from Seagate after having been burned during their debacle a decade or so ago. And for what it's worth, the Backblaze data seems to show they're not as reliable as other brands. But the Backblace data is based only on certain models; not sure if they have any EXOS disks in use.

I favored HGST disks for along time, but now that HGST has been fully amalgamated into the Western Digital Borg? I'm at a bit of a loss.

 

[Constantin]

In both cases, some transparency or a part number change would have been a better strategy, of course, but racing to assume that this is some deliberate ploy to screw the end user is probably unreasonable.

We will have to agree to disagree on that one. After the SMR debacle, WD pledged transparency to its users.

We’re committed to providing the information that can help make an informed buying decision for as many uses as possible. Thank you for letting us know how we can do better. We will update our marketing materials...

Well, once again, they didn't. Like you I don't have an issue with an OEM making part switches, as long as those part switches are fully disclosed. Release a new box, new info, alert the market. "Hey we are releasing a downgraded flash drive because we can't get parts". Or whatever. But don't keep shipping the same exact box, logo, whatever, with crummier stuff on the inside.

Both WD and Samsung have segmented the daylights out of the SSD market with multiple price points, alleged performance specifications, and so on. As such, it would be trivial for them to put out another lower tier that doesn't live up to classic WD Blue or 970EVO+ performance. Then as the current gastrointestinal distress in the SSD supply chain abates, they can either retire the new entry-level lines or keep them.