WD publishes complete list of SMR drives following user backlash

[Humza]

In context: WD, Seagate, and Toshiba were recently found - and acknowledged - using slower shingled magnetic recording (SMR) technology in some of their hard drives, without mentioning the spec in their product documentation. Among these models, WD Red drives particularly stood out, as they were being marketed for NAS consumers and caused several buyers to experience RAID rebuilding issues. The company is now being more upfront about the issue and has published the complete list of drives using SMR to clear out the confusion and help customers in deciding with their purchase.

Using SMR technology instead of conventional magnetic recording allows HDD manufacturers to put out denser drives at cheaper prices. While that also sounds like a win for consumers, these drives are not suitable for NAS environments due to their slow write speeds, with SMR being particularly concerning if not disclosed in the product documentation.

Among the recently discovered companies using SMR without informing customers, it was easier to overlook Seagate and Toshiba models as they weren't being marketed as NAS drives. Seagate, for instance, recommends its CMR-based IronWolf / IronWolf Pro series for NAS customers.

Western Digital, however, was found using the slower tech in its Red NAS drives (2TB-6TB) and soon issued a response on its official blog, an excerpt of which reads:

WD Red HDDs are ideal for home and small businesses using NAS systems. They are great for sharing and backing up files using one to eight drive bays and for a workload rate of 180 TB a year. We’ve rigorously tested this type of use and have been validated by the major NAS providers.

We typically specify the designed-for use cases and performance parameters and don’t always talk about what’s under the hood. One of those innovations is Shingled Magnetic Recording (SMR) technology.

The company also notes that its drive-managed SMR (DMSMR) models - (2TB-6TB) Red HDDs - provide optimal performance for small businesses and home customers of NAS, as their data-intensive workloads are "intermittent."

Form-Factor	Storage Capacity	WD Red	WD Red Pro	WD Blue	WD Black	WD Purple
2.5"	500GB or below	-	-	CMR	CMR	-
2.5"	1TB	CMR	-	SMR	SMR	-
2.5"	2TB	-	-	SMR	-	-
3.5"	1TB or below	CMR	CMR	CMR	CMR	CMR
3.5"	2TB - 6TB	SMR	CMR	SMR/CMR	CMR	CMR
3.5"	8TB and above	CMR	CMR	-	-	CMR

Nonetheless, it has now followed up with a complete list of SMR models that should certainly help alleviate concerns and make it easier for users to decide on their next purchase.

Unsurprisingly, SMR is present across several other WD model lineups, and the company says that it will update its marketing materials accordingly and will also share drive benchmarks and their ideal use cases in the future.

Permalink to story.

https://www.techspot.com/news/84973-wd-publishes-complete-list-smr-drives-following-user.html

 

[poohbear]

Even a WD Black model uses it? Thought the Black line was the high-end performance line?! Atleast they came clean though.

 

[veLa]

Transparency is all people wanted.

 

[wired32322]

Was just looking at NAS drives. 8 and 10tb ones they're shipping are SMR too. Can tell by the last three letters of the model. They're still lying.

 

[umbala]

Transparency is all people wanted.

Sure. This is like admitting you cheated after your wife already caught you.

 

[SpatulaCity]

Lame WD, I lost respect for that company due to this. I have been planning an upgrade for my NAS and considered WD red to be my favorite choice. Now, I have trust issues.

 

[veLa]

Sure. This is like admitting you cheated after your wife already caught you.

Yeah, obviously.

It doesn't nullify that people just wanted transparency. It should have been disclosed from the start.

 

[toooooot]

Looong have I waited.

 

[HofyPC]

So what are they doing for people who have been duped by this? Just a hi ho FU?

 

[captaincranky]

I wonder if Seagate is planning the same disclosure?

 

[Aaron Jones]

I really think the article subtitle ought to be "Check to see if you've bought an unsuitable WD model for your NAS".

 

[Zorak]

I really think the article subtitle ought to be "Check to see if you've bought an unsuitable WD model for your NAS".

Yup. I'm saving this article for checking all my disks now and in the future. It's been a while since I saved a HTML page from the web. It's still saving... still saving...

 

[OptimumSlinky]

HDD manufacturers have been scummy for a while. On platter drives, they straight up just stopped publishing whether drives are 5400rpm or 7200rpm, or when they DO publish it, it's not always accurate.

 

[Thrackerzod]

Yup. I'm saving this article for checking all my disks now and in the future. It's been a while since I saved a HTML page from the web. It's still saving... still saving...

You could just save the picture of the chart instead of the entire page.

 

[Stoatwblr]

Just to clarify (I'm the original breaker of this story)
- The A and R in EFRX/EFAX refer to cache sizes (256MB and 64MB respectively), not SMR

The larger drives 8 & 10TB are CMR, not SMR. For a while I thought "A" might mean "helium" until WD cleared that issue up

What they've done and their behaviour to date is utterly inexcusable, but they do seem to have changed their tune and it's probably because they've had their legal liabilities pointed out

Changing a product line without notice such that the replacement item appears identical but can suffer 90-98% performance degradation under conditions such as parity RAID rebuilds or scrubs and WILL suffer 40-60% performance degradation even if you handle the things "just right" is bad enough.

Concealing and denying they've done it, then gaslighting victims is "exemplary damages" territory in a courtroom.

That applies REGARDLESS of whether the product in question is a WD RED (NAS/RAID) drive that won't stay in RAIDs anymore, or a WD Desktop drive which suffers mysterious performance wipeouts.

The issue with unannounced DM-SMR on desktop/laptop drives is this:

- DM-DMR drive performance falls to pieces when faced with lots of random rewrites.
- Most filesystems (other than FAT) write a metadata update back to the disk every time a file is read, updating that file's "last accessed" time

This is the WORST POSSIBLE CASE SCENARIO for a DM-SMR drive.

It can be mitigated by setting "noatime" on a Linux mount, but on Windows, switching off access time updates is an "all-or-nothing" proposition and on all OSes, switching off access time updates on the core OS partition _WILL_ cause problems with backup, maintenance and change detection.

Owners of all drives have a right of redress on this.

 

[abysal]

Just to clarify (I'm the original breaker of this story)
- The A and R in EFRX/EFAX refer to cache sizes (256MB and 64MB respectively), not SMR

The larger drives 8 & 10TB are CMR, not SMR. For a while I thought "A" might mean "helium" until WD cleared that issue up

What they've done and their behaviour to date is utterly inexcusable, but they do seem to have changed their tune and it's probably because they've had their legal liabilities pointed out

Changing a product line without notice such that the replacement item appears identical but can suffer 90-98% performance degradation under conditions such as parity RAID rebuilds or scrubs and WILL suffer 40-60% performance degradation even if you handle the things "just right" is bad enough.

Concealing and denying they've done it, then gaslighting victims is "exemplary damages" territory in a courtroom.

That applies REGARDLESS of whether the product in question is a WD RED (NAS/RAID) drive that won't stay in RAIDs anymore, or a WD Desktop drive which suffers mysterious performance wipeouts.

The issue with unannounced DM-SMR on desktop/laptop drives is this:

- DM-DMR drive performance falls to pieces when faced with lots of random rewrites.
- Most filesystems (other than FAT) write a metadata update back to the disk every time a file is read, updating that file's "last accessed" time

This is the WORST POSSIBLE CASE SCENARIO for a DM-SMR drive.

It can be mitigated by setting "noatime" on a Linux mount, but on Windows, switching off access time updates is an "all-or-nothing" proposition and on all OSes, switching off access time updates on the core OS partition _WILL_ cause problems with backup, maintenance and change detection.

Owners of all drives have a right of redress on this.

So the 6TB EFRX/EFAX have always been SMR? I recently bought a spare 6TB Red and could only find the EFAX with 256MB cache, to replace a failing EFRX in my ZFS raidz. I'm hoping they jive well together.

 

[Stoatwblr]

So the 6TB EFRX/EFAX have always been SMR?

No. EFRX were always CMR
EFAX in 1-6TB are SMR
EFAX in 8TB and up are CMR

Don't put DM-SMR drives in a parity raid. You will wish you hadn't
- there are high chances a WD RED DM-SMR drive will throw a HARD io error during raid rebuild and be kicked out of the raid (faulty firmware)
- raid rebuild speed will take a 90-95% hit if it does manage to rebuild (12-20 hour rebuild now takes more than a week! So do raid checks)
- in a mixed CMR/SMR array, speed of the array will be dictated by the SMR drives. Whatever you're used to, you can forget about it...

 

[Zorak]

You could just save the picture of the chart instead of the entire page.

Ya missed the joke.

 

[ZorracK]

I'm only buying enterprise hdd now like HGST k6000 and Seagate exos 14
working like a bliss!

 

[mbrowne5061]

I wonder if Seagate is planning the same disclosure?

Seagate disclosed from the get-go, and when people caught WD, they went and double-checked Seagate and found everything on the up-and-up.

Honestly, this is feeling like the Intel/AMD brand inversion. I had been losing faith in WD for a while because of their business practices, and this just kind of seals the deal. Meanwhile, Seagate's reliability and performance has been improving to match WD's, and they seem to have much better business practices overall.

 

[captaincranky]

@mbrowne5061 The last 2 drives I've gotten from WD have been bad.

I buy drives ahead of time when they're on sale, to keep for later

The 1st one, a 500 GB Raptor, has a chunk out of the plastic frame. (That could have happened at Newegg's warehouse)

The 2nd one, a 1 TB blue, failed to format while installing Windows.

A suspicious 3rd drive, 640 GB WD "Black" was in a surveillance DVR in which the SATA controller failed. Don't know if the drive is bad and caused that, or the controller failed on it's own. I should probably chuck it, since I still don't know "which came first, the chicken or the egg"..

Despite all the complaining going on during Seagate's "bad runs", I've never had one fail or be DOA.

Their 3 TB drives, (with which everyone seemed to have problems), are still 7200 RPM. I only have one of these, and again, it works fine.

In fairness to WD, the drive from my 2005 eMachine was still running strong when I pulled it after 13 years of service. In hindsight, that was probably stupid, since a 160 GB SATA 1 HDD, is pretty much useless today, save for being a paperweight.

In other news, the Antec 380 watt "Earthwatts" PSU, was still working perfectly as well.