Did Western Digital did it again ?

[ChrisRJ]

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 have 8 Exos X16 16 TB SATA in a RAIDZ2 (bought in 1+5+2 batches in September/October 2020) and 1 spare (bought in April 2021, right before the prices soared). Out of the first 8 drives 3 have died since May with 6000-7000 hours of running and less than 20 power-cycles. The RMA process with Seagate here in Germany has always been flawless so far. I had sent the faulty back on Saturday with DHL (they are also the USPS equivalent here) and received the replacement via UPS on the following Wednesday or Thursday. Without a maintenance contract, that is pretty decent service in my book.

Regarding the Backblaze numbers, all I found was about just 60 drives and therefore seems to bear little significance. In my system, all drives were in different slots, so I can hopefully exclude the SATA/SAS ports. As to power I will have to double-check whether there are "overlappings" in the paths.

 

[Evertb1]

Out of the first 8 drives 3 have died since May with 6000-7000 hours of running and less than 20 power-cycles.

That aren't encouraging figures. No mather how good the RMA process works, that is a to high failure rate to be happy about. Good for you that the RMA is OK but still.

 

[joeschmuck]

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.

That is unfortunate but my relationship with Samsung products has been quite good (one exception: Refrigerator with a digital inverter compressor vibrated like crazy, thus very noisy, quiet for the first few months, noisy after the first year) so I will do my research and try to make good choices but until I am delivered a raw deal, I'll stick with Samsung SSD's for now. I personally have the 850 EVO and 860 EVO drives in my systems, they are kind of old in computer terms but they are so much better than spinning rust for my needs.

 

[HoneyBadger]

When it comes to NAND, I've never had Intel bait-and-switch a drive on me.

 

[Constantin]

That is unfortunate but my relationship with Samsung products has been quite good (one exception: Refrigerator with a digital inverter compressor vibrated like crazy, thus very noisy, quiet for the first few months, noisy after the first year

VSDs are the next big thing in fridges, right up there with VIPs. Sorry to hear yours misbehaved!

 

[JohnAtl]

Does it mean that drives advertised as "5900" or "5400" can actually be 7200 RPM?
(E.g, if smartctl tells me "Rotation Rate: 5400 rpm", the drive may in fact be spinning at 7200 RPM?)

More like monitor size classes. That is, a 24” class monitor could be 23.9”, but not 32”.
if a drive spins at 7200 rpm, they would just sell it as 7200 rpm for more money.
As I recall the whole “class” thing was due to someone (Apple?) being sued because their actual viewable area was slightly smaller than they billed it as.

So maybe a drive runs at 5350, and to keep people from losing their mind over something that doesn’t matter, they say 5400 rpm class.

 

[Etorix]

In this case, the drives measurably spin at 7200 rpm but WD marketing passes them as "5400 rpm class", to the dismay of those who bought the drives hoping they would be somewhat quieter and cooler than the 7200 rpm products from WD competitors.

Western Digital is trying to redefine the word “RPM”

Well those measurements of rotational speed are just, like, your opinion, man…

arstechnica.com

 

[Constantin]

That’s the issue with 5,400 RPM class drives: the performance of a 5,400RPM drive combined with the higher heat, lower life expectancy, higher power consumption, and higher noise of a 7,200 RPM drive.

In other words, the worst of all worlds - a 7,200 RPM drive that has been artificially knee-capped to perform like it’s slower-rotating siblings. Could be as simple as adding a few “delay” functions to the firmware.

So much for WD pledges for more “transparency” in the wake of the Red SMR Mea culpa. What makes this particularly egregious is that it affected the Red line again and per that article the firmware has been hornswoggled to report the wrong rotational speed!!!

This is why I disagree with @jgreco re: intentionality at WD re: these issues. One doesn’t make CMR->SMR changes lightly, just as one does not intentionally make firmware report an incorrect rotational speed / invent “5,400 RPM class”, just as one does not change the contents of the box without posting a notice in advance. There is a pattern of deception at WD that goes well beyond the usual levels of marketing baloney.