Buyer beware?

[GreyGeek]

[jglen490]

Always.

I bought four WD Reds for a non-profit I work with, to use in their Synology NAS. When that news came out, I was mortified. But after considerable research I found that I had bought a slightly older model of the drives (WDnnEFRX), and that none of that model are SMR, they are all CMR regardless of capacity. Good news. Lesson learned.

Still mortified that a reputable company like WD (and others) would tap-dance around spec sheets. It's their reputation that's on the line, and lying should not be an excuse to bolster bottom lines. Yeah, I know, they do it all the time.

[oshunluvr]

Interesting read but not really much of a real-world problem. Frankly, I haven't seen any reports that this issue results in data lose or drives failing prematurely. It's not like your hair is all-of-a-sudden on fire because you got one of the SMR drives. Most of the comments are just much adieu about nothing. I have 4 WD Red drives (none are SMR drives for whatever that's worth). Honestly wouldn't impact me or most anyone if you're actually using them on a small server or NAS - which is what these drives are marketed for. The network is slower than the drives anyway. Life expectancy is more important the seek times in a home server or NAS. The only real issue IMO is the consumer right to be informed of the technological differences and to get what they're paying for. However, if you're running a data center this might be a real issue. The only notable thing is WDs (other drive makers are doing the same thing BTW) comment that they did it to lower the per MB cost of the drives, but then they didn't pass that to the consumer. But since some other drive makers are doing the same thing, are you going to take your 1 year-old SMR drives and throw them away because WD hurt your feelings? That's really dumb, but hey - I'll pay the shipping and you can send them to me.

I've had very good experience using WD drives - none have failed prematurely (within warranty time) ever. I have retired them to upgrade way more often than having one die. My two "retired" drives I use as backups have 50,000 (2TB Black) and 70,000 (2TB Red) hours each with zero errors. BTW, 70,000 hours is 8 years of power-on-time for a drive warrantied for 3 years. I got my monies worth out of that drive, right?

In the olden days, I bought the best deal on a drive because the prices varied a lot. I've had Hitachi and Seagate drives die early and never a WD so I swore off the other brands. That was a long time ago. When it's time for a new drive, I will probably do my research and buy the most reliable because performance and cost per MB aren't all that far apart anymore brand to brand. My last purchase was the WD 12TB Red Pro which I choose over the non-Pro version because it had a 5-year warranty vs. 3 for the $50 cheaper non-Pro drive.

[GreyGeek]

My favorite drives for years was Toshiba. Unfortunately the two of the last three drives I purchased failed within 1000 hours. One is invisible to my hardware. It simple refuses to be seen by the BIOS during boot. The other has so many bad sectors that I've taken it off line because I no longer trust it.

[jlittle]

but not really much of a real-world problem.

There's been reports of users with NAS in some RAID configuration replacing a drive, and the rebuild being very slow, and even causing it to fail, because the RAID software decided the new drive had failed too. These reports supposedly started the controversy.

I have some skepticism; the story sounds like someone speculating about the consequences of SMR drives in RAID. But it's a real-world scenario IMO.

[oshunluvr]

There's been reports of users with NAS in some RAID configuration replacing a drive, and the rebuild being very slow, and even causing it to fail, because the RAID software decided the new drive had failed too. These reports supposedly started the controversy.

I have some skepticism; the story sounds like someone speculating about the consequences of SMR drives in RAID. But it's a real-world scenario IMO.

I did see a report claiming that if your NAS uses ZFS this could happen. Basically drive response becomes so slow during the rebuild that the file system decides the drive is no good and stops the rebuild. For sure THAT's a potential real-world problem.

I only every used RAID1 and BTRFS on my server, now no RAID at all. Honestly, unless I'm just incorrect, there's not much (or maybe any) advantage to using RAID on a small NAS. You get no performance boost at all because you access the NAS/server via a network and it's slower than the hard drive. I guess there is a storage advantage to using levels 5/6 vs. JBOD or RAID1 but also the increase in potential loss. If your goal is full time access with minimum risk of loss, isn't RAID1 the best choice?