NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
shingled magnetic recording (SMR) hard drive fiasco - inquiring on recommendations
It's been a while since I jumped on the forum but what brings me here is my NAS volume utilization is hovering around 65%. I believe it's good until 80% and then starts to bark at you about storage space and performance? So I'm looking to expand the volume but unfortunately have been reading about how WD, Segate and others have inserted shingled magnetic recording (SMR) drives into their NAS line ups without being up front about it. To the best of my knowledge, the current drives I have which were purchased in 2017 use Perpendicular Magnetic Recording (PMR) and therefore perform equally well for write and read. For reference, my current drives are: Western Digital: 8TB: WD80EFZX & 4TB: WD40EFRX.
That all said, I saw a few articles stating that WD 2-6TB are SMR while the 8TB and above are CMR. I believe CMR is essentially the same as PMR but dont' hold me to it. Here's one of those articles on SMR vs CMR: https://nascompares.com/2020/04/16/your-wd-red-nas-hard-drives-might-be-using-smr-what-you-need-to-know/amp/
In any event, I'm looking at some WD Reds on Amazon and was wondering if anyone has some experience with their 8TB and above RED HDs and can confirm they are CMR. For example:WD101EFAX or WD100EFAX. Per the aricle above, all the Pros are CMR but I rather not spend the extra money just to gurantee a CMR drive.
Thoughts?
45 Replies
Replies have been turned off for this discussion
NASguru wrote:
That all said, I saw a few articles stating that WD 2-6TB are SMR while the 8TB and above are CMR.
WD has said that, so I am thinking that applies to the WD20EFAX, WD40EFAX and the WD60EFAX. Synology shows the WD20EFAX and the WD60EFAX as SMR, but doesn't list the WD40EFAX at all.
NASguru wrote:
I believe CMR is essentially the same as PMR but dont' hold me to it.
CMR is just another name for PMR.
NASguru wrote:
I'm looking at some WD Reds on Amazon and was wondering if anyone has some experience with their 8TB and above RED HDs and can confirm they are CMR.
It's not that easy to tell. If I'm understanding what I'm reading correctly, they use soft mapping to physical sectors (similar to an SSD). And they remap updated sectors into the large cache, and don't necessarily re-write the SMR sections of the disk right away. Basically they have improved SMRs somewhat over the years.
That said, I am running two WD100EFAX drives in my main NAS, and am not seeing any issues with general write speed, scrub times, or balance times. I have four WD80EFZX running in a backup NAS, but no WD80EFAX drives..
So timing is everything, when did you buy your WD100EFAX? From what I read, they started inserting SMR drives into the red NAS line around early 2017. Therefore, it's possible you didn't get a drive using SMR depending on when you purchased it. The WD101EFAX also seems to be the replacement for it but both can still be purchased off Amazon with the WD101EFAX being about $10 more. I was also under the impression you can run read/write programs such as CrystalDiskMark to determine if the drive can write 100+MB/s. Most seem to agree that the SMR drives fall off to around 30MB/s which is the give away. I suppose droping a 30GB file onto the drive would accomplished similar results. I seen some who mentioned the ONLY way to determine is to call WD support and provide them a serial #. I could always order the drives and upon receipt call them to verify. Alternatively and if I trust WD to be honest I could just spend an extra $50-$60 per drive and get the pro version which are supposedly all CMR but then again why overpay if it's not necessary. The other method of verification I saw was to open/destroy the drive and count platters but who has money to burn and even then it's no gurantee the next one will be the same. Anyhow, rant over but was I was hoping for a simple black/white yes/no answer but alas it's a gamble either way.
NASguru wrote:
So timing is everything, when did you buy your WD100EFAX?
April/May 2019
NASguru wrote:
I was also under the impression you can run read/write programs such as CrystalDiskMark to determine if the drive can write 100+MB/s.
Not 100% sure on that. Sequential writes from beginning to the end of the disk could run at near-normal speed. And with soft sector mapping there are a lot of tricks you could play in theory, especially if part of the drive is CMR.
In any event, I did run a full erase on both drives before I added them to the NAS. While I didn't closely monitor the times, I believe it was approximately 24 hours for each (which is ~115 MB/s for 10 TB).
I also have the resilver time in an old log zip.
data resilver 2019-04-28 09:54:41 2019-04-29 17:46:39 completed
This resilvered the first inserted WD100EFAX, and completed in ~32 hours. 6 TB on the WD100EFAX would have been written, and 18 TB of data would have been read on the other three drives (all WD60EFRX). Plus the NAS was in use during the resync.
I haven't noticed any slowdown in performance on sustained writes since they were installed.
NASguru wrote:
if I trust WD to be honest
Seagate is known to have silently slipped SMR into some desktop drives. Though they don't recommend those disks for RAID, I still think that's a breach of trust.
Of course WD has now confessed that they did silently slip SMR into the 2-6 TB Red Drives. But in that confession they explicitly stated that larger Red drives (8 TB and up) used CMR. It's one thing for a company to be silent. It's another to make a false statement.
Personally I'd take WD at their word on the currently shipping drives. Honestly I don't know if I trust them to disclose SMR in new drives they introduce in the future. Their disclosure didn't include an apology or any promise of transparency in the future - instead they just asserted that their SMR Reds are fit for their purpose - which is quite debatable.
Hopefully there will be a lesson learned here for both Seagate and WD. I'll be looking for explicit statements on SMR, CMR (or HAMR) on future models I purchase.
Servethehome just did a nice article on the SMR drives in a RAID:
https://www.servethehome.com/wd-red-smr-vs-cmr-tested-avoid-red-smr/
They also have a table of SMR drives:
WD has a class action lawsuit filed against them:
I've always only purchased enterprise class 5yr warranty drives, even for desktop use. If you look properly you can find them at great prices too, so you can get a really good price/performance ratio without having to deal with 1M MTBF drives with lesser warranties as well as shenanigans.
SamirD wrote:
Servethehome just did a nice article on the SMR drives in a RAID:
https://www.servethehome.com/wd-red-smr-vs-cmr-tested-avoid-red-smr/
Yeah, I caught their youtube video on it: https://youtu.be/8hdJTwaTl8I Though they tested only with ZFS, I think their results apply to btrfs (and FWIW, to ext).
If you do want to use these drives (personally I wouldn't), then you need to be careful to not store files on the NAS while it is rebuilding or resyncing. That combination is what creates the huge performance hit in write speeds.
SamirD wrote:
They also have a table of SMR drives:
That table actually comes from WD, and is available in many other places. Seagate also has put SMR into many of their desktop (and USB) drives. One implication is that you should be very careful about using desktop drives in a RAID array (and also very careful on USB drive shucking).
I remember when drives where more expensive and were made as robust as possible so there was no 'consumer' drive variant. The reason these problems even exist is because the demand has been trying to put a consumer drive in what is otherwise an enterprise role (storage array).
Even the wd red nas drives that every datahoarder falls all over is nothing when compared to the reliability of a true enterprise drive that has design specifications calling for twice the MTBF and even carrying almost twice the warranty. Yes, it costs nearly twice as much, but then you get what you pay for--consumer drives for cheaper in an enterprise application will have higher failure rates and other issues and that's the cost tradeoff.
There was another article done by servethehome on the whole 'shucking' idea that the consumer drive inside was essentially one of the touted 'red' nas drives:
https://www.servethehome.com/wd-wd100emaz-easystore-10tb-external-backup-drive-review/
And while many similarities between the drives were found, people's real-world experiences in the comments showed the true nature of these drives:
While it's never a good idea to decieve your customers, it's also never a good idea as a consumer to try to decieve a company. I'm sure WD has warranteed many shucked drives that otherwise wouldn't have failed in their original intended use. The street goes both ways.
Late to the table with this discussion, but here goes:
In October 2019, I replaced the 1TB Toshiba drives in my ReadyNAS with Seagate BarraCuda 2TB drives, model ST2000DM008. This model is listed on the ReadyNAS compatibility list.Recently, when replacing a drive in another device, I stumbled on the information about SMR vs. CMR technology. This prompted me to go back and check on the specs for the drives that I'd installed in the ReadyNAS. Indeed, this particular model (ST2000DM008) is listed on Seagate's specs page as using SMR recording. However, in their revisions sheet, this wasn't added until May 2020.
I'd be curious to know if this Seagate model was always SMR, and they only fessed up to it in May 2020 by revising the spec sheet and documentation, or if earlier instances of these drives (from October 2019 in my case) used CMR.
Anyone?
I bet if you called Seagate and said you had 2x of these drives, one several years older and the new one is taking longer to rebuild a party drive or some other fake plausable scenario, you could trick them into revealing that the older drive was either cmr or smr.
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
