Problems with WD RED WD40EFAX

Currently reading
Problems with WD RED WD40EFAX

7
1
NAS
DS1621+
Operating system
  1. Linux
Last edited:
Hey guys,

I had three WD 4TB RED (WD40EFAX) drives in a self-built Linux file server. I replaced these into the file server for older drives that were failing.

Immediately I was getting occasional CRC errors in dmesg, especially during RAID rebuild or high-load operations. These errors were coming from all three drives... So, figured this was the file server on its way out (it's pretty old) and decided to go for a Synology DS1621+ to put them in, along with three more of the same drives, instead.

During the initial RAID build, I am seeing similar errors from the synology box, from *all six* drives. These look like:
- "Read abnormality (UNC error) is detected on Drive x"
- "Bad sector was found on disk[x]
(where x is a number between 1 and 6)

I have run an *extended* S.M.A.R.T. test on all six drives (the one that takes ~200 minues,) and they don't show any errors - and the synology box has continued building the RAID and is not showing it in a bad state.

I have heard bad things about these EFAX drives but surely they are not so bad that 6 identical drives would all be unusable? So, how much should I worry about these errors?

I am pretty sure that the synology is not bad, as the drives were doing the same thing in my old Linux box.

The drives aren't on the _official_ compatibility list for the Synology, but I figured I could get away with it. Have I just cost myself £600 in unusable storage?

Any advice on what I should do here?
 
Current WD40EFAX drives are SMR technology. Generally CMR is preferred in NAS use, but SMR is fine as long as they aren't mixed with CMR drives. Complicating this is that early WD40EFAX drives are CMR.

The first step is to test each drive on a PC using WD's Data Lifeguard extended test (allow 2 days per drive).
 
Thanks for the advice. All drives are identical, so theoretically they should all be SMR. Is there any way to be sure? (They are all generating errors, so I'm guessing so.)

That's a lot of testing! Can you test several drives at once or will I have to spend 12 days testing drives?

Does it have to be a Windows PC? I don't have one of those, just Linux, but I can't see a Linux version of Data Lifeguard on the WD site.
 
> Just a thought for consideration: Are all six drives running the same firmware version? Is it the most recent version?

Ah, good spot. Four are 82.00A82, and two are 82.00A83.

All six drives are reporting errors, but maybe they'd be happier if they were all on the latest firmware. I'll look at that.
 
UNC error means- a read attempt from the drive is damaged and can’t be fixed by ECC.
CRC error means - an issue between host and the HDD
and finally the bad sectors, .... so it’s a deadly mix of errors.

So when you have some WinOS desktop, try HDSentinel for a complex proof of the issue:

it will show you definitely more than “tuned” smartctl in Syno DSM GUI
-- post merged: --

to be sure unplug the HDDs from the NAS and plug them directly (USB/SATA adapter) to the PC for an advanced features
 
Thanks for the advice. All drives are identical, so theoretically they should all be SMR. Is there any way to be sure? (They are all generating errors, so I'm guessing so.)
They may or may not be SMR. WD submarined SMR drives into their 'Red' line starting a couple of years ago. So if your drives are several years old, there's a chance that they're CMR. Do they have similar serial numbers? I won't mention that it's IMO a bad idea to have all your drives out of the same production run, but you may only need to test one.
If they are within the warranty period, start the return process with WD. I've read of a number of people who got Red Plus drives as replacements (Red Plus is what Red used to be, definitely not SMR).
If you're actually stuck with SMR Reds, the issue is write speed. If you are going to be mostly reading them, the major issue is getting RAID sets built (lots of writing - and potential for timing out). I've run across a couple of reports of SMR working in RAID 1, so probably RAID 10 might also work. But don't even try if you're going to be writing a lot (e.g. VMs).
TLDR: get WD to replace these if you at all can. They're in the wrong here, and they seem to know it.
 
there is a much more accurate method of comparing two same PN disks for the same S/C Magnetic recording technology, explained in another thread:


otherwise you will just shoot into dark with an assumptions. And assumption are the mother of all Fups.
 
Thanks for all the advice!

> They may or may not be SMR. WD submarined SMR drives into their 'Red' line starting a couple of years ago. So if your drives are several years old, there's a chance that they're CMR. Do they have similar serial numbers? I won't mention that it's IMO a bad idea to have all your drives out of the same production run, but you may only need to test one.

None are older than about 9 months.

> try HDSentinel

I don't have access to a Windows computer, currently... I only have Linux here.

> there is a much more accurate method of comparing two same PN disks for the same S/C Magnetic recording technology,

Thanks - I took a look with hdparm and they all look the same to me, so I think that they're all SMR.

> the major issue is getting RAID sets built

Yeah, it took 5 days or so to get a RAID6 built. 🤦

> I've read of a number of people who got Red Plus drives as replacements

Right - this looks like plan A. Fingers crossed I can manage that. It is ridiculous that they badge this line as _NAS_ drives.

The disappointing thing is that I've always been a loyal WD customer at home and at work and this debacle is putting me right out. Normally their stuff has been really good. (Or, at least "not-great-but-cheap and you knew that when you bought it" in the case of Green)

> TLDR: get WD to replace these if you at all can. They're in the wrong here, and they seem to know it.

Thanks - I've had a ticket open with WD since my original post. It takes them a few days to reply, but they've now offered to RMA all six drives after a few emails. They originally asked me to run their Windows diagnostic software, but as I can't they said they would go ahead and replace, as I've confirmed that they are reporting similar errors in two different machines.

I haven't had them confirm what they will replace them with though. The tech I've been dealing with seems cagey to engage on the whole SMR discussion.
 
- to be sure - HDSentinel is running also in Lnx (Ubuntu for me)
- but extensive diagnostic and data repair tools you can find only in the Win version of this great Swiss knife for disk check/repair (relocation of data from bad blocks, identification of the exact data written in the bad blocks, ...)

Just check price of Seagate Exos and you will be surprised from the price/performance vs Red or IronWolf
 
> diagnostic and data repair tools you can find only in the Win version

Thanks - I took a look and I think the Linux version only prints out stuff you can already find from hdparm and smartctl. Doesn't look like it will do anything fancy.

> Just check price of Seagate Exos and you will be surprised from the price/performance vs Red or IronWolf

I would prefer not to have to spend more money at this point having just stumped up a load for a fancy Synology box and a bunch of drives, but thanks for the recommendation - will check this next time.

Have you used their RMA process? One of the things I always liked about WD was that you could always RMA your in-warranty drives easily with no fuss, with the new one sent out to you first so you can return the old one in the same packaging.
 
> Just check price of Seagate Exos and you will be surprised from the price/performance vs Red or IronWolf
Instead of copy/paste parts of a post and adding your own '> ' prefix it's easier to do thing when replying to part of a previous post. It's easier to read too.
1618305827395.png

I would prefer not to have to spend more money at this point having just stumped up a load for a fancy Synology box and a bunch of drives, but thanks for the recommendation - will check this next time.

You can do this as many times as you like and with different posts. They get placed at the current insertion point.

Have you used their RMA process? One of the things I always liked about WD was that you could always RMA your in-warranty drives easily with no fuss, with the new one sent out to you first so you can return the old one in the same packaging.
No-fuss is nice!
 
Follow-up on this:

I've now finally got the RMA sorted out. They've recieved my drives and sent out WD60EFRX drives to replace them - which is nice as it nets me more space over the WD40EFAX as well as using the proper technology.

It helped that I didn't have a Windows system to test them on I think, as I couldn't use any of their recommended diagnostic software so the incident got escalated more quickly. Either way, it seems like if you hassle them then they will acknowledge that EFAX are not fit-for-purpose and will do something about it. - They pretty much acknowledged that if you're trying to do Raid 6 with EFAX then you're gonna have a bad time.

YMMV of course.
 

Create an account or login to comment

You must be a member in order to leave a comment

Create account

Create an account on our community. It's easy!

Log in

Already have an account? Log in here.

Similar threads

On DSM6.2 several of My USBdisks where nicely mounted. DSM7 refuses to mount these. Disks where...
Replies
0
Views
1,572
I'm using Firefox for years and I'm still very happy with it. Works fine with my DSM, PHPMyAdmin...
Replies
1
Views
2,069
Thank you for the quick reply and sharing your experiences. Both my original 2016 WD RED drives (8 years...
Replies
2
Views
744

Welcome to SynoForum.com!

SynoForum.com is an unofficial Synology forum for NAS owners and enthusiasts.

Registration is free, easy and fast!

Back
Top