RN424 , 4x 8TB WD Red Plus failing after 120 days?

---

elaplace wrote:

 

In my machine I use the red drive, purple drive, black drive, and Seagates considering of CMR not SMR, and NETGEAR recommendation page:

 

https://kb.netgear.com/20641/ReadyNAS-Hard-Disk-Compatibility-List

 

---

Although Netgear will qualify desktop drives, I don't recommend them.  One reason is that over time the manufacturers will change them.  Barracudas for example are now SMR, but they weren't a few years ago. NAS-purposed or Enterprise (excluding those pesky SMR Reds) are safe choices, even if they aren't on the HCL (which tends to lag new drive introductions by many months).

---

elaplace wrote:

 

In my machine I use the red drive, purple drive, black drive

 

At last  in my budget I prefer to the purple drive and Seagates.

---

FWIW, I used a Black drive (quite a few years ago now), and found it was quite noisy.  But I'm sure the drive tech has changed since then.  In any event, I do check the acoustic and power specs when I am considering a new disk model.

Purple (surveillance) drives are tuned for write performance when streaming, so read performance might be somewhat lower than NAS-purposed.  They should work reliabily though, and are unlikely to be SMR.

Budget is of course a consideration. I do look at prices when I purchase new drives - my most recent purchase was an on-sale Seagate Exos which was actually cheaper than Ironwolf or WD Red Plus.  Sometimes there can be a big delta between Seagate and WD, though usually the gap is pretty narrow.

Hello Stephen B, thank you for coming back to us; @ elaplace, thank you for contributing.

Yesterday (Thursday) I carried out a low-level formatting on a 75Gb HD I took out of another PC. This morning (Friday) I inserted that disk in Slot 1, powered it up, and the RN424 created a volume as RAID1 - JBOD (see fig 1). I then changed password and NAS name, powered it down, placed the disk in Slot 2, powered it up and it tested fine (see fig 2). Ditto when insterted in Slot 3 and Slot 4. I then returned it to Slot 1… and the RN424 went incommunicado, the page timed out on three different PCs for 4 hours (but this, although strange, may not be relevant – see belo).

As the test has been successful, I have since removed the test disk. I returned the original 8TB disks in their locations (Slot 1, 3 & 4). The RN424 has NOT been repowered yet because a) I wished to be cast iron positive that it was the right thing to do without endangering the array, and b) I wish NOT to use the RN424 in a degraded state, so long I have space on the other NAS, I will try to avoid using it.

Tomorrow I will try to contact the supplier of both RN424 and see what they can do about the failed disk. [[ Update – DONE – but they are closed for the weekend ]]

If I am correct, once I have a replacement 8TB, I should

1. Power-up the RN424, and I should see the full array in a degraded state
2. “hot insert” the replacement disk in Slot 2, where it belonged
3. Wait ~24/48h until the 14TB array is fully rebuilt and the Volume return to a “redundant” state

With regards to the RN424 going incommunicado, I just realised that it was using both NIC on a bonded connection (CAT6 to a GS324T); but after the  factory default, the NIC were not bonded anymore and maybe after a reboot I did not check the IP address (which until then never changed, it was always 192.168.2.63, I was filming every reboot during the Slot 1,2,3,4 tests)

StephenB on the reply to elaplace’s comment

“FWIW, the WD80EFAX is a WD Red Plus drive, not a WD Red”

Apologies, well spotted, TY for the correction.

I am awell aware of the difference between CMR  and SMR, I returned and swapped two HDs on another NAS.

“AFAICT we are seeing just one drive failure here, not multiple.  Usually when I see multiple early failures I am thinking there might have been damage in storage or shipment.  But failures at ~120 days are somewhat unusual - in my experience they either fail out-of-the-box or run quite a bit longer.  In any event, they can and do fail at any time (which is why you need backups).”

Three points: a) all the 4 disk have the same date of manufacture (22Dec 2020) and very close SN (VGJKAN6G, VGJLL44XG, VGJJLL4G, are the three good HDs, and VGJL3SDG, is the failed one);

b) the “mode” of failure caught me by surprise, in all the failures I had in the past, the disk started showing errors which ramped up – sometimes very quickly. Not in this case, it seemed to have been a CATASTROPHIC and sudden failure. I also find odd that the WD diagnostics did not even SEE the disk

c) backups – YES I need a backup strategy. But with 25TB (and growing) original data, even a new RN214 with 4 x8TB would not be enough for the first backup (and then there would be incremental), I really do not know which startegy is open to me. Any suggestions welcomed. Also, I don’t think I really understood “snapshots” and where and when use them.

elaplace

“The NAS players around me accept it very early, and are not friendly to the WD RED”

…”not friendly”? I am not sure what you mean…..

I mean more RED drives used and more complains to it than other drives on the forum.

---

elaplace wrote:

I mean more RED drives used and more complains to it than other drives on the forum.

---

FWIW, I haven't seen more complaints about WD Red Plus drives than other models here.  There have certainly been complaints about SMR drives (both Seagate and WD) over the years.

---

berillio wrote:

 

c) backups – YES I need a backup strategy. But with 25TB (and growing) original data, even a new RN214 with 4 x8TB would not be enough for the first backup (and then there would be incremental), I really do not know which startegy is open to me. Any suggestions welcomed. Also, I don’t think I really understood “snapshots” and where and when use them.

Larger drives are of course available, and cost-per-TB is approximately level up to 14 TB models at the moment.  Using JBOD on the backup NAS is another option, though you do need to shift shares around occasionly to make sure you have enough free space on all volumes.  And expansion is a pain, since you need to destroy a volume, and then re-do the backup.

My volume isn't as large as yours (I have about 14 TiB at the moment).  I do use other ReadyNAS for backup (I happen to use more than one).  One is an RN202 that uses JBOD (a 14 TB drive and a 6 TB drive at the moment).

---

berillio wrote:

 

 I then returned it to Slot 1… and the RN424 went incommunicado, the page timed out on three different PCs for 4 hours (but this, although strange, may not be relevant – see below).

...

With regards to the RN424 going incommunicado, I just realised that it was using both NIC on a bonded connection (CAT6 to a GS324T); but after the  factory default, the NIC were not bonded anymore and maybe after a reboot I did not check the IP address (which until then never changed, it was always 192.168.2.63, I was filming every reboot during the Slot 1,2,3,4 tests)

Loss of bonding could have been a factor, especially since the switch was still assuming a LAG. That's not a scenario I've tested.

It might be good to grab the log zip file from that setup, if you still have that disk available.

---

berillio wrote:

 

Tomorrow I will try to contact the supplier of both RN424 and see what they can do about the failed disk. [[ Update – DONE – but they are closed for the weekend ]]

 It sounds like you purchased the NAS with disks installed - is that the case?   Were the disks provided by Netgear or the reseller?  Either way, you should contact the seller before trying to exchange the disk with WD.

---

berillio wrote:

 

If I am correct, once I have a replacement 8TB, I should

1. Power-up the RN424, and I should see the full array in a degraded state
2. “hot insert” the replacement disk in Slot 2, where it belonged
3. Wait ~24/48h until the 14TB array is fully rebuilt and the Volume return to a “redundant” state

 

---

 Correct.

---

berillio wrote:

the “mode” of failure caught me by surprise, in all the failures I had in the past, the disk started showing errors which ramped up – sometimes very quickly. Not in this case, it seemed to have been a CATASTROPHIC and sudden failure. I also find odd that the WD diagnostics did not even SEE the disk

I have seen this sometimes.  If the controller circuitry on the disk fails, then it can be catastrophic - a different failure mode than the more common one of running into increasing bad sector counts or timeouts.  A significant percentage of drive failures happen with no warning whatsoever (36% in a rather old Google study, the numbers might be somewhat better now).

stephen,

“It might be good to grab the log zip file from that setup, if you still have that disk available.”

I do, just DL the zip with the logs

When I tested the 424 with the spare disk, I changed the NAS name and I can see on the log page. You can see the consecutive shutdown when I was swapping slots (in the pic)

But I did not change the time to UK, all the entries started @3:46 am, that was probably 9:46am

And I also have the log of the 424 after it failed from which you could see the bonding I had ( cannot remember now ).

Incidentally, I spent quite a bit of time tying to get the double bonding sorted out, without any success. I had the bond (partially) working on the 424, meaning that I had no increase in performance from the the single NIC; the 214s, instead, would simply drop out of line within half a minute the moment I set it, so I am using a single NIC.

And that was while following all the advice I could find, from  three or four posts (mostly yours) – but it never worked.

It is all documented somewhere ( right now I don’t remember any of that ).

I planned to start a new discussion trying to get to the bottom it, but I never had the time.

---

berillio wrote:

I do, just DL the zip with the logs

 

---

You can send me a PM (private message) using the envelope icon at the top right of the forum page, with a download link to the logs (google drive, dropbox, etc), and I could take a look at it.  Don't post the download link publicly.

---

berillio wrote:

I had the bond (partially) working on the 424, meaning that I had no increase in performance from the single NIC

---

What bonding mode are you using?

It's hard to see any consistent gain in performance unless you have a lot of clients accessing the NAS simultaneously.  Two often isn't enough.

Just a quick update – I got in touch with the vendor (from where I bought the RN424 and the 4 HDs), they recommend doing the RMA directly with WD, which I did, got the reply with the RMA number and RMA Date: 08/11/2021 03:51 PM) and sent them my failed disk that afternoon, WD should have received next day.

Update #2  WD dispatched a replacement disk on 21 August (Saturday morning @5:02). unfortunately UPS failed to recognise the tracking number during the weekend, and also on Monday morning, but on Monday afternoon I received more mails following the despatching, with delivery expected by the end of 24 August.

Update #3 Received the replacement from WD on Wednesday 25 August 2021~9:00 am

I think I have read somewhere StephenB saying that he tests his replacement HDs, so I put on test using DLGDIAG – Estimated time required ~66h

Update #4 Friday 27 ~19:30. DLGDIAG fails the replacement HD  “Too many bad sectors detected”. I immediately filed a “Question” on the WD Support (could not find any specific way of “connecting” to the RMA).

StephenB, Thank You !!!

Update #5 Wednesday 1 September ~19:05 “Incident Update” reply from WD

“……..I am sorry to hear that your replacement device was faulty well; relevant teams are informed about this issue and we will continue assisting you until we provide a solution.

It may take time to hear back from relevant teams about how to proceed but we will do our best to conclude this issue as fast as possible. ………”

Update #6  Thursday 2 September 18:05 (UK time)

On 30 July , when the incident occurred, the situation was:

RN424 has ~14.03TB used and 7.74TB free.

RN214a has ~7.71TB used and 3.19TB free

RN214b has ~4.22TB used, 4.97TB of snapshots and 1.77TB free

On 2 September, the situation is

RN424 – unprotected volume, NAS is offline. I did not copy/move any conspicuous amount of data, just  the end of the folders in use, so the usage can be considered to be the same as above

RN214a – 8.35TB data, 138GB snapshots, 2.42 TB free of 10.90TB

RN214b – 5.73TB data, 955.54GB snapshots, 4.24 TB free of 10.90TB

Update #7 Monday 4 October (yes, it has been a month)

WD is still trying to sort out a replacement disk for me – there have been some wishful emails on their part expecially because my spare NAS has gone to 30% “space left”, then to 20% “space left”.

But I cannot wait any longer so I bought another 8TB WD Red Plus, CMR 7200rpm (as well as UPS, which was needed)

Update #8 Tuesday 5 October – Parcel delivered Hurray…..

Unfortunately I just checked and the WD80EFBX (supposedly the replacement for the WD80EFAX) is NOT in the compatibilty list…….

The compatibility list is a guide, and has not been updated in a while.  The WD80EFBX is completel;y compatible -- one of my NAS has 4 of them.

Brilliant, thanks, pfeeww

I'll power the 424, which has been offline since the event, and "hot insert" the new disk.

Cheers !

Hi all, sorry for the delays

Tue 5 Oct at 21:55

Disk Model:WDC WD80EFBX-68AZZN0 Serial:VRGRW69K was added to Channel 2 of the head unit.

Tue 5 Oct at 21:55

Resyncing started for Volume data.

Wed 6 Oct at 15:30

Volume data is resynced.

Wed 6 Oct at 15:30

Volume data health changed from Degraded to Redundant.

Pfeww Hurray

Thank you to all for the support

Added note – WD has not replaced the failed disk yet.

When I will receive it (if I do), I could get another NAS, maybe a 2 bay one, and I could twin it with another 8TB on RAID 1, purely for backup duties.

Unfortunately NETGEAR seems to have disappared from the UK market, for reason I cannot really fathom.

One simple question – in the event of a NAS failure – can a RAID1 disk be read by any LINUX machine or would I need recovery software (which is kind of pricey)?

---

berillio wrote:

in the event of a NAS failure – can a RAID1 disk be read by any LINUX machine or would I need recovery software (which is kind of pricey)?

---

Yes.  You would need to install mdadm and btrfs on the linux system.  But no paid software packges are needed. 

@@Stephen B wrote:

Yes.  You would need to install mdadm and btrfs……….

Thank you :-)

Updates

Wednesday 13 October 2021 ~ 15:30

UPS parcel, WD80EFAX  S.N. VGGW8JKK ; 256 MB cache 5400 rpm (? I thought it was 7200rpm..?)

Disk goes on test with DLGDIAG, Quick test  Passed, starting Extended test (on my USB system ~60-70 h)

Thursday 14 October 2021 14:02 – Too Many Bad Sectors (after 22h10m, 44h49m remaining)

Mailed WD with the screencap of DLGDIAG

Did somebody put a Jinx on me ???

   :-(  :-(  :-(

---

berillio wrote:

WD80EFAX  S.N. VGGW8JKK ; 256 MB cache 5400 rpm (? I thought it was 7200rpm..?)

 

---

It is 7200 rpm.  https://www.pcmag.com/news/western-digital-confuses-everyone-with-5400-rpm-class-hard-drives

Note after this came out, WD put 7200 rpm on their datasheet (with a footnote talking about the performance class).  Not sure if they updated the firmware to report the actual speed.

---

berillio wrote:

 

Mailed WD with the screencap of DLGDIAG

Did somebody put a Jinx on me ???

---

Ouch.  That possibility is why I also test mine before putting them in my NAS.