The short answer: In October 2025, the Berlin laboratory 030 Datenrettung warned of a firmware mapping failure affecting 2–6 TB WD Blue and WD Red SMR drives (WD20EFAX through WD60EZAZ). A drive may click, disappear from the BIOS, or freeze even though its platters, heads, and motor remain undamaged. The problem lies in the translation table between logical addresses and the data’s actual locations on the platters. CHKDSK, fsck, and recovery software cannot repair it; they merely write to the only existing copy of the data. If a drive behaves this way, the first step is to turn it off—not run another test.
Do you own one of these drives? What matters is whether it still works normally
The warning was published by 030 Datenrettung Berlin on 21 October 2025 and concerns selected 2–6 TB WD Blue and WD Red models that use shingled magnetic recording. The laboratory reported that the drive mechanics were undamaged in 95% of the drives from this family that it had examined.
That figure is easy to misinterpret. It is not the failure rate of all drives sold—it is the proportion within a non-random sample. A data recovery laboratory receives drives that are already malfunctioning, not those that have been running quietly in a NAS for years. The only useful conclusion is that, with this group of models, the cause may not be mechanical, and the drive should be handled accordingly.
For a business owner, this creates two very different situations:
- The drive is working without errors. First copy important data to another physical medium, then schedule a replacement. Do not begin with a surface test that takes several hours.
- The drive clicks, disappears, freezes, or has slowed dramatically. Turn it off. Do not run CHKDSK, fsck, a long SMART test, or recovery software.
Which models are named in the warning
The full model number matters—not the WD Blue or WD Red branding on the label. Over the years, multiple designs and recording technologies have been sold under the same product name.
| Series | Model | Capacity |
|---|---|---|
| WD Red | WD20EFAX | 2 TB |
| WD Red | WD30EFAX | 3 TB |
| WD Red | WD40EFAX | 4 TB |
| WD Red | WD60EFAX | 6 TB |
| WD Blue | WD20EZAZ | 2 TB |
| WD Blue | WD40EZAZ | 4 TB |
| WD Blue | WD60EZAZ | 6 TB |
| WD Blue | WD20EDAZ | 2 TB |
| WD Blue | WD40EDAZ | 4 TB |
| WD Blue | WD60EDAZ | 6 TB |
You can find the model on the drive’s label, in the BIOS, in CrystalDiskInfo on Windows, or in storage management within Synology DSM and QNAP QTS. Beware of one trap: if the drive is already clicking or causing the entire computer to freeze, do not turn it on merely to read its model number in an application. Copy the number directly from the drive’s label or from a previously saved NAS inventory.
The reverse also applies: the warning names specific models. It does not mean that every SMR drive on the market has this problem, nor does it automatically extend to CMR models or capacities not included in the list.
Why some of these models were missing from WD’s old list
In April 2020—following the controversy over undisclosed SMR technology in NAS drives—Western Digital published an overview of the internal SMR models it was shipping at the time. Among the 3.5-inch WD Blue drives, it listed the WD20EZAZ and WD60EZAZ. The WD40EZAZ and EDAZ variants were not included.
Their absence, however, does not prove that they use CMR. The document described the product range being shipped at that time, not a complete historical catalogue. The 2025 warning explicitly names the WD40EZAZ and the EDAZ models. For a second-hand drive, therefore, an old capacity-based table is not enough—check the exact model number.
The WD Red EFAX series also had legal repercussions: in 2021, a $2.7 million settlement fund was established in a US class action for owners of the WD20EFAX through WD60EFAX models. The settlement was not an admission of liability, and after the controversy, WD moved NAS drives to the CMR-based Red Plus range. That dispute, however, concerned a different issue from the current warning—slow writes and RAID rebuilds that seemingly never finished. The present issue is a corrupted internal drive map.
What fails when the drive mechanics are still sound
The operating system addresses data logically: it requests block X without knowing which track or which part of a platter contains it. The drive’s firmware handles this translation. With conventional CMR recording, the translation is almost direct—a logical address corresponds to a fixed physical location, except for remapped bad sectors.
Drive-managed SMR does not work that way. Its tracks overlap like roof shingles, so the drive cannot rewrite one without damaging neighbouring tracks. It works around this by first storing incoming writes in a reserved area recorded conventionally, then reorganising them into shingled zones later, while idle. To keep track of where everything is actually stored, the drive maintains its own translation table—a data structure that changes with virtually every write and is stored on the drive itself, in its service area. The laboratory refers to the VeniceR platform and corruption of the so-called secondary translator.
This structure is precisely the weak point. If it becomes corrupted, the drive may have readable platters and functioning heads and motor, yet no longer know where its own data is located. From the outside, this looks like a mechanical failure: the drive stalls while loading the service area, repeatedly recalibrates, moves the heads back and forth, and produces rhythmic clicking. Yet nothing inside may have physically broken in a way that could be replaced with a screwdriver.
This also leads to the key point for data recovery: CHKDSK repairs NTFS structures that sit above the drive. The translator sits below them, inside the firmware, and no Windows utility can access it. The difference between shingled and conventional recording is explained in the CMR and SMR overview.
The symptoms look like a slow computer, not a failing drive
The failure does not begin with a message saying „drive failure." Reported symptoms include:
- copying that slows dramatically or stops completely;
- File Explorer, applications, or the entire system freezing;
- a file that opens once but not the next time;
- a drive that occasionally disappears from the BIOS, DSM, or QTS;
- rhythmic clicking after power-on;
- a NAS marking the drive as faulty or switching the volume to a degraded state.
SMART may reveal another problem, but it cannot reliably predict a translator failure because it monitors different parameters. Clean SMART values are therefore not a reason to keep using one of the named models without a backup.
What to do based on the drive’s condition
| Drive condition | Safe next step |
|---|---|
| Working normally and the data is backed up | Verify that the backup can actually be restored, then schedule a replacement |
| Working, but it holds the only copy of the data | Copy irreplaceable folders to another physical medium first; handle the complete migration afterwards |
| Copying freezes or the drive disappears | Stop trying and turn it off; further full-surface reads may make the situation worse |
| Clicking rhythmically | Turn it off immediately and do not power it on again—not even „just for a moment" to run a test |
| NAS reports degraded or failed | Stop write operations and do not automatically click Repair or Rebuild |
| The data exists elsewhere and has been verified | Retire the drive; do not repair it for continued business use |
What not to do—and why ordinary software is not enough
chkdsk /f and fsck write changes to the file system. They cannot repair the secondary translator, yet they modify the only available copy of the metadata. On a drive returning inconsistent data, this is the fastest way to turn a difficult situation into a hopeless one.
Recuva, TestDisk, and PhotoRec operate above the sectors made available by the drive itself. If the firmware returns incorrect data, stops responding, or cannot locate a block at all, no user-level application can bypass that layer.
Even ddrescue is not a magic button. In experienced hands, it is a useful tool for controlled reading, but it is unsuitable as the first DIY attempt on a clicking VeniceR drive. Poorly configured error retries can keep the drive running for hours without addressing the corrupted mapping at all.
Then there are the classics from internet forums: putting the drive in a freezer introduces a risk of condensation, while tapping or shaking it can turn a firmware fault into a mechanical one. Both worsen the condition in which the drive reaches a laboratory.
When the at-risk drive is installed in a Synology or QNAP NAS
First, record the models, serial numbers, and drive-bay positions in DSM or QTS. Also save a screenshot of the storage pool and RAID group status. Do not move drives between bays—their order is part of the array’s configuration information.
If shared folders are still accessible, minimise writes: stop backup jobs, indexing, antivirus scans, and scheduled extended SMART tests. Copy data to independent storage in order of importance—accounting records, work in progress, and operational data first.
An automatic rebuild involves prolonged, continuous reading from every remaining drive and writing to a new one. If another array member is reporting errors or is the same at-risk model, the rebuild itself may turn a recoverable situation into a complete array failure. This is explained in detail in why a rebuild can kill a RAID array.
What to replace an at-risk drive with
For a NAS, choose a specific CMR model from the NAS manufacturer’s compatibility list—the list published by Synology or QNAP, not a retailer’s general recommendation. Current options include WD Red Plus and Red Pro, Seagate IronWolf, and Toshiba N300. The product-line name alone is not enough: verify the exact SKU, capacity, sector format, and support for your specific NAS model. The differences between particular product lines are covered in the comparison of WD Red, IronWolf, and Toshiba N300.
Even a new CMR drive is not a backup. A NAS needs a separate copy that can be restored after an array failure, accidental deletion, or ransomware—meaning a copy that is not connected to the same network as the system it backs up.
How much recovery costs and how to recognise a fair quote
Sound mechanics do not mean inexpensive software-based recovery. The work may require access to the service area, bypassing the corrupted translator, controlled reading, and subsequent reconstruction of logical sectors from whatever can be recovered. Internal encryption complicates the process in some variants.
A fair price therefore cannot be determined from a WD60EFAX model number alone. Before approving a quote, ask:
- whether the price is based on diagnostics of your specific drive or simply on a capacity-based price list;
- what exactly will count as successfully recovered data;
- whether you will also pay for recovering files you do not need;
- whether the company performs the recovery itself or sends the drive elsewhere;
- whether a working copy is created first and the original is protected;
- how confidentiality, data transfer, and subsequent deletion are handled.
If you do not want to risk working on a clicking drive—and when it contains the only copy of business data, that is a sensible decision—we will diagnose it first and give you the price in advance. How we calculate the cost of data recovery also explains what you will not be charged for.
Status as of 28 August 2026
The warning dates from October 2025. As of 28 August 2026, we had not found a conclusive technical analysis or a corrective firmware update for the named drives in publicly available materials.
Firmware availability for internal drives varies from model to model, so you cannot plan on a fix becoming available for your particular drive. Even if one were released, flashing an update to a faulty drive is a write operation. It should not be attempted on a medium holding the only copy of the data until that data has been secured elsewhere.
Frequently asked questions
Does the 95% figure mean that 95% of these drives will fail?
No. It is the proportion of mechanically undamaged drives among the cases analysed by one laboratory. Laboratories receive drives that already have a problem, so this type of sample cannot be used to calculate the failure rate of all units sold.
Is the WD40EZAZ at risk even though it was missing from WD’s old SMR drive list?
The October 2025 warning explicitly names it. The 2020 list covered models being shipped at the time, not a complete historical catalogue. With a used drive, always check the full model number, not merely the product line and capacity.
Can CrystalDiskInfo detect a translator failure in advance?
It displays the model, firmware version, and SMART values, but it may not detect this failure in advance because SMART does not monitor it directly. If the drive is already clicking or freezing the computer, do not turn it on merely to check it in CrystalDiskInfo.
Can I simply replace a WD Red EFAX in my NAS with a Red Plus?
Technically, this is often possible, but first verify the exact model’s compatibility, the required capacity, and the condition of the other drives in the array. On a degraded volume, do not start a rebuild automatically if another member is reporting errors or is the same at-risk model.
Is there a firmware update that fixes the problem?
As of 28 August 2026, we had not found a publicly available corrective update for the listed models. Firmware availability for internal drives also varies by model. Even if such an update were released, flashing it to a faulty drive would be a write operation—the data must be secured elsewhere first.
Can data be recovered from a clicking VeniceR drive?
It may be possible because clicking in this drive family does not necessarily mean that the platters or heads are damaged. Recovery cannot be guaranteed without diagnostics. What genuinely improves the chances is turning the drive off immediately and preserving it in its current condition.
Where to find us
ITHOPE is a company based in Brno, Czechia: we do data recovery from disks, RAID and NAS, PC and laptop service and IT for business, all in our own lab at Viniční 3587/216, 615 00 Brno.
Phone +420 775 556 063 (24/7 line). You can bring the device in person, or send it by courier (Zásilkovna, PPL).