When Refurbished Drives Fail: Inside an I.NORYS NAS HDD

We recently received a hard drive from a completely unfamiliar brand, I.NORYS, with unusual model markings: INO-IHDD4000S3-D1-5964, which had failed while operating in a NAS (Network Attached Storage). After a brief Internet search, we found a legend explaining these part codes: IHDD – Internal HDD, 4000 – a capacity of 4TB, S3 – SATA3, D1 – 3.5″ format, 59 – 5900 RPM, and 64 – 64MB cache. Comparing it with familiar brands and models, we concluded that its electronics and build were the same as those of a Seagate drive, though there was
no official branding. Even before opening the unit, it stood out for one unusual reason: the sound.
Upon powering it up, we heard a noise unlike anything we had encountered before. It wasn’t the typical head clicking or beeping. Instead, it was a sudden, sharp scraping sound that occurred intermittently, leaving us puzzled. We had no idea where it was coming from, only that it was not good. After opening the drive, things became clearer and more concerning. Inside, we discovered that one of the six screws securing the platters was not fastened. In fact, it wasn’t screwed in at all. It was just sitting in place, slightly protruding and dangerously close to falling out. Upon further inspection, we concluded that the thread where the screw should go had never been fully cut at the factory, so the screw had only been machine-tightened as far as the existing thread allowed. This was not a case of a screw coming loose during use. It had never been installed correctly. As a result, the screw had scratched the inner cover of the drive, producing aluminum shavings that posed a serious threat to the heads and platters (picture 1).

(picture 1: protruding screw and scratched inner lid)

We removed the screw entirely, replaced the cover, and began with standard head swap and cloning procedures. The original heads were stuck under the ramp, and one platter surface showed signs of minor surface damage and dust contamination. After the first cloning attempt, the new heads also became contaminated, most likely by residual particles left from the mechanical damage, and failed mid-process. A second head replacement was required. This time, we were able to recover around two-thirds of the user data. The HddSurgery FileLister showed
fewer than 30% of files were affected by bad sectors or header damage.
(picture 2)

(picture 2: file list statistics generated by FileLister)

Let’s talk about how this can happen.

Companies that refurbish drives usually acquire failed units in bulk, often thousands at a time, and try to bring some of them back to life using internal rework methods. If a drive fails Seagate’s or WD’s internal QA testing, for example vibration, read stability, or power management, it never reaches the market. Refurbishers, however, do not always follow the same standards. A drive that failed in vibration testing, like this one may have, could slip through if the refurbishment process skips that check. And so this drive, which clearly once worked, was returned to the market under a new name and failed in the field when it was likely bumped or relocated during operation, which probably caused the protruding screw to make contact. We are not here to condemn the entire concept of refurbished drives. Refurbishment, when done properly, can serve a valid purpose.

But we are here to say one thing clearly:
If the data is important, always keep a backup.

It does not matter if your drive is new, used, or refurbished; mechanical failure does not discriminate. And when it comes to your business or personal data, the risk just is not worth it.

HelpDisc, Belgrade