My Seagate HDD failed last week. It has 3 partitions: 1 Linux, 1 windows, and 1 for data.

I was working in Linux and suddenly when I was saving a text file the system complained saying the filesystem was read only. Then the whole Mate desktop disappeared and I had to restart.

When I restarted the HDD was not listed in the POST screen. Neither in the BIOS. I changed the connector to another SATA, changed the cable for a known working one, but it was dead.

After some time I mounted the HDD in a USB adapter box to make sure it was not the PSU. Amazingly it worked. I only had backup of the most important data, because I have too many files in the dying HDD and as of today I dont have enough backup storage. So I tried to backup the bulk of the remaining files in the data partition to a laptop, but after 4 minutes of copying the disk was gone again.

I turned the adapter on ond off several times to no avail.

After an hour or so I came back for my lost data again, and once more it turned on at the first try, but after 15 minutes it died in the same manner. And again once it dies, you turn it on and off several times but does not work.

Now it occurred to me that the HDD degrades with time, so I though it might be the temperature. I put it in the fridge for 20 minutes, then connected it again with an external fan, and I was able to recover 32 GB of my precious data. Just after that it died again.

On my 4th try, I repeated the fridge and fan technique but this time I placed a frozen bag on top of the metallic case of the HDD. Again I extracted several GBs of data.

I ran a S.M.A.R.T app and it says the HDD is at risk (too many moved sectors or something like that). This could indicate the disks are bad, but on the other hand the HDD seems to work better at 20ºC than at 35ºC, which could indicate the electronics in the controller are damaged.

Could this be possible? How should I proceed now: should I try a full clone while covered in ice? Should I try to repair it using software? Should I send it to a data recovery service (it will cost several hundreds, up to a thousand bucks).

  • 2
    without being a definitive answer, metal expands when heated; if the controller/head mechanism is out of kilter, it may prevent it following that expansion correctly. Recovery might entail such extremes as keeping the drive in the fridge whilst attempting it, or giving up & handing it to a recovery specialist. [Edit: or, of course, just recovering from your backup...
    – Tetsujin
    Oct 26, 2014 at 18:21
  • The backup is incomplete. There are so many hours of config spent in both windows and linux partitions I'd be willing to pay if there were a way of recovering it all. Oct 26, 2014 at 18:30
  • I've never actually needed a recovery specialist, but they charge by the MB & are to all accounts expensive. More a corporate solution than home user, I would think... unless your pockets are very deep... Maybe try the fridge first, but that would come with no guarantees & might prove more damaging, vs 'professional recovery'. Choice is ultimately yours :-(
    – Tetsujin
    Oct 26, 2014 at 18:33
  • Yes, that ice-based trick is a very good idea, +1. It indicates a mechanical problem, and do this until your hd won't be fully recovered.
    – peterh
    Oct 26, 2014 at 18:58
  • if the drive is dying, you pays your money, you takes your choice. Other than expensive pro recovery, everything else is fraught with danger.... but free til it breaks. A definite hint to invest in a backup strategy in future.
    – Tetsujin
    Oct 26, 2014 at 21:12

1 Answer 1


I finally got the damn thing cloned. I tried putting the HDD in the fridge for some hours, and also putting a block of ice on top of it while reading, but the temperature didn't seem to make a difference. But there is for sure something that starts malfunctioning at 10 to 15 minutes after the disk is turned on.

It was an odd failure: the platters were spinning, there wasn't any periodic clicking, nor any scratching noise. It just dies after some time, and then it becames undetectable by the OS and even the POST screen. Then you let it rest for an hour or so and ready to go again. At first I thought that some bad sectors were causing the disk to hang, but on the other hand I was able to extract 100% of the files and they were ok. I refuse to believe it is programmed obsolescence, so I guess it is just crappy electronics.

As I couldn't extract more than 10GB each time, it took me several attempts to recover all the files. rsync was of great help here.

As I said, no reason for it to break this soon. I don't make a heavy use of it, there's no vibration in the room, no humidity, and the PC has a UPS since always, so no voltage spikes (the PSU is two years old, good brand)- I've bought a replacement disk guaranteed for 5 years.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .