Odd SMART Behavior

[TraTech]

A 320GB WD drive I had been using as scratch space for a while now (2008) started to act up. Transfers are capping at 700 KB/s.

Speedfan S.M.A.R.T. data:


The reallocated sector count should be something other than zero if there are so many bad sectors, right? I tried HDDTune's utility for drive zeroing, and, moving at 700 KB/s, it would have taken 5 days. Same thing with DBAN, except it reported 25 days to completion. Spinrite hung upon trying to detect it. Any benchmarking on the drive failed with "read error."

The odd thing is that I placed a 720p movie (1.4GB took 30 minutes to transfer) on the drive to see if it could play back. Lo and behold, it played back just fine throughout the entire length of it.

Seems odd to me, and no big loss if the consensus is to toss it. I was planning on partitioning a part of my new-er 1TB WD Black for scratch space anyway. Just wondering why is acts as it does and if there is any saving it. Thanks for your input.

If it helps, this is the rig it is stuffed in:
i5-2500k | 8GB DDR3 1600 | ASUS HD6950 2GB | ASUS P8Z68-V Pro | 128GB Crucial M4 | 1TB WD FAEX

[Vulcan]

Here are my comments, in case they are useful. If you don't like them, then please ignore

The reallocated sector count should be something other than zero if there are so many bad sectors, right?

Not necessarily.

Within the limits of the SMART attributes and what they can show, I don't see anything really "odd" about this info. It's either a sick drive, or has poor quality power and/or vibration, causing it to have problems.

The odd thing is that I placed a 720p movie (1.4GB took 30 minutes to transfer) on the drive to see if it could play back. Lo and behold, it played back just fine throughout the entire length of it.

Depending on whether that SMART data was collected before, or after, copying that movie to the drive, the movie played OK either because it forced some reallocations when it was written to the drive, or because that part of the drive was OK anyway.

If you want to try continuing to use that drive (no guarantee it won't fail at a critical time!), and assuming you don't need the data on the drive now, I would zero-fill it, collect the SMART data, then read scan it (if you get any fatal read errors here, then scrap the drive), collect the SMART data again (see if any new pending sectors are recorded at this point), and then run a benchmark to see if the speed is acceptable. However this is probably in vain, and in any case, I wouldn't use the drive for anything other than scratch space which I could afford to lose.

Have fun

[TraTech]

Thanks! F.Y.I the SMART data was taken after the movie, and after about a day (20% in HDTune) of zero-filling, oddly enough. Is it common for the zero-fill to move so slowly? What could be causing this? My previous experience with drives is black or white. i.e either the drive is running normally, or the drive is dead (not recognized/unreadable/wont spin up etc.). Never have I had a drive simply under perform and spit random errors. It has really got me curious as to what specifically can cause this. Is this type of "failure" rare?

Thanks again!

(Oh, and I'll put it in my mates Windows box NAS unit for a week and zero it out properly; worth a try, I guess; it is his if it works )

[Vulcan]

Thanks! F.Y.I the SMART data was taken after the movie, and after about a day (20% in HDTune) of zero-filling, oddly enough.

It would be interesting to know whether, on that zero-filled approx 20% of the disk, the read speed had improved compared to before that zero-fill.

Is it common for the zero-fill to move so slowly?

Only when there is a problem

What could be causing this?

IMHO accurate diagnosis would require the drive to be in front of someone with the necessary equipment. Remote diagnosis would be a guess. As I said before, power & other factors like vibration can also affect drive behaviour, and should be eliminated (by measurement or experimentation) to avoid unnecessarily blaming the drive.

My previous experience with drives is black or white. i.e either the drive is running normally, or the drive is dead (not recognized/unreadable/wont spin up etc.). Never have I had a drive simply under perform and spit random errors.

There are many "shades of grey" with drive problems.