HDD failure testing, theory questions?

Question

Software testing

Good afternoon!
Such a non-trivial question. Usually in connection with hard drives, people have a question about finding a means to test the surface condition, SMART and so on. In this matter, such programs as Victoria, MHDD and others like them reign. Less often there is a question about performance testing, IOPS and such. IOMeter helps here.
I was "lucky" to get the task to highlight the issue of assessing fault tolerance. That is, a disk is taken, stress tested for failure (for example, multiple writes to one sector), on the basis of which a conclusion is made about the quality of the model / batch of disks. (We look at SMART, what has changed).
Could someone give me a tip, can I get foreign sources on how to do this? That is, I must somehow understand what is the standard load on the HDD during office work, in stressful conditions. How to imitate? Let's say I wrote down 256k 300 gigabytes of data in 1 LBA sector. Got relocate. How many months of regular work does this correspond to? And further in this spirit. For otherwise you will have to broadcast to the authorities that the mission is impossible (

Answer 1

Testing disks for "fault tolerance" is not an easy thing and requires a lot of hardware and, as a result, money. You can't just take one or two disks and, based on the results of their work, draw conclusions about the fault tolerance of this model. To do this, you need to have at least a few (tens/hundreds) of disks from different batches + develop a methodology that will take into account statistical errors, etc.
I would advise looking for existing data and testing methods, but this is extremely rare in the public domain. The only thing that comes to mind is Backblaze with their quarterly and annual reports.
Try starting from there.
https://www.backblaze.com/b2/hard-drive-test-data.html
Perhaps someone like Google (something seems to come to mind) or Amazon shares similar statistics, but I did not look for it purposefully.
Well, a bit of bare theory .

Answer 2

In terms of overwriting sectors, HDDs are quite stable, unlike SSDs. They fail due to hardware malfunctions - dust got in, knocked, high humidity, the read / write head failed, etc. You can build a vibrating stand, stick it in the stove, spill water ... I don’t know what other stress tests are. Writing several terabytes to one sector will most likely survive.

Answer 3

dubious tests and conclusions
, taking into account the fact that the disks are guaranteed for two years, you can not bathe, everything that is strong with marriage will die before this period.