Run the HDD manufacturers diagnostics, including the full
read/write surface scan. Same for scandisk, surface
Fill it with data, then copy it off again. Personally I
never trust a new drive for at least a few weeks, only
mirrored data goes onto it.
As for one-system-one-drive type of burn in, not really, at
most you can again run the manufacturers diagnostics if
they'll run, and try installing windows/other-OS.
If the drive has a jumper labelled SS (Self Seek), this
will give it a jolly good work through with just the power
connected. Unfortunately, this has become quite rare on
disks nowadays, whereas it was once quite standard.
The theory being that burning it in will reveal faults (that won't show up
until used a bit) in areas that can then be locked away from use. Or, if a
significant number, trigger the return of the drive to the vendor.
Not really. Infant mortality for HDDs is pretty low these days, so
burn-in does not help much. Same is generally true for semiconductors.
It used to be different.
Hiwever if you really want to burn in, then just put the drive
under higher load for some time. I used to do this by compiling
Linux kernels in a loop.
Just crank up some memory hog application that handles files bigger
than your ram and you will thrash the hell out of the drive as it
pages the data in and out.
Sound Forge with a big audio file or some video editor springs to
I extend the mistrust to the disks entire llifespan...
It's volatile memory...
backup procedures, redundancy etc..
Are there any drives that one could submit to diagnostics ALL the
time, that transmit "condition data" to the mobo/OS constantly ?
With devices such as disk drives, if they don't fail within an hour or
two, they'll probably run for years. Vendors exercise drives to
reduce the incidence of the former. As a result, drives that survive
a very brief infancy will likely remain reliable for a very long time.