Ric Wheeler wrote:

The hard thing is to figure out how to test this kind of scenario 
without dropping power.

Apart from using virtual machines, you mean?

To expose the failure mode, it might be 
sufficient to drop power to a drive with smartctl (or, if you have hot 
swap bays, just pull them).

I wonder, does sending any kind of reset commands to drives cause them
to discard their write cache?

If the hot swap power switch is accessible from software, that would
be a nice trick :-)

-- Jamie