On Thu, 27 Dec 2007, Justin Piszcz wrote:

With that high of a stripe size the stripe_cache_size needs to be greater than
the default to handle it.

i'd argue that any deadlock is a bug...

regardless i'm still seeing deadlocks with the default chunk_size of 64k 
and stripe_cache_size of 256... in this case it's with a workload which is 
untarring 34 copies of the linux kernel at the same time.  it's a variant 
of doug ledford's memtest, and i've attached it.

-dean