Approaching 2000 random IOps with small queue depths and rotating hard disks? I don't think so.

Individually the drives are 285 IOps. With small queue depths you can't do much better than that.

I call ********