The FastT has a global pool of hot spare disks, hot spares do not need to be in the same drawer as the failing disk. It will first look for any hot spares defined in the same drawer as the failed drive, then it looks at the last hot spares defined.
I always define at least one hot spare in each drawer. If one should fail, the system will use the drive in that drawer. Also, make sure that the drive assigned as a hot spare is one of the largest drives in that drawer. If the hot spare is too small, the system will start looking at hot spares in other drawers as candidates.
Where all of this could potentially bite you, though I have never seen it, is if a drive failed in a drawer with either no hot spare defined or a hot spare that is too small. If the last drawers added were slower disks, like SATA, the system could potentially use one of those disks as a hot spare for an array of fibre channel disks. I don’t know if the FastT is smart enough not to do mix fibre channel and SATA disks, but I’m not going to test it. 🙂 Just bite the bullet and define one hot spare per drawer, possibly two for large SATA disks.