I have Isilon A2000 28-node cluster with protection (+2d:1n). One node is failed and we want to evict it. At the same time, one disk on another node is  failed and is in REPLACE state.  So my question is can I replace the disk first and then smart fail the node? Just curious as I read pulling out SLED is equivalent to one drive failure. Or do I need to do smart fail and then go for disk replacement?