8-10 filers behind a rainstoreage cluster
10M files max to move on a sync between mirrors

must set level for syncs: max, nonintrusive, variable in between
1 sync per filer, rest in queue (up to 300 in queue?)

rainstorage is a router for file requests

NFS is sync then broken when 100% and switched to new plex (use automounter)
CIFS is sync then stays that way until you move over users to new box,
	then you break the mirror and destroy the old primary plex (use DFS or batch)

keeps track of addresses that use a particular qtree

layer 2 switch, works like a bridge
need a management IP
have 8 GB ethernet
needs a new VLAN
use GB fiber (can do copper)
use SC
we have 3 subnets

DL380 HP server w/ Redhat , 4GB, 2-72 mirrored, redundant PS
	either 2u or 5u: only diff is 5u has more PCI for fibre boards 
		(2 conn percard, copper has 4 conn per card)
	cpu perform limiting factor

licenses
	ulimited cap, unlimited move, unlimited sites
	done by unit and protocols on unit
		rainstore app
		NFS
		CIFS

can also work with other NFS and CIFS other than Netapp