High Availability architecture is one main feature of netezza

there will be two hosts ( active and standby) which runs High availability Linux and uses the Distributed Replicated Block Device (DRDB) which mirrors the contents between the hosts. any problem detected, a failover will be initiated to the standby host

Coming to the main data, each disk consists of primary data slice and a mirror partition which is an identical copy of another disk's primary data slice. if a disk fails, regeneration task will be triggered automatically where the spare disks will regenerate the failed disk by using its mirror copy. during this failover and regeneration queries fired on the failed disk will not be effected as the disk storing its mirror copy will respond to the query.

you can find more on these at the webcourses at https://www.syberworks.com/netezza/Register.htm