I did something similar in a migration of a one node cluster from an EMC array to XP array. Mirroring is the way to go. However I'd make the suggestion that after you vgextend your new disks (command should include all the new disks you are going to extend) the next step would be another vgextend command with the alternate paths. For example:
vgcreate -l 20 -s 16 -e 26222 /dev/vg01 /dev/dsk/c4t10d4 /dev/dsk/c4t10d5
vgextend /dev/vg01 /dev/dsk/c6t4d1 /dev/dsk/c6t5d6
vgdisplay will then show the alternate paths.
Also do all the volume group work with the cluster running, that is do your vgextends, lvextend with the mirror, lvreduce and vgreduce the old disks. Then stop your cluster do to the lock disk edits. If you stop your cluster before the LVM work you will have to activate the volume groups and then deactivate the volume groups before in order to complete the LVM work before stopping the cluster. Here's a link for reference:
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1342480Lastly don't forget to do a rmsf -H to remove the hardware paths from the old array.