This is the procedure:
1) Upgrade your existing cluster to the latest version of CanIt. This is required so the version matches that of our ISO image.
2) Build a new machine with the ISO image.
3) Transfer the database over from the old to the new machine. This will involve downtime; it involved doing a pg_dump on the old machine and restoring on the new machine, as specified in the Administration Guide.
4) Point all your existing cluster members to the new DB.
5) Make a second machine with the ISO and join it to the new DB.
6) Make the second machine a read/write Storage Manager node and make canIt storage a read-only node. Wait 30-60 days until there's no more relevant data left on CanIt storage; at that point, you can decommission it by first re-running the Storage Manager wizard and removing it, then removing it from the Cluster Members table.
7) At your leisure,remove the old machines from the cluster.