On your XtremIO box you will just need the journal volumes if you want them to come from that array. The journal volumes can come from either of the arrays at this site since you will have multiple arrays. In terms of size for failover, that technote doc will also talk about the changes in replication mode upon failback when you use snap based replication at the source and vnx splitter at the target so take a look at that so you know what mode to flip to in a failover. When it is snap based replication the XtremIO as a target still needs a journal, but it can be minimal for the config and be 10GB.
If this is an existing RP cluster there isn't a requirement for a repository volume because one already exists for the cluster, likely on your VNX. If you plan to remove the VNX from the environment then the XtremIO will need to hold the repo volume and you will have to migrate it there.
Let me know if this helps.