We currently run nodetool repair with the -pr option on every single node every weekend , and as the literature suggests run the repairs with the full option once every month on all nodes .
Is it enough to run the repair with the full option for one node per datacenter ?
Currently running on AWS Ec2 instances in a single region with 2 datacenters.
it is always best to run
nodetool repair --partitioner-range(or-prin shorthand) one node at a time across the entire cluster.here is another article that explains about repairs.