Each recovery point consists of one data file and one index file. Both files start with the volume GUID, followed by an underscore, and the epoch number. The index file then has an "i". And both end with a ".ReplayEPS" extension.
During replication there are also two checksum files--one for the data and one for the index--that have the same names as the recovery point files, but end in an ".A32" extension. These are used to verify the recovery point files were replicated intact. These four files are transferred to the target core's "repl" folder. A comparison is run between the checksum and recovery point files, and upon verifying the data is valid, the A32 files are deleted and the recovery point files are moved into the main repository folder.
If replication gets interrupted before all four files are replicated, it can sometimes cause replication to hang at that point.