I tend to believe that a backup (pg_dump) in custom format (-F c) using multiple jobs (parallel) -> restore (pg_restore) also with multiple concurrent jobs would be better.
I have a relatively large table (100m rows) that I want to move to a new box with more resources. The table isn't doing anything...i.e its not being updated or read from. Which approach would be faster to move the data over:
a). Use pg_fdw and do "create local_table as select * from foreign_table". b). setup logical replication between the two servers.