Ensure copy-only and copy-and-stream are resumable just like stream mode

[shayonj]

If a the bulk copy/load experiences a network disruption today or a deployment comes in, the entire process would need to be started over again. Lets make it resumable like streaming replicator.

Further clarification

• Streaming is fully resumable, say if you have to shutdown the container, or a SIGTERM comes, it will flush, save state, exit.
• The copy modes are not resumable today, so if you need to interrupt the copy, for now its best to dump/delete the table in the target DB and start the process again. It parallelizes the data syncs, so should be fast, depending on table size.
• Once the operation moves from copy to stream (which happens automatically when you use copy-and-stream) the operation is resumable-able again, since the initial sync is complete and the new data is flowing from the replication slot.