riak-repl fullsync might stop on network errors

[cipy]

If the tcp or ssl socket has an error or closes during a fullsync, the coordinator exits abnormally. This means a full sync would need to be completely restarted if there's just a network blip. Given that a full sync can take a very long time, it would be favorable to attempt to re-establish the connection instead while keeping as much state alive as possible.

tracked also as riak-repl #625

[bsparrow435]

Doesnt really matter but this should be filed against https://github.com/basho/riak_ee-issues