Couchdb replication stalled with one record left

I have two couchdb instances with continuous replication setup from A to B. The replication proceeds fine when initiated, but then stalls at 93%. When I dig deeper I see that the update_seq is 15, but I'm stalled at 14 (hence the 93%).

If I look at the database on instance A, it shows:

{
  "db_name":"dummy",
  "doc_count":14,
  "doc_del_count":0,
  "update_seq":15,
  "purge_seq":0,
  "compact_running":false,
  "disk_size":262254,
  "data_size":218586,
  "instance_start_time":"1453487824616597",
  "disk_format_version":6,
  "committed_update_seq":15
}

and if I look at the changes stream, I show:

"last_seq":14

If I look at _active_tasks on node B I see:

{
  "pid":"<0.2413.0>",
  "checkpoint_interval":5000,
  "checkpointed_source_seq":14,
  "continuous":true,"doc_id":null,
  "doc_write_failures":0,
  "docs_read":14,
  "docs_written":14,
  "missing_revisions_found":14,
  "progress":93,
  "replication_id":"69db4b589016b28b20b1fdedb695c286+continuous+create_target",
  "revisions_checked":14,
  "source":"https://noneofyourbusiness/dummy/",
  "source_seq":15,
  "started_on":1459378067,
  "target":"dummy",
  "type":"replication",
  "updated_on":1459379195
}

I am at a loss to explain why I'm not able to get a progress of 100. The only think I can think is that node A was previously on the receiving end of replication and perhaps had that replication process terminated in an incomplete state.

I really appreciate any thoughts on what might cause this situation and how I can recover from it.

Thanks