Couchbase Server

inability to reach some node should not cause entire per-bucket supervisor to fail [was: Rebalance 5->4 nodes is failed with reason bulk_set_vbucket_state_failed]