swap rebalance 2 nodes 1.8.1 in 1.8.0 cluster with defaut bucket 2 replica hang in windows

Install couchbase server 1.8.0 on 4 node cluster (21, 22, 23 and 24)
Create a default bucket with 2 replica.
Load 400 K items to this default bucket
Install couchbase server 1.8.1 on 2 nodes (25 and 73)
Add node 25 and 73 to 1.8.0 cluster but not rebalance yet.
Apply the hotfix to node 25 and 73 to enable swap rebalance.
Remove node 21 and 22
Start rebalance.
Rebalance hang. Tried to stop and do rebalance again but it hang and failed

Rebalance exited with reason {exited,
{'EXIT',<0.13654.2>,
{replicator_died,
{'EXIT',<15204.32734.0>,timeout}}}}