GoXDCR: Adding back a hard-failed-over node results in "Unexpected server error"

Build
------
4.0.0-1650

May be related to xdcr checkpointing and could be a duplicate of MB-13968. Wanted to confirm the same with ns_server.

Testcase
--------
./testrunner -i -p demand_encryption=1 -t xdcr.biXDCR.bidirectional.load_with_failover_then_add_back,items=10000,ctopology=chain,rdirection=bidirection,standard_buckets=2,expires=60,update=C1-C2,delete=C1-C2,failover=C1,timeout=150

Steps

1. C1 [.45,.46] <--> C2 [.47,.48], 1 default and 2 standard buckets, checkpoint_interval=60s
2. Load 10K keys on C1 and C2.
3. Hard failover .46 and add back soon after.

[2015-03-25 16:14:44,247] - [xdcrnewbasetests:1749] INFO - Starting failover for nodes:[ip:172.23.106.46 port:8091 ssh_username:root] at C1 cluster 172.23.106.45
[2015-03-25 16:14:46,072] - [task:2803] INFO - Failing over 172.23.106.46:8091 with graceful=False
[2015-03-25 16:14:51,342] - [rest_client:1103] INFO - fail_over node ns_1@172.23.106.46 successful
[2015-03-25 16:14:51,342] - [task:2783] INFO - 0 seconds sleep after failover, for nodes to go pending....
[2015-03-25 16:14:57,887] - [rest_client:730] ERROR - http://172.23.106.45:8091/controller/reAddNode error 500 reason: unknown ["Unexpected server error, request logged."]
[2015-03-25 16:14:57,887] - [rest_client:1138] ERROR - add_back_node ns_1@172.23.106.46 error : ["Unexpected server error, request logged."]
ERROR

Adding logs from .46 and .45