GoXDCR crashes on master after a node is rebalanced out of cluster

Build
------
4.0.0-1767

Testcase
--------
./testrunner -i 3clusters.ini -p get-cbcollect-info=True,get-logs=False,stop-on-failure=False,get-coredumps=True,replication_type=xmem -t xdcr.xdcrTopologyTests.XDCRTopologyTest.load_with_rebalance_out,items=10000,chain_length=3,rdirection=unidirection,ctopology=chain,standard_buckets=1,default_bucket=False,rebalance=C2,num_rebalance=1,replication_type=xmem

1. C1 --> C2[.188,.189] --> C3
2. Load 10K keys on to C1 and C2.
3. Rebalance out .189 on C2. goxdcr crashes on .188 soon after .189 goes down.

Attaching logs from .188 and .189.

Port server goxdcr on node 'babysitter_of_ns_1@127.0.0.1' exited with status 1. Restarting. Messages: github.com/couchbase/gomemcached/client.(*UprFeed).runFeed(0xc20878c2d0, 0xc208c8e210)
/home/couchbase/jenkins/workspace/sherlock-unix/godeps/src/github.com/couchbase/gomemcached/client/upr_feed.go:398 +0x162
created by github.com/couchbase/gomemcached/client.(*UprFeed).StartFeed
/home/couchbase/jenkins/workspace/sherlock-unix/godeps/src/github.com/couchbase/gomemcached/client/upr_feed.go:328 +0x65
[goport] 2015/04/08 07:52:13 /opt/couchbase/bin/goxdcr terminated: exit status 2 ns_log000 ns_1@10.3.4.188 07:52:13 - Wed Apr 8, 2015