OFAgent container crash when some etcd nodes go down.

I set up the bbsim environment with minimal configuration and three etcd pods by running following command
EXTRA_HELM_FLAGS="--set defaults.image_tag=master --set voltha-etcd-cluster.clusterSize=3 --set images.bbsim.tag=0.0.4" TYPE=minimal WITH_RADIUS=y WITH_BBSIM=y INSTALL_ONOS_APPS=y CONFIG_SADIS=y ./voltha up.

Following is my testing steps.
1. setup voltha
2. run make sanity-kind (pass)
3. scale down the etcd from 3 pods to 1 pods by the following command. (simulate some etcd nodes go down)

kubectl -n voltha patch etcdclusters.etcd.database.coreos.com voltha-etcd-cluster --type='merge' -p '{"spec":{"size":1}}'

4. run make sanity-kind(fail).

I found the voltha-ofagent container crash after I scale down the etcd cluster.
voltha-ofagent-65f9b885d5-2bkkt                             0/1     Running   0          11m
 

I removed the teardown step in the second test to keep the testing data and use the voltha dump to dump the debug info. Please see the attachment.