netmaster/netplugin restart

[rchirakk]

install/install/k8s/install.sh

Line 202 in 6a0bea7

cat $contiv_yaml_template >>$contiv_yaml

start etcd before netplugin/netmaster

cat $contiv_yaml_template >>$contiv_yaml

if [ "$cluster_store" = "" ]; then
	cat $contiv_etcd_template >>$contiv_yaml
fi

[rchirakk]

Another change to make:
start netmaster before netplugin

[dseevr]

Per the private issue we had to track this, starting etcd before the other services actually made the problem worse. Both nodes in the demo environment come up in the wrong mode consistently.

We should probably have an issue to track making netplugin resilient against an unready data store in general. EDIT: added an issue: contiv/netplugin#915