Make verify-flags-underscore.py return error on failure

We were running it in shippable and travis, but since it didn't return
an error it wasn't actually catching things.

Author: eparis

examples/mysql-wordpress-pd/README.md

@@ -330,7 +330,7 @@ Now, we can visit the running WordPress app.
 Use the external IP that you obtained above, and visit it on port 80:
 
 ```
-http://<external_ip>
+http://<external-ip>
 ```
 
 You should see the familiar WordPress init page.

hack/verify-flags-underscore.py

@@ -236,6 +236,7 @@ def main():
         bad_lines.sort()
         for (relname, line) in bad_lines:
             print("%s:%s" % (relname, line))
+        return 1
 
 if __name__ == "__main__":
   sys.exit(main())

hack/verify-flags/exceptions.txt

@@ -1,4 +1,6 @@
 cluster/addons/cluster-monitoring/README.md:Heapster enables monitoring of Kubernetes Clusters using [cAdvisor](https://github.com/google/cadvisor). The kubelet will communicate with an instance of cAdvisor running on localhost and proxy container stats to Heapster. Kubelet will attempt to connect to cAdvisor on port 4194 by default but this port can be configured with kubelet's `-cadvisor_port` run flag. Detailed information about heapster can be found [here](https://github.com/GoogleCloudPlatform/heapster).
+cluster/addons/registry/images/Dockerfile:ADD run_proxy.sh /usr/bin/run_proxy
+cluster/addons/registry/images/Dockerfile:CMD ["/usr/bin/run_proxy"]
 cluster/aws/templates/salt-minion.sh:# We set the hostname_override to the full EC2 private dns name
 cluster/gce/coreos/helper.sh:# cloud_config yaml file should be passed
 cluster/juju/charms/trusty/kubernetes-master/files/controller-manager.upstart.tmpl:     --address=%(bind_address)s \
@@ -32,8 +34,7 @@ cluster/juju/charms/trusty/kubernetes/hooks/network-relation-changed:        api
 cluster/juju/charms/trusty/kubernetes/hooks/network-relation-changed:    api_servers = get_rel_hosts('api', rels, ('hostname', 'port'))
 cluster/juju/charms/trusty/kubernetes/hooks/network-relation-changed:    etcd_servers = get_rel_hosts('etcd', rels, ('hostname', 'port'))
 cluster/juju/charms/trusty/kubernetes/hooks/network-relation-changed:    for k in ('etcd_servers', 'kubeapi_server'):
-cluster/libvirt-coreos/util.sh:        public_ip=$MASTER_IP
-cluster/libvirt-coreos/util.sh:      public_ip=${MINION_IPS[$i]}
+cluster/mesos/docker/common/bin/util-ssl.sh:  local cluster_domain="cluster.local"
 cluster/saltbase/salt/kube-addons/kube-addons.sh:# Create admission_control objects if defined before any other addon services. If the limits
 cluster/saltbase/salt/kube-admission-controls/init.sls:{% if 'LimitRanger' in pillar.get('admission_control', '') %}
 cluster/saltbase/salt/kube-apiserver/kube-apiserver.manifest:{% set params = address + " " + etcd_servers + " " + cloud_provider + " " + cloud_config + " " + runtime_config + " " + admission_control + " " + service_cluster_ip_range + " " + client_ca_file + " " + basic_auth_file + " " + min_request_timeout -%}
@@ -80,3 +81,7 @@ pkg/util/oom/oom_linux.go:// Writes 'value' to /proc/<pid>/oom_score_adj. PID =
 test/e2e/es_cluster_logging.go:		Failf("No cluster_name field in Elasticsearch response: %v", esResponse)
 test/e2e/es_cluster_logging.go:	// Check to see if have a cluster_name field.
 test/e2e/es_cluster_logging.go:	clusterName, ok := esResponse["cluster_name"]
+test/e2e/host_path.go:			fmt.Sprintf("--file_content_in_loop=%v", filePath),
+test/e2e/host_path.go:			fmt.Sprintf("--retry_time=%d", retryDuration),
+test/images/mount-tester/mt.go:	flag.IntVar(&retryDuration, "retry_time", 180, "Retry time during the loop")
+test/images/mount-tester/mt.go:	flag.StringVar(&readFileContentInLoopPath, "file_content_in_loop", "", "Path to read the file content in loop from")