Merge pull request #4342 from kubernetes/revert-4341-helm/tls_hostname

Revert "fix(helm): add `--tls-hostname` flag to tls flags."

cmd/helm/helm.go

@@ -40,7 +40,6 @@ import (
 )
 
 var (
-	tlsServerName string // overrides the server name used to verify the hostname on the returned certificates from the server.
 	tlsCaCertFile string // path to TLS CA certificate file
 	tlsCertFile   string // path to TLS certificate file
 	tlsKeyFile    string // path to TLS key file
@@ -281,13 +280,8 @@ func newClient() helm.Interface {
 		if tlsKeyFile == "" {
 			tlsKeyFile = settings.Home.TLSKey()
 		}
-		debug("Host=%q, Key=%q, Cert=%q, CA=%q\n", tlsKeyFile, tlsCertFile, tlsCaCertFile)
-		tlsopts := tlsutil.Options{
-			ServerName:         tlsServerName,
-			KeyFile:            tlsKeyFile,
-			CertFile:           tlsCertFile,
-			InsecureSkipVerify: true,
-		}
+		debug("Key=%q, Cert=%q, CA=%q\n", tlsKeyFile, tlsCertFile, tlsCaCertFile)
+		tlsopts := tlsutil.Options{KeyFile: tlsKeyFile, CertFile: tlsCertFile, InsecureSkipVerify: true}
 		if tlsVerify {
 			tlsopts.CaCertFile = tlsCaCertFile
 			tlsopts.InsecureSkipVerify = false
@@ -307,7 +301,6 @@ func newClient() helm.Interface {
 func addFlagsTLS(cmd *cobra.Command) *cobra.Command {
 
 	// add flags
-	cmd.Flags().StringVar(&tlsServerName, "tls-hostname", settings.TillerHost, "the server name used to verify the hostname on the returned certificates from the server")
 	cmd.Flags().StringVar(&tlsCaCertFile, "tls-ca-cert", tlsCaCertDefault, "path to TLS CA certificate file")
 	cmd.Flags().StringVar(&tlsCertFile, "tls-cert", tlsCertDefault, "path to TLS certificate file")
 	cmd.Flags().StringVar(&tlsKeyFile, "tls-key", tlsKeyDefault, "path to TLS key file")

docs/tiller_ssl.md

@@ -284,21 +284,6 @@ the host name that Helm connects to matches the host name on the certificate. In
 some cases this is awkward, since Helm will connect over localhost, or the FQDN is
 not available for public resolution.
 
-*If I use `--tls-verify` on the client, I get `Error: x509: cannot validate certificate for 127.0.0.1 because it doesn't contain any IP SANs`*
-
-By default, the Helm client connects to Tiller via tunnel (i.e. kube proxy) at 127.0.0.1. During the TLS handshake,
-a target, usually provided as a hostname (e.g. example.com), is checked against the subject and subject alternative
-names of the certificate (i.e. hostname verficiation). However, because of the tunnel, the target is an IP address.
-Therefore, to validate the certificate, the IP address 127.0.0.1 must be listed as an IP subject alternative name
-(IP SAN) in the Tiller certificate.
-
-For example, to list 127.0.0.1 as an IP SAN when generating the Tiller certificate:
-
-```console
-$ echo subjectAltName=IP:127.0.0.1 > extfile.cnf
-$ openssl x509 -req -CA ca.cert.pem -CAkey ca.key.pem -CAcreateserial -in tiller.csr.pem -out tiller.cert.pem -days 365 -extfile extfile.cnf
-```
-
 *If I use `--tls-verify` on the client, I get `Error: x509: certificate has expired or is not yet valid`*
 
 Your helm certificate has expired, you need to sign a new certificate using your private key and the CA (and consider increasing the number of days)

pkg/tlsutil/cfg.go

@@ -33,9 +33,6 @@ type Options struct {
 	CertFile string
 	// Client-only options
 	InsecureSkipVerify bool
-	// Overrides the server name used to verify the hostname on the returned
-	// certificates from the server.
-	ServerName string
 	// Server-only options
 	ClientAuth tls.ClientAuthType
 }
@@ -58,12 +55,8 @@ func ClientConfig(opts Options) (cfg *tls.Config, err error) {
 			return nil, err
 		}
 	}
-	cfg = &tls.Config{
-		InsecureSkipVerify: opts.InsecureSkipVerify,
-		Certificates:       []tls.Certificate{*cert},
-		ServerName:         opts.ServerName,
-		RootCAs:            pool,
-	}
+
+	cfg = &tls.Config{InsecureSkipVerify: opts.InsecureSkipVerify, Certificates: []tls.Certificate{*cert}, RootCAs: pool}
 	return cfg, nil
 }