Adds basic support for node IDs.

api/agent_test.go

@@ -657,7 +657,7 @@ func TestAgent_Monitor(t *testing.T) {
 	// Wait for the first log message and validate it
 	select {
 	case log := <-logCh:
-		if !strings.Contains(log, "[INFO] raft: Initial configuration") {
+		if !strings.Contains(log, "[INFO]") {
 			t.Fatalf("bad: %q", log)
 		}
 	case <-time.After(10 * time.Second):

command/agent/agent.go

@@ -224,7 +224,6 @@ func Create(config *Config, logOutput io.Writer, logWriter *logger.LogWriter,
 		shutdownCh:     make(chan struct{}),
 		endpoints:      make(map[string]string),
 	}
-
 	if err := agent.resolveTmplAddrs(); err != nil {
 		return nil, err
 	}
@@ -236,6 +235,12 @@ func Create(config *Config, logOutput io.Writer, logWriter *logger.LogWriter,
 	}
 	agent.acls = acls
 
+	// Retrieve or generate the node ID before setting up the rest of the
+	// agent, which depends on it.
+	if err := agent.setupNodeID(config); err != nil {
+		return nil, fmt.Errorf("Failed to setup node ID: %v", err)
+	}
+
 	// Initialize the local state.
 	agent.state.Init(config, agent.logger)
 
@@ -303,6 +308,9 @@ func (a *Agent) consulConfig() *consul.Config {
 		base = consul.DefaultConfig()
 	}
 
+	// This is set when the agent starts up
+	base.NodeID = a.config.NodeID
+
 	// Apply dev mode
 	base.DevMode = a.config.DevMode
 
@@ -600,6 +608,67 @@ func (a *Agent) setupClient() error {
 	return nil
 }
 
+// setupNodeID will pull the persisted node ID, if any, or create a random one
+// and persist it.
+func (a *Agent) setupNodeID(config *Config) error {
+	// If they've configured a node ID manually then just use that, as
+	// long as it's valid.
+	if config.NodeID != "" {
+		if _, err := uuid.ParseUUID(string(config.NodeID)); err != nil {
+			return err
+		}
+
+		return nil
+	}
+
+	// For dev mode we have no filesystem access so just make a GUID.
+	if a.config.DevMode {
+		id, err := uuid.GenerateUUID()
+		if err != nil {
+			return err
+		}
+
+		config.NodeID = types.NodeID(id)
+		a.logger.Printf("[INFO] agent: Generated unique node ID %q for this agent (will not be persisted in dev mode)", config.NodeID)
+		return nil
+	}
+
+	// Load saved state, if any. Since a user could edit this, we also
+	// validate it.
+	fileID := filepath.Join(config.DataDir, "node-id")
+	if _, err := os.Stat(fileID); err == nil {
+		rawID, err := ioutil.ReadFile(fileID)
+		if err != nil {
+			return err
+		}
+
+		nodeID := strings.TrimSpace(string(rawID))
+		if _, err := uuid.ParseUUID(nodeID); err != nil {
+			return err
+		}
+
+		config.NodeID = types.NodeID(nodeID)
+	}
+
+	// If we still don't have a valid node ID, make one.
+	if config.NodeID == "" {
+		id, err := uuid.GenerateUUID()
+		if err != nil {
+			return err
+		}
+		if err := lib.EnsurePath(fileID, false); err != nil {
+			return err
+		}
+		if err := ioutil.WriteFile(fileID, []byte(id), 0600); err != nil {
+			return err
+		}
+
+		config.NodeID = types.NodeID(id)
+		a.logger.Printf("[INFO] agent: Generated unique node ID %q for this agent (persisted)", config.NodeID)
+	}
+	return nil
+}
+
 // setupKeyrings is used to initialize and load keyrings during agent startup
 func (a *Agent) setupKeyrings(config *consul.Config) error {
 	fileLAN := filepath.Join(a.config.DataDir, serfLANKeyring)

command/agent/agent_test.go

@@ -18,6 +18,8 @@ import (
 	"github.com/hashicorp/consul/consul/structs"
 	"github.com/hashicorp/consul/logger"
 	"github.com/hashicorp/consul/testutil"
+	"github.com/hashicorp/consul/types"
+	"github.com/hashicorp/go-uuid"
 	"github.com/hashicorp/raft"
 	"strings"
 )
@@ -308,6 +310,62 @@ func TestAgent_ReconnectConfigSettings(t *testing.T) {
 	}()
 }
 
+func TestAgent_NodeID(t *testing.T) {
+	c := nextConfig()
+	dir, agent := makeAgent(t, c)
+	defer os.RemoveAll(dir)
+	defer agent.Shutdown()
+
+	// The auto-assigned ID should be valid.
+	id := agent.consulConfig().NodeID
+	if _, err := uuid.ParseUUID(string(id)); err != nil {
+		t.Fatalf("err: %v", err)
+	}
+
+	// Set an invalid ID via config.
+	c.NodeID = types.NodeID("nope")
+	err := agent.setupNodeID(c)
+	if err == nil || !strings.Contains(err.Error(), "uuid string is wrong length") {
+		t.Fatalf("err: %v", err)
+	}
+
+	// Set a valid ID via config.
+	newID, err := uuid.GenerateUUID()
+	if err != nil {
+		t.Fatalf("err: %v", err)
+	}
+	c.NodeID = types.NodeID(newID)
+	if err := agent.setupNodeID(c); err != nil {
+		t.Fatalf("err: %v", err)
+	}
+	if id := agent.consulConfig().NodeID; string(id) != newID {
+		t.Fatalf("bad: %q vs. %q", id, newID)
+	}
+
+	// Set an invalid ID via the file.
+	fileID := filepath.Join(c.DataDir, "node-id")
+	if err := ioutil.WriteFile(fileID, []byte("adf4238a!882b!9ddc!4a9d!5b6758e4159e"), 0600); err != nil {
+		t.Fatalf("err: %v", err)
+	}
+	c.NodeID = ""
+	err = agent.setupNodeID(c)
+	if err == nil || !strings.Contains(err.Error(), "uuid is improperly formatted") {
+		t.Fatalf("err: %v", err)
+	}
+
+	// Set a valid ID via the file.
+	if err := ioutil.WriteFile(fileID, []byte("adf4238a-882b-9ddc-4a9d-5b6758e4159e"), 0600); err != nil {
+		t.Fatalf("err: %v", err)
+	}
+	c.NodeID = ""
+	if err := agent.setupNodeID(c); err != nil {
+		t.Fatalf("err: %v", err)
+	}
+	if id := agent.consulConfig().NodeID; string(id) != "adf4238a-882b-9ddc-4a9d-5b6758e4159e" {
+		t.Fatalf("bad: %q vs. %q", id, newID)
+	}
+}
+
 func TestAgent_AddService(t *testing.T) {
 	dir, agent := makeAgent(t, nextConfig())
 	defer os.RemoveAll(dir)

command/agent/command.go

@@ -92,6 +92,7 @@ func (c *Command) readConfig() *Config {
 
 	cmdFlags.StringVar(&cmdConfig.LogLevel, "log-level", "", "log level")
 	cmdFlags.StringVar(&cmdConfig.NodeName, "node", "", "node name")
+	cmdFlags.StringVar((*string)(&cmdConfig.NodeID), "node-id", "", "node ID")
 	cmdFlags.StringVar(&dcDeprecated, "dc", "", "node datacenter (deprecated: use 'datacenter' instead)")
 	cmdFlags.StringVar(&cmdConfig.Datacenter, "datacenter", "", "node datacenter")
 	cmdFlags.StringVar(&cmdConfig.DataDir, "data-dir", "", "path to the data directory")
@@ -1115,6 +1116,7 @@ func (c *Command) Run(args []string) int {
 
 	c.Ui.Output("Consul agent running!")
 	c.Ui.Info(fmt.Sprintf("       Version: '%s'", c.HumanVersion))
+	c.Ui.Info(fmt.Sprintf("       Node ID: '%s'", config.NodeID))
 	c.Ui.Info(fmt.Sprintf("     Node name: '%s'", config.NodeName))
 	c.Ui.Info(fmt.Sprintf("    Datacenter: '%s'", config.Datacenter))
 	c.Ui.Info(fmt.Sprintf("        Server: %v (bootstrap: %v)", config.Server, config.Bootstrap))

command/agent/config.go

@@ -14,6 +14,7 @@ import (
 
 	"github.com/hashicorp/consul/consul"
 	"github.com/hashicorp/consul/lib"
+	"github.com/hashicorp/consul/types"
 	"github.com/hashicorp/consul/watch"
 	"github.com/mitchellh/mapstructure"
 )
@@ -312,6 +313,10 @@ type Config struct {
 	// LogLevel is the level of the logs to putout
 	LogLevel string `mapstructure:"log_level"`
 
+	// Node ID is a unique ID for this node across space and time. Defaults
+	// to a randomly-generated ID that persists in the data-dir.
+	NodeID types.NodeID `mapstructure:"node_id"`
+
 	// Node name is the name we use to advertise. Defaults to hostname.
 	NodeName string `mapstructure:"node_name"`
 
@@ -1273,6 +1278,9 @@ func MergeConfig(a, b *Config) *Config {
 	if b.Protocol > 0 {
 		result.Protocol = b.Protocol
 	}
+	if b.NodeID != "" {
+		result.NodeID = b.NodeID
+	}
 	if b.NodeName != "" {
 		result.NodeName = b.NodeName
 	}

command/agent/config_test.go

@@ -60,7 +60,7 @@ func TestDecodeConfig(t *testing.T) {
 	}
 
 	// Without a protocol
-	input = `{"node_name": "foo", "datacenter": "dc2"}`
+	input = `{"node_id": "bar", "node_name": "foo", "datacenter": "dc2"}`
 	config, err = DecodeConfig(bytes.NewReader([]byte(input)))
 	if err != nil {
 		t.Fatalf("err: %s", err)
@@ -70,6 +70,10 @@ func TestDecodeConfig(t *testing.T) {
 		t.Fatalf("bad: %#v", config)
 	}
 
+	if config.NodeID != "bar" {
+		t.Fatalf("bad: %#v", config)
+	}
+
 	if config.Datacenter != "dc2" {
 		t.Fatalf("bad: %#v", config)
 	}
@@ -1532,6 +1536,7 @@ func TestMergeConfig(t *testing.T) {
 		DataDir:                "/tmp/foo",
 		Domain:                 "basic",
 		LogLevel:               "debug",
+		NodeID:                 "bar",
 		NodeName:               "foo",
 		ClientAddr:             "127.0.0.1",
 		BindAddr:               "127.0.0.1",
@@ -1586,6 +1591,7 @@ func TestMergeConfig(t *testing.T) {
 		},
 		Domain:           "other",
 		LogLevel:         "info",
+		NodeID:           "bar",
 		NodeName:         "baz",
 		ClientAddr:       "127.0.0.2",
 		BindAddr:         "127.0.0.2",

consul/client.go

@@ -14,6 +14,7 @@ import (
 	"github.com/hashicorp/consul/consul/agent"
 	"github.com/hashicorp/consul/consul/servers"
 	"github.com/hashicorp/consul/consul/structs"
+	"github.com/hashicorp/consul/lib"
 	"github.com/hashicorp/serf/coordinate"
 	"github.com/hashicorp/serf/serf"
 )
@@ -144,6 +145,7 @@ func (c *Client) setupSerf(conf *serf.Config, ch chan serf.Event, path string) (
 	conf.NodeName = c.config.NodeName
 	conf.Tags["role"] = "node"
 	conf.Tags["dc"] = c.config.Datacenter
+	conf.Tags["id"] = string(c.config.NodeID)
 	conf.Tags["vsn"] = fmt.Sprintf("%d", c.config.ProtocolVersion)
 	conf.Tags["vsn_min"] = fmt.Sprintf("%d", ProtocolVersionMin)
 	conf.Tags["vsn_max"] = fmt.Sprintf("%d", ProtocolVersionMax)
@@ -156,7 +158,7 @@ func (c *Client) setupSerf(conf *serf.Config, ch chan serf.Event, path string) (
 	conf.RejoinAfterLeave = c.config.RejoinAfterLeave
 	conf.Merge = &lanMergeDelegate{dc: c.config.Datacenter}
 	conf.DisableCoordinates = c.config.DisableCoordinates
-	if err := ensurePath(conf.SnapshotPath, false); err != nil {
+	if err := lib.EnsurePath(conf.SnapshotPath, false); err != nil {
 		return nil, err
 	}
 	return serf.Create(conf)

consul/config.go

@@ -8,6 +8,7 @@ import (
 	"time"
 
 	"github.com/hashicorp/consul/tlsutil"
+	"github.com/hashicorp/consul/types"
 	"github.com/hashicorp/memberlist"
 	"github.com/hashicorp/raft"
 	"github.com/hashicorp/serf/serf"
@@ -66,6 +67,9 @@ type Config struct {
 	// DevMode is used to enable a development server mode.
 	DevMode bool
 
+	// NodeID is a unique identifier for this node across space and time.
+	NodeID types.NodeID
+
 	// Node name is the name we use to advertise. Defaults to hostname.
 	NodeName string
 

consul/server.go

@@ -20,6 +20,7 @@ import (
 	"github.com/hashicorp/consul/consul/agent"
 	"github.com/hashicorp/consul/consul/state"
 	"github.com/hashicorp/consul/consul/structs"
+	"github.com/hashicorp/consul/lib"
 	"github.com/hashicorp/consul/tlsutil"
 	"github.com/hashicorp/raft"
 	"github.com/hashicorp/raft-boltdb"
@@ -308,6 +309,7 @@ func (s *Server) setupSerf(conf *serf.Config, ch chan serf.Event, path string, w
 	}
 	conf.Tags["role"] = "consul"
 	conf.Tags["dc"] = s.config.Datacenter
+	conf.Tags["id"] = string(s.config.NodeID)
 	conf.Tags["vsn"] = fmt.Sprintf("%d", s.config.ProtocolVersion)
 	conf.Tags["vsn_min"] = fmt.Sprintf("%d", ProtocolVersionMin)
 	conf.Tags["vsn_max"] = fmt.Sprintf("%d", ProtocolVersionMax)
@@ -337,7 +339,7 @@ func (s *Server) setupSerf(conf *serf.Config, ch chan serf.Event, path string, w
 	// When enabled, the Serf gossip may just turn off if we are the minority
 	// node which is rather unexpected.
 	conf.EnableNameConflictResolution = false
-	if err := ensurePath(conf.SnapshotPath, false); err != nil {
+	if err := lib.EnsurePath(conf.SnapshotPath, false); err != nil {
 		return nil, err
 	}
 
@@ -390,7 +392,7 @@ func (s *Server) setupRaft() error {
 	} else {
 		// Create the base raft path.
 		path := filepath.Join(s.config.DataDir, raftState)
-		if err := ensurePath(path, true); err != nil {
+		if err := lib.EnsurePath(path, true); err != nil {
 			return err
 		}
 

consul/util.go

@@ -4,8 +4,6 @@ import (
 	"encoding/binary"
 	"fmt"
 	"net"
-	"os"
-	"path/filepath"
 	"runtime"
 	"strconv"
 
@@ -64,14 +62,6 @@ func init() {
 	privateBlocks[5] = block
 }
 
-// ensurePath is used to make sure a path exists
-func ensurePath(path string, dir bool) error {
-	if !dir {
-		path = filepath.Dir(path)
-	}
-	return os.MkdirAll(path, 0755)
-}
-
 // CanServersUnderstandProtocol checks to see if all the servers in the given
 // list understand the given protocol version. If there are no servers in the
 // list then this will return false.

lib/path.go

@@ -0,0 +1,14 @@
+package lib
+
+import (
+	"os"
+	"path/filepath"
+)
+
+// EnsurePath is used to make sure a path exists
+func EnsurePath(path string, dir bool) error {
+	if !dir {
+		path = filepath.Dir(path)
+	}
+	return os.MkdirAll(path, 0755)
+}

types/node_id.go

@@ -0,0 +1,4 @@
+package types
+
+// NodeID is a unique identifier for a node across space and time.
+type NodeID string

website/source/docs/agent/http/agent.html.markdown

@@ -143,6 +143,7 @@ It returns a JSON body like this:
     "DNSRecursors": [],
     "Domain": "consul.",
     "LogLevel": "INFO",
+    "NodeID": "40e4a748-2192-161a-0510-9bf59fe950b5",
     "NodeName": "foobar",
     "ClientAddr": "127.0.0.1",
     "BindAddr": "0.0.0.0",
@@ -183,6 +184,7 @@ It returns a JSON body like this:
     "Tags": {
       "bootstrap": "1",
       "dc": "dc1",
+      "id": "40e4a748-2192-161a-0510-9bf59fe950b5",
       "port": "8300",
       "role": "consul",
       "vsn": "1",

website/source/docs/agent/options.html.markdown

@@ -282,6 +282,15 @@ will exit with an error at startup.
 * <a name="_node"></a><a href="#_node">`-node`</a> - The name of this node in the cluster.
   This must be unique within the cluster. By default this is the hostname of the machine.
 
+* <a name="_node_id"></a><a href="#_node_id">`-node-id`</a> - Available in Consul 0.7.3 and later, this
+  is a unique identifier for this node across all time, even if the name of the node or address
+  changes. This must be in the form of a hex string, 36 characters long, such as
+  `adf4238a-882b-9ddc-4a9d-5b6758e4159e`. If this isn't supplied, which is the most common case, then
+  the agent will generate an identifier at startup and persist it in the <a href="#_data_dir">data directory</a>
+  so that it will remain the same across agent restarts. This is currently only exposed via the agent's
+  <a href="/docs/agent/http/agent.html#agent_self">/v1/agent/self</a> endpoint, but future versions of
+  Consul will use this to better manage cluster changes, especially for Consul servers.
+
 * <a name="_node_meta"></a><a href="#_node_meta">`-node-meta`</a> - Available in Consul 0.7.3 and later,
   this specifies an arbitrary metadata key/value pair to associate with the node, of the form `key:value`.
   This can be specified multiple times. Node metadata pairs have the following restrictions:
@@ -695,6 +704,9 @@ Consul will not enable TLS for the HTTP API unless the `https` port has been ass
 * <a name="log_level"></a><a href="#log_level">`log_level`</a> Equivalent to the
   [`-log-level` command-line flag](#_log_level).
 
+* <a name="node_id"></a><a href="#node_id">`node_id`</a> Equivalent to the
+  [`-node-id` command-line flag](#_node_id).
+
 * <a name="node_name"></a><a href="#node_name">`node_name`</a> Equivalent to the
   [`-node` command-line flag](#_node).