Merge pull request #163 from squaremo/integrate_weavedns

Integrate weavedns

.gitignore

@@ -0,0 +1,30 @@
+# Compiled Object files, Static and Dynamic libs (Shared Objects)
+*.o
+*.a
+*.so
+
+# Folders
+_obj
+_test
+
+# Architecture specific extensions/prefixes
+*.[568vq]
+[568vq].out
+
+*.cgo1.go
+*.cgo2.c
+_cgo_defun.c
+_cgo_gotypes.go
+_cgo_export.*
+
+_testmain.go
+
+*.exe
+*.test
+*.prof
+
+# Emacs backup files
+*~
+
+# Project specific
+weavedns/weavedns

Makefile

@@ -5,14 +5,17 @@
 SUDO=sudo
 
 WEAVER_EXE=weaver/weaver
+WEAVEDNS_EXE=weavedns/weavedns
 WEAVER_IMAGE=zettio/weave
+WEAVEDNS_IMAGE=zettio/weavedns
 WEAVER_EXPORT=/var/tmp/weave.tar
+WEAVEDNS_EXPORT=/var/tmp/weavedns.tar
 
-all: $(WEAVER_EXPORT)
+all: $(WEAVER_EXPORT) $(WEAVEDNS_EXPORT)
 
-$(WEAVER_EXE): router/*.go weaver/main.go
-	go get -tags netgo ./weaver
-	go build -ldflags '-extldflags "-static"' -tags netgo -o $@ ./weaver
+$(WEAVER_EXE) $(WEAVEDNS_EXE):
+	go get -tags netgo ./$(shell dirname $@)
+	go build -ldflags '-extldflags "-static"' -tags netgo -o $@ ./$(shell dirname $@)
 	@strings $@ | grep cgo_stub\\\.go >/dev/null || { \
 		rm $@; \
 		echo "\nYour go standard library was built without the 'netgo' build tag."; \
@@ -21,15 +24,25 @@ $(WEAVER_EXE): router/*.go weaver/main.go
 		false; \
 	}
 
+$(WEAVER_EXE): router/*.go weaver/main.go
+$(WEAVEDNS_EXE): nameserver/*.go weavedns/main.go
+
 $(WEAVER_EXPORT): weaver/Dockerfile $(WEAVER_EXE)
 	$(SUDO) docker build -t $(WEAVER_IMAGE) weaver
 	$(SUDO) docker save $(WEAVER_IMAGE):latest > $@
 
+$(WEAVEDNS_EXPORT): weavedns/Dockerfile $(WEAVEDNS_EXE)
+	$(SUDO) docker build -t $(WEAVEDNS_IMAGE) weavedns
+	$(SUDO) docker save $(WEAVEDNS_IMAGE):latest > $@
+
 publish: $(WEAVER_EXPORT)
 	$(SUDO) docker tag $(WEAVER_IMAGE) $(WEAVER_IMAGE):git-`git rev-parse --short=12 HEAD`
 	$(SUDO) docker push $(WEAVER_IMAGE):latest
 	$(SUDO) docker push $(WEAVER_IMAGE):git-`git rev-parse --short=12 HEAD`
+	$(SUDO) docker tag $(WEAVEDNS_IMAGE) $(WEAVEDNS_IMAGE):git-`git rev-parse --short=12 HEAD`
+	$(SUDO) docker push $(WEAVEDNS_IMAGE):latest
+	$(SUDO) docker push $(WEAVEDNS_IMAGE):git-`git rev-parse --short=12 HEAD`
 
 clean:
-	-$(SUDO) docker rmi $(WEAVER_IMAGE)
-	rm -f $(WEAVER_EXE) $(WEAVER_EXPORT)
+	-$(SUDO) docker rmi $(WEAVER_IMAGE) $(WEAVEDNS_IMAGE)
+	rm -f $(WEAVER_EXE) $(WEAVEDNS_EXE) $(WEAVER_EXPORT) $(WEAVEDNS_EXPORT)

nameserver/dns.go

@@ -0,0 +1,55 @@
+package nameserver
+
+import (
+	"github.com/miekg/dns"
+	"net"
+)
+
+const (
+	localTTL uint32 = 300 // somewhat arbitrary; we don't expect anyone
+	// downstream to cache results
+)
+
+func makeHeader(r *dns.Msg, q *dns.Question) *dns.RR_Header {
+	return &dns.RR_Header{
+		Name: q.Name, Rrtype: q.Qtype,
+		Class: dns.ClassINET, Ttl: localTTL}
+}
+
+func makeReply(r *dns.Msg, as []dns.RR) *dns.Msg {
+	m := new(dns.Msg)
+	m.SetReply(r)
+	m.RecursionAvailable = true
+	m.Answer = as
+	return m
+}
+
+func makeAReply(r *dns.Msg, q *dns.Question, addrs []net.IP) *dns.Msg {
+	answers := make([]dns.RR, len(addrs))
+	header := makeHeader(r, q)
+	count := 0
+	for _, addr := range addrs {
+		switch q.Qtype {
+		case dns.TypeA:
+			if ip4 := addr.To4(); ip4 != nil {
+				answers[count] = &dns.A{*header, addr}
+				count++
+			}
+		case dns.TypeAAAA:
+			if ip4 := addr.To4(); ip4 == nil {
+				answers[count] = &dns.AAAA{*header, addr}
+				count++
+			}
+		}
+	}
+	return makeReply(r, answers[:count])
+}
+
+func makePTRReply(r *dns.Msg, q *dns.Question, names []string) *dns.Msg {
+	answers := make([]dns.RR, len(names))
+	header := makeHeader(r, q)
+	for i, name := range names {
+		answers[i] = &dns.PTR{*header, name}
+	}
+	return makeReply(r, answers)
+}

nameserver/http.go

@@ -0,0 +1,107 @@
+package nameserver
+
+import (
+	"errors"
+	"fmt"
+	"github.com/miekg/dns"
+	"log"
+	"net"
+	"net/http"
+	"strings"
+)
+
+// Parse a URL of the form PUT /name/<identifier>/<ip-address>
+func parseUrl(url string) (identifier string, ipaddr string, err error) {
+	parts := strings.Split(url, "/")
+	if len(parts) != 4 {
+		return "", "", errors.New("Unable to parse url: " + url)
+	}
+	return parts[2], parts[3], nil
+}
+
+func httpErrorAndLog(level *log.Logger, w http.ResponseWriter, msg string,
+	status int, logmsg string, logargs ...interface{}) {
+	http.Error(w, msg, status)
+	level.Printf(logmsg, logargs...)
+}
+
+func ListenHttp(domain string, db Zone, port int) {
+	http.HandleFunc("/name/", func(w http.ResponseWriter, r *http.Request) {
+
+		reqError := func(msg string, logmsg string, logargs ...interface{}) {
+			httpErrorAndLog(Warning, w, msg, http.StatusBadRequest,
+				logmsg, logargs...)
+		}
+
+		switch r.Method {
+		case "PUT":
+			ident, weaveIPstr, err := parseUrl(r.URL.Path)
+			name := r.FormValue("fqdn")
+			localIPstr := r.FormValue("local_ip")
+			if ident == "" || weaveIPstr == "" || name == "" || localIPstr == "" {
+				reqError("Invalid request", "Invalid request: %s, %s", r.URL, r.Form)
+				return
+			}
+
+			localIP := net.ParseIP(localIPstr)
+			if localIP == nil {
+				reqError("Invalid IP in request", "Invalid IP in request: %s", localIPstr)
+				return
+			}
+
+			weaveIP := net.ParseIP(weaveIPstr)
+			if weaveIP == nil {
+				reqError("Invalid weave IP", "Invalid weave IP in request: %s", weaveIPstr)
+				return
+			}
+
+			if dns.IsSubDomain(domain, name) {
+				Info.Printf("Adding %s (%s) -> %s", name, localIPstr, weaveIPstr)
+				if err = db.AddRecord(ident, name, localIP, weaveIP); err != nil {
+					if dup, ok := err.(DuplicateError); !ok {
+						httpErrorAndLog(
+							Error, w, "Internal error", http.StatusInternalServerError,
+							"Unexpected error from DB: %s", err)
+						return
+					} else if dup.Ident != ident {
+						http.Error(w, err.Error(), http.StatusConflict)
+						return
+					} // else we are golden
+				}
+			} else {
+				Info.Printf("Ignoring name %s, not in %s", name, domain)
+			}
+
+		case "DELETE":
+			ident, weaveIPstr, err := parseUrl(r.URL.Path)
+			if ident == "" || weaveIPstr == "" {
+				reqError("Invalid Request", "Invalid request: %s, %s", r.URL, r.Form)
+				return
+			}
+
+			weaveIP := net.ParseIP(weaveIPstr)
+			if weaveIP == nil {
+				reqError("Invalid IP in request", "Invalid IP in request: %s", weaveIPstr)
+				return
+			}
+			Info.Printf("Deleting %s (%s)", ident, weaveIPstr)
+			if err = db.DeleteRecord(ident, weaveIP); err != nil {
+				if _, ok := err.(LookupError); !ok {
+					httpErrorAndLog(
+						Error, w, "Internal error", http.StatusInternalServerError,
+						"Unexpected error from DB: %s", err)
+					return
+				}
+			}
+		default:
+			msg := "Unexpected http method: " + r.Method
+			reqError(msg, msg)
+			return
+		}
+	})
+
+	address := fmt.Sprintf(":%d", port)
+	if err := http.ListenAndServe(address, nil); err != nil {
+		Error.Fatal("Unable to create http listener: ", err)
+	}
+}

nameserver/http_test.go

@@ -0,0 +1,82 @@
+package nameserver
+
+import (
+	"fmt"
+	"math/rand"
+	"net"
+	"net/http"
+	"net/url"
+	"strings"
+	"testing"
+	"time"
+)
+
+func genForm(method string, url string, data url.Values) (resp *http.Response, err error) {
+	req, err := http.NewRequest(method, url, strings.NewReader(data.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	return http.DefaultClient.Do(req)
+}
+
+func TestHttp(t *testing.T) {
+	var (
+		containerID     = "deadbeef"
+		testDomain      = "weave.local."
+		successTestName = "test1." + testDomain
+		testAddr1       = "10.0.2.1/24"
+		dockerIP        = "9.8.7.6"
+	)
+
+	var zone = new(ZoneDb)
+	port := rand.Intn(10000) + 32768
+	fmt.Println("Http test on port", port)
+	go ListenHttp(testDomain, zone, port)
+
+	time.Sleep(100 * time.Millisecond) // Allow for http server to get going
+
+	// Ask the http server to add our test address into the database
+	addrParts := strings.Split(testAddr1, "/")
+	addrUrl := fmt.Sprintf("http://localhost:%d/name/%s/%s", port, containerID, addrParts[0])
+	resp, err := genForm("PUT", addrUrl,
+		url.Values{"fqdn": {successTestName}, "local_ip": {dockerIP}, "routing_prefix": {addrParts[1]}})
+	assertNoErr(t, err)
+	assertStatus(t, resp.StatusCode, http.StatusOK, "http response")
+
+	// Check that the address is now there.
+	ip, err := zone.MatchLocal(successTestName)
+	assertNoErr(t, err)
+	weaveIP, _, _ := net.ParseCIDR(testAddr1)
+	if !ip.Equal(weaveIP) {
+		t.Fatal("Unexpected result for", successTestName, ip)
+	}
+
+	// Adding exactly the same address should be OK
+	resp, err = genForm("PUT", addrUrl,
+		url.Values{"fqdn": {successTestName}, "local_ip": {dockerIP}, "routing_prefix": {addrParts[1]}})
+	assertNoErr(t, err)
+	assertStatus(t, resp.StatusCode, http.StatusOK, "http success response for duplicate add")
+
+	// Now try adding the same address again with a different ident - should fail
+	otherUrl := fmt.Sprintf("http://localhost:%d/name/%s/%s", port, "other", addrParts[0])
+	resp, err = genForm("PUT", otherUrl,
+		url.Values{"fqdn": {successTestName}, "local_ip": {dockerIP}, "routing_prefix": {addrParts[1]}})
+	assertNoErr(t, err)
+	assertStatus(t, resp.StatusCode, http.StatusConflict, "http response")
+
+	// Delete the address
+	resp, err = genForm("DELETE", addrUrl, nil)
+	assertNoErr(t, err)
+	assertStatus(t, resp.StatusCode, http.StatusOK, "http response")
+
+	// Check that the address is not there now.
+	_, err = zone.MatchLocal(successTestName)
+	assertErrorType(t, err, (*LookupError)(nil), "nonexistent lookup")
+
+	// Delete the address again, it should accept this
+	resp, err = genForm("DELETE", addrUrl, nil)
+	assertNoErr(t, err)
+	assertStatus(t, resp.StatusCode, http.StatusOK, "http response")
+
+	// Would like to shut down the http server at the end of this test
+	// but it's complicated.
+	// See https://groups.google.com/forum/#!topic/golang-nuts/vLHWa5sHnCE
+}

nameserver/logging.go

@@ -0,0 +1,37 @@
+package nameserver
+
+import (
+	"io"
+	"io/ioutil"
+	"log"
+	"os"
+)
+
+const (
+	standardLogFlags = log.Ldate | log.Ltime | log.Lmicroseconds
+)
+
+// Largely taken from
+// http://www.goinggo.net/2013/11/using-log-package-in-go.html
+
+var (
+	Debug   *log.Logger
+	Info    *log.Logger
+	Warning *log.Logger
+	Error   *log.Logger
+)
+
+func InitLogging(debugHandle io.Writer,
+	infoHandle io.Writer,
+	warningHandle io.Writer,
+	errorHandle io.Writer) {
+
+	Debug = log.New(debugHandle, "DEBUG: ", standardLogFlags)
+	Info = log.New(infoHandle, "INFO: ", standardLogFlags)
+	Warning = log.New(warningHandle, "WARNING: ", standardLogFlags)
+	Error = log.New(errorHandle, "ERROR: ", standardLogFlags)
+}
+
+func init() {
+	InitLogging(ioutil.Discard, os.Stdout, os.Stdout, os.Stderr)
+}

nameserver/mdns_client.go

@@ -0,0 +1,251 @@
+package nameserver
+
+import (
+	"github.com/miekg/dns"
+	"math"
+	"net"
+	"time"
+)
+
+// Portions of this code taken from github.com/armon/mdns
+
+const (
+	ipv4mdns = "224.0.0.251" // link-local multicast address
+	mdnsPort = 5353          // mDNS assigned port
+	// We wait this long to hear a response from other mDNS servers on
+	// the network.
+	mDNSTimeout = 500 * time.Millisecond
+	MaxDuration = time.Duration(math.MaxInt64)
+	MailboxSize = 16
+)
+
+var (
+	ipv4Addr = &net.UDPAddr{
+		IP:   net.ParseIP(ipv4mdns),
+		Port: mdnsPort,
+	}
+)
+
+type ResponseA struct {
+	Name string
+	Addr net.IP
+	Err  error
+}
+
+type responseInfo struct {
+	timeout time.Time // if no answer by this time, give up
+	ch      chan<- *ResponseA
+}
+
+// Represents one query that we have sent for one name.
+// If we, internally, get several requests for the same name while we have
+// a query in flight, then we don't want to send more queries out.
+// Invariant on responseInfos: they are in ascending order of timeout.
+type inflightQuery struct {
+	name          string
+	id            uint16 // the DNS message ID
+	responseInfos []*responseInfo
+}
+
+type MDNSClient struct {
+	server    *dns.Server
+	conn      *net.UDPConn
+	addr      *net.UDPAddr
+	inflight  map[string]*inflightQuery
+	queryChan chan<- *MDNSInteraction
+}
+
+type mDNSQueryInfo struct {
+	name       string
+	querytype  uint16
+	responseCh chan<- *ResponseA
+}
+
+func NewMDNSClient() (*MDNSClient, error) {
+	if conn, err := net.ListenUDP("udp4", &net.UDPAddr{IP: net.IPv4zero, Port: 0}); err != nil {
+		return nil, err
+	} else {
+		return &MDNSClient{
+			conn:     conn,
+			addr:     ipv4Addr,
+			inflight: make(map[string]*inflightQuery)}, nil
+	}
+}
+
+func (c *MDNSClient) Start(ifi *net.Interface) error {
+	multicast, err := LinkLocalMulticastListener(ifi)
+	if err != nil {
+		return err
+	}
+
+	handleMDNS := func(w dns.ResponseWriter, r *dns.Msg) {
+		// Don't want to handle queries here, so filter anything out that isn't a response
+		if len(r.Answer) > 0 {
+			c.ResponseCallback(r)
+		}
+	}
+
+	c.server = &dns.Server{Listener: nil, PacketConn: multicast, Handler: dns.HandlerFunc(handleMDNS)}
+	go c.server.ActivateAndServe()
+
+	queryChan := make(chan *MDNSInteraction, MailboxSize)
+	c.queryChan = queryChan
+	go c.queryLoop(queryChan)
+
+	return nil
+}
+
+func LinkLocalMulticastListener(ifi *net.Interface) (net.PacketConn, error) {
+	return net.ListenMulticastUDP("udp", ifi, ipv4Addr)
+}
+
+// ACTOR client API
+
+const (
+	CSendQuery       = iota
+	CShutdown        = iota
+	CMessageReceived = iota
+)
+
+type MDNSInteraction struct {
+	code       int
+	resultChan chan<- interface{}
+	payload    interface{}
+}
+
+// Async
+func (c *MDNSClient) Shutdown() {
+	c.queryChan <- &MDNSInteraction{code: CShutdown}
+}
+
+// Async
+func (c *MDNSClient) SendQuery(name string, querytype uint16, responseCh chan<- *ResponseA) {
+	c.queryChan <- &MDNSInteraction{
+		code:    CSendQuery,
+		payload: mDNSQueryInfo{name, querytype, responseCh},
+	}
+}
+
+// Async - called from dns library multiplexer
+func (c *MDNSClient) ResponseCallback(r *dns.Msg) {
+	c.queryChan <- &MDNSInteraction{code: CMessageReceived, payload: r}
+}
+
+// ACTOR server
+
+// Check all in-flight queries, close all that have already timed out,
+// and return the duration until the next timeout
+func (c *MDNSClient) checkInFlightQueries() time.Duration {
+	now := time.Now()
+	after := MaxDuration
+	for name, query := range c.inflight {
+		// Invariant on responseInfos: they are in ascending order of timeout.
+		numClosed := 0
+		for _, item := range query.responseInfos {
+			duration := item.timeout.Sub(now)
+			if duration <= 0 { // timed out
+				close(item.ch)
+				numClosed++
+			} else {
+				if duration < after {
+					after = duration
+				}
+				break // don't need to look at any more for this query
+			}
+		}
+		// Remove timed-out items from the slice
+		query.responseInfos = query.responseInfos[numClosed:]
+		if len(query.responseInfos) == 0 {
+			delete(c.inflight, name)
+		}
+	}
+	return after
+}
+
+func (c *MDNSClient) queryLoop(queryChan <-chan *MDNSInteraction) {
+	timer := time.NewTimer(MaxDuration)
+	run := func() {
+		timer.Reset(c.checkInFlightQueries())
+	}
+
+	terminate := false
+	for !terminate {
+		select {
+		case query, ok := <-queryChan:
+			if !ok {
+				break
+			}
+			switch query.code {
+			case CShutdown:
+				c.server.Shutdown()
+				terminate = true
+			case CSendQuery:
+				c.handleSendQuery(query.payload.(mDNSQueryInfo))
+				run()
+			case CMessageReceived:
+				c.handleResponse(query.payload.(*dns.Msg))
+				run()
+			}
+		case <-timer.C:
+			run()
+		}
+	}
+
+	// Close all response channels at termination
+	for _, query := range c.inflight {
+		for _, item := range query.responseInfos {
+			close(item.ch)
+		}
+	}
+}
+
+func (c *MDNSClient) handleSendQuery(q mDNSQueryInfo) {
+	query, found := c.inflight[q.name]
+	if !found {
+		m := new(dns.Msg)
+		m.SetQuestion(q.name, q.querytype)
+		m.RecursionDesired = false
+		buf, err := m.Pack()
+		if err != nil {
+			q.responseCh <- &ResponseA{Err: err}
+			close(q.responseCh)
+			return
+		}
+		query = &inflightQuery{
+			name: q.name,
+			id:   m.Id,
+		}
+		if _, err = c.conn.WriteTo(buf, c.addr); err != nil {
+			q.responseCh <- &ResponseA{Err: err}
+			close(q.responseCh)
+			return
+		}
+		c.inflight[q.name] = query
+	}
+	info := &responseInfo{
+		ch:      q.responseCh,
+		timeout: time.Now().Add(mDNSTimeout),
+	}
+	// Invariant on responseInfos: they are in ascending order of timeout.
+	// Since we use a fixed interval from Now(), this must be after all existing timeouts.
+	query.responseInfos = append(query.responseInfos, info)
+}
+
+func (c *MDNSClient) handleResponse(r *dns.Msg) {
+	for _, answer := range r.Answer {
+		switch rr := answer.(type) {
+		case *dns.A:
+			name := rr.Hdr.Name
+			if query, found := c.inflight[name]; found {
+				for _, resp := range query.responseInfos {
+					resp.ch <- &ResponseA{Name: rr.Hdr.Name, Addr: rr.A}
+					close(resp.ch)
+				}
+				delete(c.inflight, name)
+			} else {
+				// We've received a response that didn't match a query
+				// Do we want to cache it?
+			}
+		}
+	}
+}

nameserver/mdns_client_test.go

@@ -0,0 +1,153 @@
+package nameserver
+
+import (
+	"github.com/miekg/dns"
+	"log"
+	"net"
+	"testing"
+	"time"
+)
+
+var (
+	successTestName = "test1.weave."
+	failTestName    = "test2.weave."
+	testAddr        = net.ParseIP("9.8.7.6")
+)
+
+func minimalServer(w dns.ResponseWriter, req *dns.Msg) {
+	//log.Println("minimalServer received:", req)
+	if len(req.Answer) > 0 {
+		return // Only interested in questions.
+	}
+	if len(req.Question) != 1 {
+		return // We only handle single-question messages
+	}
+	if req.Question[0].Name != successTestName {
+		return // This is not the DNS record you are looking for
+	}
+
+	m := new(dns.Msg)
+	m.SetReply(req)
+
+	hdr := dns.RR_Header{Name: m.Question[0].Name, Rrtype: dns.TypeA,
+		Class: dns.ClassINET, Ttl: 3600}
+	a := &dns.A{hdr, testAddr}
+	m.Answer = append(m.Answer, a)
+
+	buf, err := m.Pack()
+	checkFatal(err)
+	if buf == nil {
+		log.Fatal("Nil buffer")
+	}
+	//log.Println("minimalServer sending:", buf)
+	// This is a bit of a kludge - per the RFC we should send responses from 5353, but that doesn't seem to work
+	sendconn, err := net.DialUDP("udp", nil, ipv4Addr)
+	checkFatal(err)
+
+	_, err = sendconn.Write(buf)
+	sendconn.Close()
+	checkFatal(err)
+}
+
+func RunLocalMulticastServer() (*dns.Server, error) {
+	multicast, err := net.ListenMulticastUDP("udp", nil, ipv4Addr)
+	if err != nil {
+		return nil, err
+	}
+	server := &dns.Server{Listener: nil, PacketConn: multicast, Handler: dns.HandlerFunc(minimalServer)}
+	go server.ActivateAndServe()
+	return server, nil
+}
+
+func setup(t *testing.T) (*MDNSClient, *dns.Server, error) {
+	mdnsClient, err := NewMDNSClient()
+	assertNoErr(t, err)
+	err = mdnsClient.Start(nil)
+	assertNoErr(t, err)
+
+	server, err := RunLocalMulticastServer()
+	if err != nil {
+		t.Fatalf("Unable to run test server: %s", err)
+	}
+	return mdnsClient, server, err
+}
+
+type testContext struct {
+	receivedAddr  net.IP
+	receivedCount int
+	channel       chan *ResponseA
+}
+
+func newTestContext() *testContext {
+	return &testContext{channel: make(chan *ResponseA, 4)}
+}
+
+func (c *testContext) checkResponse(t *testing.T, channelOk bool, resp *ResponseA) {
+	if !channelOk {
+		c.channel = nil
+		return
+	}
+	assertNoErr(t, resp.Err)
+	log.Printf("Got address response %s addr %s", resp.Name, resp.Addr)
+	c.receivedAddr = resp.Addr
+	c.receivedCount++
+}
+
+func TestSimpleQuery(t *testing.T) {
+	log.Println("TestSimpleQuery starting")
+	mdnsClient, server, _ := setup(t)
+	defer mdnsClient.Shutdown()
+	defer server.Shutdown()
+
+	context := newTestContext()
+
+	// First, a test we expect to succeed
+	mdnsClient.SendQuery(successTestName, dns.TypeA, context.channel)
+	for resp := range context.channel {
+		context.checkResponse(t, true, resp)
+	}
+
+	if !context.receivedAddr.Equal(testAddr) || context.receivedCount != 1 {
+		t.Fatal("Unexpected result for", successTestName, context.receivedAddr, context.receivedCount)
+	}
+
+	// Now, a test we expect to time out with no responses
+	context = newTestContext()
+	mdnsClient.SendQuery("test2.weave.", dns.TypeA, context.channel)
+	for resp := range context.channel {
+		context.checkResponse(t, true, resp)
+	}
+
+	if context.receivedCount > 0 {
+		t.Fatal("Unexpected result for test2.weave", context.receivedAddr)
+	}
+}
+
+func TestParallelQuery(t *testing.T) {
+	log.Println("TestParallelQuery starting")
+	mdnsClient, server, _ := setup(t)
+	defer mdnsClient.Shutdown()
+	defer server.Shutdown()
+
+	context1 := newTestContext()
+	context2 := newTestContext()
+
+	go mdnsClient.SendQuery(successTestName, dns.TypeA, context1.channel)
+	go mdnsClient.SendQuery(successTestName, dns.TypeA, context2.channel)
+	timeout := time.After(2 * time.Second)
+outerloop:
+	for context1.channel != nil || context2.channel != nil {
+		select {
+		case resp, ok := <-context1.channel:
+			context1.checkResponse(t, ok, resp)
+		case resp, ok := <-context2.channel:
+			context2.checkResponse(t, ok, resp)
+		case <-timeout:
+			break outerloop
+		}
+	}
+
+	if !context1.receivedAddr.Equal(testAddr) || !context2.receivedAddr.Equal(testAddr) || context1.receivedCount != 1 || context2.receivedCount != 1 {
+		t.Fatal("Unexpected result for", successTestName, context1.receivedAddr, context2.receivedAddr, context1.receivedCount, context2.receivedCount)
+	}
+}

nameserver/mdns_server.go

@@ -0,0 +1,79 @@
+package nameserver
+
+import (
+	"github.com/miekg/dns"
+	"net"
+)
+
+type MDNSServer struct {
+	localAddrs []net.Addr
+	sendconn   *net.UDPConn
+	zone       Zone
+}
+
+func NewMDNSServer(zone Zone) (*MDNSServer, error) {
+	// This is a bit of a kludge - per the RFC we should send responses from 5353, but that doesn't seem to work
+	sendconn, err := net.ListenUDP("udp4", &net.UDPAddr{IP: net.IPv4zero, Port: 0})
+	if err != nil {
+		return nil, err
+	}
+	return &MDNSServer{sendconn: sendconn, zone: zone}, nil
+}
+
+// Return true if testaddr is a UDP address with IP matching my local i/f
+func (s *MDNSServer) addrIsLocal(testaddr net.Addr) bool {
+	if udpaddr, ok := testaddr.(*net.UDPAddr); ok {
+		for _, localaddr := range s.localAddrs {
+			if ipnetlocal, ok := localaddr.(*net.IPNet); ok {
+				if ipnetlocal.IP.Equal(udpaddr.IP) {
+					return true
+				}
+			}
+		}
+	}
+	return false
+}
+
+func (s *MDNSServer) Start(ifi *net.Interface) error {
+	conn, err := LinkLocalMulticastListener(ifi)
+	if err != nil {
+		return err
+	}
+
+	if ifi == nil {
+		s.localAddrs, err = net.InterfaceAddrs()
+	} else {
+		s.localAddrs, err = ifi.Addrs()
+	}
+	if err != nil {
+		return err
+	}
+
+	handleMDNS := func(w dns.ResponseWriter, r *dns.Msg) {
+		// Ignore answers to other questions
+		if len(r.Answer) == 0 && len(r.Question) > 0 {
+			q := r.Question[0]
+			if ip, err := s.zone.MatchLocal(q.Name); err == nil {
+				m := makeAReply(r, &q, []net.IP{ip})
+				if err = s.sendResponse(m); err != nil {
+					Warning.Printf("Error writing to %s", w)
+				}
+			} else if s.addrIsLocal(w.RemoteAddr()) {
+				// ignore this - it's our own query received via multicast
+			} else {
+				Debug.Printf("Failed MDNS lookup for %s", q.Name)
+			}
+		}
+	}
+	go dns.ActivateAndServe(nil, conn, dns.HandlerFunc(handleMDNS))
+	return err
+}
+
+func (s *MDNSServer) sendResponse(m *dns.Msg) error {
+	buf, err := m.Pack()
+	if err != nil {
+		return err
+	}
+	_, err = s.sendconn.WriteTo(buf, ipv4Addr)
+	return err
+}

nameserver/mdns_server_test.go

@@ -0,0 +1,88 @@
+package nameserver
+
+import (
+	"github.com/miekg/dns"
+	"log"
+	"net"
+	"testing"
+	"time"
+)
+
+var (
+	containerID = "deadbeef"
+	testAddr1   = "10.0.2.1/24"
+)
+
+func sendQuery(name string, querytype uint16) error {
+	m := new(dns.Msg)
+	m.SetQuestion(name, querytype)
+	m.RecursionDesired = false
+	buf, err := m.Pack()
+	if err != nil {
+		return err
+	}
+	conn, err := net.ListenUDP("udp4", &net.UDPAddr{IP: net.IPv4zero, Port: 0})
+	if err != nil {
+		return err
+	}
+	_, err = conn.WriteTo(buf, ipv4Addr)
+	return err
+}
+
+func TestServerSimpleQuery(t *testing.T) {
+	log.Println("TestServerSimpleQuery starting")
+	var zone = new(ZoneDb)
+	dockerIP := net.ParseIP("9.8.7.6")
+	weaveIP, _, _ := net.ParseCIDR(testAddr1)
+	zone.AddRecord(containerID, "test.weave.", dockerIP, weaveIP)
+
+	mdnsServer, err := NewMDNSServer(zone)
+	assertNoErr(t, err)
+	err = mdnsServer.Start(nil)
+	assertNoErr(t, err)
+
+	var receivedAddr net.IP
+	receivedCount := 0
+
+	// Implement a minimal listener for responses
+	multicast, err := LinkLocalMulticastListener(nil)
+	assertNoErr(t, err)
+
+	handleMDNS := func(w dns.ResponseWriter, r *dns.Msg) {
+		// Only handle responses here
+		if len(r.Answer) > 0 {
+			for _, answer := range r.Answer {
+				switch rr := answer.(type) {
+				case *dns.A:
+					receivedAddr = rr.A
+					receivedCount++
+				}
+			}
+		}
+	}
+
+	server := &dns.Server{Listener: nil, PacketConn: multicast, Handler: dns.HandlerFunc(handleMDNS)}
+	go server.ActivateAndServe()
+	defer server.Shutdown()
+
+	time.Sleep(100 * time.Millisecond) // Allow for server to get going
+
+	sendQuery("test.weave.", dns.TypeA)
+
+	time.Sleep(time.Second)
+
+	if receivedCount != 1 {
+		t.Fatal("Unexpected result count for test.weave", receivedCount)
+	}
+	if !receivedAddr.Equal(weaveIP) {
+		t.Fatal("Unexpected result for test.weave", receivedAddr)
+	}
+
+	receivedCount = 0
+
+	sendQuery("testfail.weave.", dns.TypeA)
+
+	if receivedCount != 0 {
+		t.Fatal("Unexpected result count for testfail.weave", receivedCount)
+	}
+}

nameserver/server.go

@@ -0,0 +1,152 @@
+package nameserver
+
+import (
+	"fmt"
+	"github.com/miekg/dns"
+	"net"
+)
+
+const (
+	LOCAL_DOMAIN = "weave.local."
+	RDNS_DOMAIN  = "in-addr.arpa."
+)
+
+// +1 to also exclude a dot
+var rdnsDomainLen = len(RDNS_DOMAIN) + 1
+
+func checkFatal(e error) {
+	if e != nil {
+		Error.Fatal(e)
+	}
+}
+
+func checkWarn(e error) {
+	if e != nil {
+		Warning.Println(e)
+	}
+}
+
+func makeDNSFailResponse(r *dns.Msg) *dns.Msg {
+	m := new(dns.Msg)
+	m.SetReply(r)
+	m.RecursionAvailable = true
+	m.Rcode = dns.RcodeNameError
+	return m
+}
+
+func queryHandler(zone Zone, mdnsClient *MDNSClient) dns.HandlerFunc {
+	return func(w dns.ResponseWriter, r *dns.Msg) {
+		q := r.Question[0]
+		Debug.Printf("Local query: %+v", q)
+		if q.Qtype == dns.TypeA {
+			if ip, err := zone.MatchLocal(q.Name); err == nil {
+				m := makeAReply(r, &q, []net.IP{ip})
+				w.WriteMsg(m)
+			} else {
+				Debug.Printf("Failed lookup for %s; sending mDNS query", q.Name)
+				// We don't know the answer; see if someone else does
+				channel := make(chan *ResponseA)
+				replies := make([]net.IP, 0)
+				go func() {
+					for resp := range channel {
+						Debug.Printf("Got address response %s to query %s addr %s", resp.Name, q.Name, resp.Addr)
+						replies = append(replies, resp.Addr)
+					}
+					var responseMsg *dns.Msg
+					if len(replies) > 0 {
+						responseMsg = makeAReply(r, &q, replies)
+					} else {
+						responseMsg = makeDNSFailResponse(r)
+					}
+					w.WriteMsg(responseMsg)
+				}()
+				mdnsClient.SendQuery(q.Name, dns.TypeA, channel)
+			}
+		} else {
+			Warning.Printf("Local query not handled: %+v", q)
+			m := makeDNSFailResponse(r)
+			w.WriteMsg(m)
+		}
+		return
+	}
+}
+
+func rdnsHandler(zone Zone, mdnsClient *MDNSClient) dns.HandlerFunc {
+	return func(w dns.ResponseWriter, r *dns.Msg) {
+		q := r.Question[0]
+		Debug.Printf("Local rdns query: %+v", q)
+		if q.Qtype == dns.TypePTR {
+			if ip := net.ParseIP(q.Name[:len(q.Name)-rdnsDomainLen]); ip != nil {
+				ip4 := ip.To4()
+				revIP := []byte{ip4[3], ip4[2], ip4[1], ip4[0]}
+				Debug.Printf("Looking for address: %+v", revIP)
+				if name, err := zone.MatchLocalIP(revIP); err == nil {
+					Debug.Printf("Found name: %s", name)
+					m := makePTRReply(r, &q, []string{name})
+					w.WriteMsg(m)
+				} else {
+					Debug.Printf("Failed lookup for %s; sending mDNS query", q.Name)
+					// We don't know the answer; see if someone else does
+					// TODO
+				}
+			}
+		}
+	}
+}
+
+/* When we receive a request for a name outside of our '.weave' domain, call
+   the underlying lookup mechanism and return the answer(s) it gives.
+   Unfortunately, this means that TTLs from a real DNS server are lost - FIXME.
+*/
+func notUsHandler() dns.HandlerFunc {
+	return func(w dns.ResponseWriter, r *dns.Msg) {
+		q := r.Question[0]
+		Debug.Printf("Non-local query: %+v", q)
+		var responseMsg *dns.Msg
+		if q.Qtype == dns.TypeA {
+			if addrs, err := net.LookupIP(q.Name); err == nil {
+				responseMsg = makeAReply(r, &q, addrs)
+			} else {
+				responseMsg = makeDNSFailResponse(r)
+				Debug.Print("Failed fallback lookup ", err)
+			}
+		} else {
+			Warning.Printf("Non-local query not handled: %+v", q)
+			responseMsg = makeDNSFailResponse(r)
+		}
+		w.WriteMsg(responseMsg)
+	}
+}
+
+func StartServer(zone Zone, iface *net.Interface, dnsPort int, httpPort int, wait int) error {
+	go ListenHttp(LOCAL_DOMAIN, zone, httpPort)
+
+	mdnsClient, err := NewMDNSClient()
+	checkFatal(err)
+
+	ifaceName := "default interface"
+	if iface != nil {
+		ifaceName = iface.Name
+	}
+	Info.Printf("Using mDNS on %s", ifaceName)
+	err = mdnsClient.Start(iface)
+	checkFatal(err)
+
+	LocalServeMux := dns.NewServeMux()
+	LocalServeMux.HandleFunc(LOCAL_DOMAIN, queryHandler(zone, mdnsClient))
+	LocalServeMux.HandleFunc(RDNS_DOMAIN, rdnsHandler(zone, mdnsClient))
+	LocalServeMux.HandleFunc(".", notUsHandler())
+
+	mdnsServer, err := NewMDNSServer(zone)
+	checkFatal(err)
+
+	err = mdnsServer.Start(iface)
+	checkFatal(err)
+
+	address := fmt.Sprintf(":%d", dnsPort)
+	Info.Printf("Listening for DNS on %s", address)
+	err = dns.ListenAndServe(address, "udp", LocalServeMux)
+	checkFatal(err)
+
+	return nil
+}

nameserver/testing_utils.go

@@ -0,0 +1,32 @@
+package nameserver
+
+import (
+	"reflect"
+	"testing"
+)
+
+func assertNoErr(t *testing.T, err error) {
+	if err != nil {
+		t.Fatal(err)
+	}
+}
+
+func assertStatus(t *testing.T, got int, wanted int, desc string) {
+	if got != wanted {
+		t.Fatalf("Expected %s %d but got %d", desc, wanted, got)
+	}
+}
+
+func assertErrorInterface(t *testing.T, got interface{}, wanted interface{}, desc string) {
+	gotT, wantedT := reflect.TypeOf(got), reflect.TypeOf(wanted).Elem()
+	if !gotT.Implements(wantedT) {
+		t.Fatalf("Expected %s but got %s (%s)", wantedT.String(), gotT.String(), desc)
+	}
+}
+
+func assertErrorType(t *testing.T, got interface{}, wanted interface{}, desc string) {
+	gotT, wantedT := reflect.TypeOf(got), reflect.TypeOf(wanted).Elem()
+	if gotT != wantedT {
+		t.Fatalf("Expected %s but got %s (%s)", wantedT.String(), gotT.String(), desc)
+	}
+}

nameserver/updater.go

@@ -0,0 +1,34 @@
+package nameserver
+
+import (
+	"github.com/fsouza/go-dockerclient"
+)
+
+func StartUpdater(apiPath string, zone Zone) error {
+	client, err := docker.NewClient(apiPath)
+	if err != nil {
+		Error.Fatalf("Unable to connect to Docker API on %s: %s", apiPath, err)
+	}
+
+	events := make(chan *docker.APIEvents)
+	client.AddEventListener(events)
+
+	Info.Printf("Using Docker API on %s", apiPath)
+
+	go func() {
+		for event := range events {
+			handleEvent(zone, event, client)
+		}
+	}()
+	return nil
+}
+
+func handleEvent(zone Zone, event *docker.APIEvents, client *docker.Client) error {
+	switch event.Status {
+	case "die":
+		id := event.ID
+		Info.Printf("Container %s down. Removing records", id)
+		zone.DeleteRecordsFor(id)
+	}
+	return nil
+}

nameserver/zone.go

@@ -0,0 +1,115 @@
+package nameserver
+
+import (
+	"github.com/miekg/dns"
+	"net"
+	"sync"
+)
+
+type Zone interface {
+	AddRecord(ident string, name string, localIP net.IP, weaveIP net.IP) error
+	DeleteRecord(ident string, weaveIp net.IP) error
+	DeleteRecordsFor(ident string) error
+	MatchLocal(name string) (net.IP, error)
+	MatchLocalIP(ip net.IP) (string, error)
+}
+
+type Record struct {
+	Ident   string
+	Name    string
+	LocalIP net.IP
+	WeaveIP net.IP
+}
+
+// Very simple data structure for now, with linear searching.
+// TODO: make more sophisticated to improve performance.
+type ZoneDb struct {
+	mx   sync.RWMutex
+	recs []Record
+}
+
+type LookupError string
+
+func (ops LookupError) Error() string {
+	return "Unable to find " + string(ops)
+}
+
+type DuplicateError struct {
+	Name    string
+	WeaveIP net.IP
+	Ident   string
+}
+
+func (err DuplicateError) Error() string {
+	return "Duplicate " + err.Name + "," + err.WeaveIP.String() + " for identity " + err.Ident
+}
+
+func (zone *ZoneDb) indexOf(match func(Record) bool) int {
+	for i, r := range zone.recs {
+		if match(r) {
+			return i
+		}
+	}
+	return -1
+}
+
+func (zone *ZoneDb) MatchLocal(name string) (net.IP, error) {
+	zone.mx.RLock()
+	defer zone.mx.RUnlock()
+	for _, r := range zone.recs {
+		if r.Name == name {
+			return r.WeaveIP, nil
+		}
+	}
+	return nil, LookupError(name)
+}
+
+func (zone *ZoneDb) MatchLocalIP(ip net.IP) (string, error) {
+	zone.mx.RLock()
+	defer zone.mx.RUnlock()
+	for _, r := range zone.recs {
+		if r.WeaveIP.Equal(ip) {
+			return r.Name, nil
+		}
+	}
+	return "", LookupError(ip.String())
+}
+
+func (zone *ZoneDb) AddRecord(ident string, name string, localIP net.IP, weaveIP net.IP) error {
+	zone.mx.Lock()
+	defer zone.mx.Unlock()
+	fqdn := dns.Fqdn(name)
+	if index := zone.indexOf(
+		func(r Record) bool { return r.Name == fqdn && r.WeaveIP.Equal(weaveIP) }); index != -1 {
+		return DuplicateError{fqdn, weaveIP, zone.recs[index].Ident}
+	}
+	zone.recs = append(zone.recs, Record{ident, fqdn, localIP, weaveIP})
+	return nil
+}
+
+func (zone *ZoneDb) DeleteRecord(ident string, weaveIP net.IP) error {
+	zone.mx.Lock()
+	defer zone.mx.Unlock()
+	if index := zone.indexOf(
+		func(r Record) bool { return r.Ident == ident && r.WeaveIP.Equal(weaveIP) }); index == -1 {
+		return LookupError(ident)
+	} else {
+		zone.recs = append(zone.recs[:index], zone.recs[index+1:]...)
+	}
+	return nil
+}
+
+func (zone *ZoneDb) DeleteRecordsFor(ident string) error {
+	zone.mx.Lock()
+	defer zone.mx.Unlock()
+	w := 0 // write index
+
+	for _, r := range zone.recs {
+		if r.Ident != ident {
+			zone.recs[w] = r
+			w++
+		}
+	}
+	zone.recs = zone.recs[:w]
+	return nil
+}

nameserver/zone_test.go

@@ -0,0 +1,84 @@
+package nameserver
+
+import (
+	"net"
+	"testing"
+)
+
+func TestZone(t *testing.T) {
+	var (
+		containerID     = "deadbeef"
+		successTestName = "test1.weave."
+		testAddr1       = "10.0.2.1/24"
+		dockerIP        = "9.8.7.6"
+	)
+
+	var zone = new(ZoneDb)
+
+	ip := net.ParseIP(dockerIP)
+	weaveIP, _, _ := net.ParseCIDR(testAddr1)
+	err := zone.AddRecord(containerID, successTestName, ip, weaveIP)
+	assertNoErr(t, err)
+
+	// Add a few more records to make the job harder
+	err = zone.AddRecord("abcdef0123", "adummy.weave.", net.ParseIP("1.1.1.1"), net.ParseIP("10.0.0.1"))
+	assertNoErr(t, err)
+	err = zone.AddRecord("0123abcdef", "zdummy.weave.", net.ParseIP("2.2.2.2"), net.ParseIP("10.0.0.2"))
+	assertNoErr(t, err)
+
+	// Check that the address is now there.
+	foundIP, err := zone.MatchLocal(successTestName)
+	assertNoErr(t, err)
+
+	if !foundIP.Equal(weaveIP) {
+		t.Fatal("Unexpected result for", successTestName, foundIP)
+	}
+
+	// See if we can find the address by IP.
+	foundName, err := zone.MatchLocalIP(weaveIP)
+	assertNoErr(t, err)
+
+	if foundName != successTestName {
+		t.Fatal("Unexpected result for", weaveIP, foundName)
+	}
+
+	// Now try to add the same address again
+	err = zone.AddRecord(containerID, successTestName, ip, weaveIP)
+	assertErrorType(t, err, (*DuplicateError)(nil), "duplicate add")
+
+	// Now delete the record
+	err = zone.DeleteRecord(containerID, weaveIP)
+	assertNoErr(t, err)
+
+	// Check that the address is not there now.
+	_, err = zone.MatchLocal(successTestName)
+	assertErrorType(t, err, (*LookupError)(nil), "after deleting record")
+
+	// Delete a record that isn't there
+	err = zone.DeleteRecord(containerID, net.ParseIP("0.0.0.0"))
+	assertErrorType(t, err, (*LookupError)(nil), "when deleting record that doesn't exist")
+}
+
+func TestDeleteFor(t *testing.T) {
+	var (
+		id       = "foobar"
+		name     = "foo.weave."
+		addr1    = "10.1.2.3/24"
+		addr2    = "10.2.7.8/24"
+		dockerIP = "172.16.0.4"
+	)
+	zone := new(ZoneDb)
+	ip := net.ParseIP(dockerIP)
+	for _, addr := range []string{addr1, addr2} {
+		weaveIP, _, _ := net.ParseCIDR(addr)
+		err := zone.AddRecord(id, name, ip, weaveIP)
+		assertNoErr(t, err)
+	}
+
+	_, err := zone.MatchLocal(name)
+	assertNoErr(t, err)
+
+	err = zone.DeleteRecordsFor(id)
+	_, err = zone.MatchLocal(name)
+	assertErrorType(t, err, (*LookupError)(nil), "after deleting records for ident")
+}

net/if.go

@@ -0,0 +1,28 @@
+package net
+
+import (
+	"fmt"
+	"net"
+	"time"
+)
+
+func EnsureInterface(ifaceName string, wait int) (iface *net.Interface, err error) {
+	if iface, err = findInterface(ifaceName); err == nil || wait == 0 {
+		return
+	}
+	for ; err != nil && wait > 0; wait -= 1 {
+		time.Sleep(1 * time.Second)
+		iface, err = findInterface(ifaceName)
+	}
+	return
+}
+
+func findInterface(ifaceName string) (iface *net.Interface, err error) {
+	if iface, err = net.InterfaceByName(ifaceName); err != nil {
+		return iface, fmt.Errorf("Unable to find interface %s", ifaceName)
+	}
+	if 0 == (net.FlagUp & iface.Flags) {
+		return iface, fmt.Errorf("Interface %s is not up", ifaceName)
+	}
+	return
+}

weave

@@ -3,14 +3,15 @@ set -e
 
 usage() {
     echo "Usage:"
-    echo "weave launch  [-password <password>] <peer> ..."
-    echo "weave connect <peer>"
-    echo "weave run     <cidr> <docker run args> ..."
-    echo "weave start   <cidr> <container_id>"
-    echo "weave attach  <cidr> <container_id>"
-    echo "weave detach  <cidr> <container_id>"
-    echo "weave expose  <cidr>"
-    echo "weave hide    <cidr>"
+    echo "weave launch     [-password <password>] <peer> ..."
+    echo "weave launch-dns <cidr>"
+    echo "weave connect    <peer>"
+    echo "weave run        <cidr> <docker run args> ..."
+    echo "weave start      <cidr> <container_id>"
+    echo "weave attach     <cidr> <container_id>"
+    echo "weave detach     <cidr> <container_id>"
+    echo "weave expose     <cidr>"
+    echo "weave hide       <cidr>"
     echo "weave ps"
     echo "weave status"
     echo "weave version"
@@ -36,12 +37,15 @@ command_exists () {
 [ $# -gt 0 ] || usage
 
 IMAGE=zettio/weave
+DNS_IMAGE=zettio/weavedns
 CONTAINER_NAME=weave
+DNS_CONTAINER_NAME=weavedns
 BRIDGE=weave
 CONTAINER_IFNAME=ethwe
 MTU=65535
 PORT=6783
 HTTP_PORT=6784
+DNS_HTTP_PORT=6785
 
 COMMAND=$1
 
@@ -263,29 +267,72 @@ detach() {
     ip link del $LOCAL_IFNAME type veth
 }
 
+# Call url $4 on container $2 at port $3 with http verb $1
+http_call() {
+    http_verb="$1"
+    if ! DOCKERIP=$(docker inspect --format='{{ .NetworkSettings.IPAddress }}' $2 2>/dev/null) ; then
+        echo "$2 is not running." >&2
+        exit 1
+    fi
+    if [ -z "$DOCKERIP" ] ; then
+        echo "$2 is not running." >&2
+        exit 1
+    fi
+    port="$3"
+    url="$4"
+    shift 3
+    if command_exists curl ; then
+        curl -s -X $http_verb "$@" http://$DOCKERIP:$port$url
+    else
+        echo "Cannot find curl, which is required for this operation." >&2
+        exit 1
+    fi
+}
+
+# Perform operation $1 on container ID $2 to local DNS database at address $3
+# This function is only called where we know $2 is a valid container name
+tell_dns() {
+    if ! DNSID=$(docker inspect --format='{{ .Id }}' $DNS_CONTAINER_NAME 2>/dev/null) || [ -z "$DNSID" ] ; then
+        # weavedns not running - silently return
+        return
+    fi
+    # get the long form of the container ID
+    CONTAINER=$(docker inspect --format '{{.Id}}' $2 2>/dev/null)
+    # extract IP address and routing prefix from CIDR
+    WEAVE_ADDR_IP=$(echo $3 | sed -e 's/\([^/]*\)\/\(.*\)/\1/')
+    WEAVE_ADDR_PR=$(echo $3 | sed -e 's/\([^/]*\)\/\(.*\)/\2/')
+    MORE_ARGS=$(docker inspect --format '--data-urlencode local_ip={{ .NetworkSettings.IPAddress }} --data-urlencode fqdn={{ .Config.Hostname }}.{{ .Config.Domainname }}.' $CONTAINER 2>/dev/null) && true
+    http_call $1 $DNS_CONTAINER_NAME $DNS_HTTP_PORT /name/$CONTAINER/$WEAVE_ADDR_IP $MORE_ARGS
+}
+
+# Check that a container named $1 with image $2 is not running
+check_not_running() {
+    case $(docker inspect --format '{{ .State.Running }} {{ .Config.Image }}' $1 2>/dev/null) in
+        "true $2")
+            echo "$1 is already running." >&2
+            exit 1
+            ;;
+        "false $2")
+            docker rm $1 >/dev/null
+            ;;
+        true*)
+            echo "Found another running container named '$1'. Aborting." >&2
+            exit 1
+            ;;
+        false*)
+            echo "Found another container named '$1'. Aborting." >&2
+            exit 1
+            ;;
+    esac
+}
+
 container_weave_addrs() {
     ip netns exec $NETNS ip addr show dev $CONTAINER_IFNAME
 }
 
 case "$COMMAND" in
     launch)
-        case $(docker inspect --format '{{ .State.Running }} {{ .Config.Image }}' $CONTAINER_NAME 2>/dev/null) in
-            "true $IMAGE")
-                echo "Weave is already running." >&2
-                exit 1
-                ;;
-            "false $IMAGE")
-                docker rm $CONTAINER_NAME >/dev/null
-                ;;
-            true*)
-                echo "Found another running container named '$CONTAINER_NAME'. Aborting." >&2
-                exit 1
-                ;;
-            false*)
-                echo "Found another container named '$CONTAINER_NAME'. Aborting." >&2
-                exit 1
-                ;;
-        esac
+        check_not_running $CONTAINER_NAME $IMAGE
         create_bridge
         # We set the router name to the bridge mac since that is
         # stable across re-creations of the containers.
@@ -306,27 +353,22 @@ case "$COMMAND" in
         with_container_netns $CONTAINER launch >/dev/null
         echo $CONTAINER
         ;;
+    launch-dns)
+        [ $# -gt 0 ] || usage
+        CIDR=$1
+        shift 1
+        check_not_running $DNS_CONTAINER_NAME $DNS_IMAGE
+        create_bridge
+        CONTAINER=$(docker run -d --name=$DNS_CONTAINER_NAME -v /var/run/docker.sock:/var/run/docker.sock $DNS_IMAGE "$@")
+        with_container_netns $CONTAINER attach $CIDR >/dev/null
+        echo $CONTAINER
+        ;;
     connect)
         [ $# -eq 1 ] || usage
-        docker_ip
-        if command_exists curl ; then
-            curl -d peer=$1 $DOCKERIP:$HTTP_PORT/connect
-        else
-            echo "Cannot find curl. Install it for 'weave connect' to function." >&2
-            exit 1
-        fi
+        http_call POST $CONTAINER_NAME $HTTP_PORT /connect -d "peer=$1"
         ;;
     status)
-        [ $# -eq 0 ] || usage
-        docker_ip
-        if command_exists wget ; then
-            wget -O - -q $DOCKERIP:$HTTP_PORT/status
-        elif command_exists curl ; then
-            curl $DOCKERIP:$HTTP_PORT/status
-        else
-            echo "Cannot find wget or curl. Install one of them for 'weave status' to function." >&2
-            exit 1
-        fi
+        http_call GET $CONTAINER_NAME $HTTP_PORT /status
         ;;
     ps)
         [ $# -eq 0 ] || usage
@@ -355,8 +397,10 @@ case "$COMMAND" in
         CIDR=$1
         shift 1
         create_bridge
-        CONTAINER=$(docker run -d "$@" | tail -n 1)
+        DNS_ARG=$(docker inspect --format '--dns {{ .NetworkSettings.IPAddress }}' $DNS_CONTAINER_NAME 2>/dev/null) || true
+        CONTAINER=$(docker run $DNS_ARG -d "$@" | tail -n 1)
         with_container_netns $CONTAINER attach $CIDR >/dev/null
+        tell_dns PUT $CONTAINER $CIDR
         echo $CONTAINER
         ;;
     start)
@@ -365,6 +409,7 @@ case "$COMMAND" in
         create_bridge
         CONTAINER=$(docker start $2)
         with_container_netns $CONTAINER attach $1 >/dev/null
+        tell_dns PUT $CONTAINER $1
         echo $CONTAINER
         ;;
     attach)
@@ -372,11 +417,13 @@ case "$COMMAND" in
         validate_cidr $1
         create_bridge
         with_container_netns $2 attach $1 >/dev/null
+        tell_dns PUT $2 $1
         ;;
     detach)
         [ $# -eq 2 ] || usage
         validate_cidr $1
         with_container_netns $2 detach $1 >/dev/null
+        tell_dns DELETE $2 $1
         ;;
     expose)
         [ $# -eq 1 ] || usage
@@ -405,7 +452,9 @@ case "$COMMAND" in
         if ! docker kill $CONTAINER_NAME >/dev/null 2>&1 ; then
             echo "Weave is not running."
         fi
+
         docker rm -f $CONTAINER_NAME >/dev/null 2>&1 || true
+        docker rm -f $DNS_CONTAINER_NAME >/dev/null 2>&1 || true
         check_command_and_run conntrack conntrack -D -p udp --dport $PORT >/dev/null 2>&1 || true
         ;;
     reset)

weavedns/Dockerfile

@@ -0,0 +1,5 @@
+FROM scratch
+# EXPOSE 53/udp
+WORKDIR /home/weave
+ADD ./weavedns /home/weave/
+ENTRYPOINT ["/home/weave/weavedns", "-wait", "5", "-iface", "ethwe"]

weavedns/README.md

@@ -0,0 +1,97 @@
+# Weave DNS server
+
+The Weave DNS server answers name queries in a Weave network. It is
+run per-host, to be supplied as the nameserver for containers on that
+host. It is then told about hostnames for the local containers. For
+other names it will ask the other weave hosts, or fall back to using
+the host's configured name server.
+
+## Using weaveDNS
+
+The weave script command `launch-dns` starts the DNS container, and
+giving any other container a hostname in the domain `.weave.local`
+will register it in DNS. For example:
+
+```bash
+$ weave launch
+$ weave launch-dns 10.1.0.2/16
+$ weave run 10.1.1.25/24 -ti -h pingme.weave.local ubuntu /bin/bash
+$ shell1=$(weave run 10.1.1.26 -ti ubuntu.weave.local ubuntu /bin/bash)
+$ docker attach $shell1
+
+# ping pingme.weave.local
+...
+```
+
+## Doing things more manually
+
+If a weaveDNS container is running, `weave run` will automatically
+supply it as the DNS server to the new container. Similarly, both
+`weave run` and `weave attach` will register the hostname of the given
+container against the given weave network IP address.
+
+In some circumstances, you may not want to use the `weave`
+command. You can still take advantage of a running weaveDNS, using the
+HTTP API.
+
+### Supplying the DNS server
+
+If you want to use `docker run` to start a container, rather than
+`weave run`, you can supply the weaveDNS IP as the `--dns` option to
+make it use weaveDNS:
+
+```bash
+$ dns_ip=$(docker inspect --format='{{ .NetworkSettings.IPAddress }}' weavedns)
+$ shell2=$(docker run --dns=$dns_ip -ti ubuntu /bin/bash)
+$ weave attach 10.1.1.27/24 $shell2
+```
+
+This isn't very useful unless the container is also attached to the
+weave network (as in the last line above).
+
+### Adding containers to DNS
+
+If DNS is started after you've attached a container to the weave
+network, or you want to give the container a name in DNS *other* than
+its hostname, you can register it using the HTTP API:
+
+```bash
+$ docker start $shell2
+$ shell2_ip=$(docker inspect --format='{{ .NetworkSettings.IPAddress }}' $shell2)
+$ curl -X PUT "http://$dns_ip:6785/name/$shell2/10.1.1.27" -d local_ip=$shell2_ip -d fqdn=shell2.weave.local
+```
+
+### Not watching docker events
+
+By default, the server will watch docker events and remove entries for
+any containers that die. You can tell it not to, by adding
+`--watch=false` to the container args:
+
+```bash
+$ weave launch-dns 10.1.0.2/16 --watch=false
+```
+
+You can manually delete entries for a host, by poking weaveDNS's HTTP
+API with e.g., `curl`:
+
+```bash
+$ docker stop $shell2
+$ dns_ip=$(docker inspect --format='{{ .NetworkSettings.IPAddress }}' weavedns)
+$ curl -X DELETE "http://$dns_ip:6785/name/$shell2/10.1.1.27"
+```
+
+## Present limitations
+
+ * The server will not know about restarted containers, but if you
+   re-attach a restarted container to the weave network, it will be
+   re-registered with weaveDNS.
+ * The server will currently forget names if it is itself restarted,
+   and otherwise not know about containers running when it starts. In
+   the future it will look at existing container hostnames upon
+   starting up.
+ * The server may give unreachable IPs as answers, since it doesn't
+   try to filter by reachability. If you use subnets, align your
+   hostnames with the subnets.
+ * We use UDP multicast to find out about remote names (from Weave DNS
+   servers on other hosts); this likely won't scale well beyond a
+   certain point T.B.D., so we'll have to come up with another scheme.

weavedns/design.md

@@ -0,0 +1,110 @@
+# Weave DNS (service discovery) design
+
+The model is that each host has a service that is notified of
+hostnames and weave addresses for containers on the host. It binds to
+the host bridge to answer DNS queries from local containers; anything
+it can't answer, it asks with mDNS on the weave network.
+
+The service is divided into two components: the first is the DNS
+server, which serves the records it is told about, and otherwise asks
+via mDNS on the weave network. The other component is the DNS updater,
+specific to Docker, which monitors the running containers so to add
+and remove records from the DNS server.
+
+## DNS server API
+
+The DNS server accepts HTTP requests on the following URL (patterns)
+and methods:
+
+`PUT /name/<identifier>/<ip-address>`
+
+Put a record for an IP, bound to a host-scoped identifier (e.g., a
+container ID), in the DNS database.
+
+`DELETE /name/<identifier>/<ip-address>`
+
+Remove a specific record for an IP and host-scoped identifier.
+
+`DELETE /name/<identifier>`
+
+Remove all records for the host-scoped identifier.
+
+`GET /name/`
+
+List all the records in the database.
+
+`GET /name/<identifier>`
+
+List all the records for a host-scoped identifier.
+
+`GET /status`
+
+Give the server status.
+
+### Record structure
+
+The record structure is *either* form fields (useful for scripting
+PUTs), or a JSON object with the same fields. The fields are:
+
+```js
+    {
+        "fqdn": string,
+        "routing_prefix": number,
+        "local_ip": string
+    }
+```
+
+> Also TTL?
+
+## DNS server behaviour
+
+The DNS server listens for the HTTP requests as above, and maintains a
+database of host-local names accordingly.
+
+It also listens on the host-local network for DNS queries (port
+53). Anything that is in its database it responds to immediately. For
+other requests, it asks using multicast DNS on the weave network.
+
+Meanwhile, the DNS server is listening for mDNS queries, and responds
+if it the queried name is in its database.
+
+### Caching mDNS answers
+
+The DNS server will hear the answers to its own and other servers'
+mDNS queries sent on the weave network. It can cache these, to respond
+to local DNS requests; the cache is consulted before using mDNS.
+
+## DNS updater
+
+The updater component uses the Docker remote API to monitor containers
+coming and going, and tells the DNS server to update its records via
+its HTTP interface. It does not need to be attached to the weave
+network.
+
+The updater starts by subscribing to the events, and getting a list of
+the current containers. Any containers given a domain ending with
+".weave" are considered for inclusion in the name database.
+
+When it sees a container start or stop, the updater checks the weave
+network attachment of the container, and updates the DNS server.
+
+> How does it check the network attachment from within a container?
+
+> Will it need to delay slightly so that `attach` has a chance to run?
+> Perhaps it could put containers on a watch list when it's noticed
+> them.
+
+## Container startup
+
+> TBD
+
+## Grouping and load balancing
+
+> TBD
+
+## Extensions
+
+ * An updater that checks health more specifically, e.g., by testing
+   that a particular port is being listened to, or that it can get a
+   200 OK. These could be encoded in the container environment
+   variables, as per registrator.

weavedns/main.go

@@ -0,0 +1,63 @@
+package main
+
+import (
+	"flag"
+	weavedns "github.com/zettio/weave/nameserver"
+	weavenet "github.com/zettio/weave/net"
+	"io/ioutil"
+	"net"
+	"os"
+)
+
+func main() {
+	var (
+		ifaceName string
+		apiPath   string
+		dnsPort   int
+		httpPort  int
+		wait      int
+		watch     bool
+		debug     bool
+	)
+
+	flag.StringVar(&ifaceName, "iface", "", "name of interface to use for multicast")
+	flag.StringVar(&apiPath, "api", "unix:///var/run/docker.sock", "Path to Docker API socket")
+	flag.IntVar(&wait, "wait", 0, "number of seconds to wait for interface to be created and come up")
+	flag.IntVar(&dnsPort, "dnsport", 53, "port to listen to dns requests")
+	flag.IntVar(&httpPort, "httpport", 6785, "port to listen to HTTP requests")
+	flag.BoolVar(&watch, "watch", true, "watch the docker socket for container events")
+	flag.BoolVar(&debug, "debug", false, "output debugging info to stderr")
+	flag.Parse()
+
+	debugOut := ioutil.Discard
+	if debug {
+		debugOut = os.Stderr
+	}
+	weavedns.InitLogging(debugOut, os.Stdout, os.Stdout, os.Stderr)
+
+	var zone = new(weavedns.ZoneDb)
+
+	if watch {
+		err := weavedns.StartUpdater(apiPath, zone)
+		if err != nil {
+			weavedns.Error.Fatal("Unable to start watcher", err)
+		}
+	}
+
+	var iface *net.Interface = nil
+	if ifaceName != "" {
+		var err error
+		weavedns.Info.Println("Waiting for interface", ifaceName, "to come up")
+		iface, err = weavenet.EnsureInterface(ifaceName, wait)
+		if err != nil {
+			weavedns.Error.Fatal(err)
+		} else {
+			weavedns.Info.Println("Interface", ifaceName, "is up")
+		}
+	}
+
+	err := weavedns.StartServer(zone, iface, dnsPort, httpPort, wait)
+	if err != nil {
+		weavedns.Error.Fatal("Failed to start server", err)
+	}
+}

weaver/main.go

@@ -6,6 +6,7 @@ import (
 	"flag"
 	"fmt"
 	"github.com/davecheney/profile"
+	weavenet "github.com/zettio/weave/net"
 	weave "github.com/zettio/weave/router"
 	"io"
 	"log"
@@ -15,36 +16,8 @@ import (
 	"os/signal"
 	"runtime"
 	"syscall"
-	"time"
 )
 
-func ensureInterface(ifaceName string, wait int) (iface *net.Interface, err error) {
-	iface, err = findInterface(ifaceName)
-	if err == nil || wait == 0 {
-		return
-	}
-	log.Println("Waiting for interface", ifaceName, "to come up")
-	for ; err != nil && wait > 0; wait -= 1 {
-		time.Sleep(1 * time.Second)
-		iface, err = findInterface(ifaceName)
-	}
-	if err == nil {
-		log.Println("Interface", ifaceName, "is up")
-	}
-	return
-}
-
-func findInterface(ifaceName string) (iface *net.Interface, err error) {
-	iface, err = net.InterfaceByName(ifaceName)
-	if err != nil {
-		return iface, fmt.Errorf("Unable to find interface %s", ifaceName)
-	}
-	if 0 == (net.FlagUp & iface.Flags) {
-		return iface, fmt.Errorf("Interface %s is not up", ifaceName)
-	}
-	return
-}
-
 func main() {
 
 	log.SetPrefix(weave.Protocol + " ")
@@ -86,7 +59,7 @@ func main() {
 		fmt.Println("Missing required parameter 'iface'")
 		os.Exit(1)
 	}
-	iface, err := ensureInterface(ifaceName, wait)
+	iface, err := weavenet.EnsureInterface(ifaceName, wait)
 	if err != nil {
 		log.Fatal(err)
 	}
@@ -141,7 +114,7 @@ func handleHttp(router *weave.Router) {
 	http.HandleFunc("/status", func(w http.ResponseWriter, r *http.Request) {
 		io.WriteString(w, router.Status())
 	})
-	http.HandleFunc("/connect", func (w http.ResponseWriter, r *http.Request) {
+	http.HandleFunc("/connect", func(w http.ResponseWriter, r *http.Request) {
 		peer := r.FormValue("peer")
 		if addr, err := net.ResolveTCPAddr("tcp4", weave.NormalisePeerAddr(peer)); err == nil {
 			router.ConnectionMaker.InitiateConnection(addr.String())