update stopper

Gopkg.lock

@@ -152,7 +152,7 @@
     "stopOnce",
     "util"
   ]
-  revision = "a656ad8a1456310d6ca712098826d02950e46e0b"
+  revision = "2a6ea528bdd66de4f3c707304e26f69dcf003909"
 
 [[projects]]
   branch = "master"
@@ -172,6 +172,12 @@
   packages = ["."]
   revision = "e2103e2c35297fb7e17febb81e49b312087a2372"
 
+[[projects]]
+  branch = "master"
+  name = "github.com/sebdah/goldie"
+  packages = ["."]
+  revision = "8784dd1ab561dcf43d141f6678e9e41f3d0dff55"
+
 [[projects]]
   branch = "master"
   name = "github.com/sirupsen/logrus"
@@ -244,6 +250,6 @@
 [solve-meta]
   analyzer-name = "dep"
   analyzer-version = 1
-  inputs-digest = "b3d2f1578da4f4fca9a15d04df16cdb7f76f83c7916bab11fe1e50f65334a2c8"
+  inputs-digest = "bfa7ee41b88f515ef386012a9dd28e045f9108de93c1dcfa66d9361340cf7eec"
   solver-name = "gps-cdcl"
   solver-version = 1

cluster/cluster.go

@@ -3,13 +3,84 @@ package cluster
 import (
 	"io/ioutil"
 	baselog "log"
+	"os"
+	"os/signal"
+	"strconv"
+	"sync"
+	"syscall"
 
+	"github.com/davecgh/go-spew/spew"
+	"github.com/lbryio/lbry.go/crypto"
 	"github.com/lbryio/lbry.go/errors"
+	"github.com/lbryio/reflector.go/cluster"
 
 	"github.com/hashicorp/serf/serf"
 	log "github.com/sirupsen/logrus"
 )
 
+type Cluster struct {
+	s       *serf.Serf
+	eventCh <-chan serf.Event
+}
+
+func New() {
+	c := &Cluster{}
+	var err error
+
+	nodeName := crypto.RandString(12)
+	clusterAddr := "127.0.0.1:" + strconv.Itoa(clusterPort)
+	if args[0] == clusterStart {
+		c.s, c.eventCh, err = cluster.Connect(nodeName, clusterAddr, clusterPort)
+	} else {
+		c.s, c.eventCh, err = cluster.Connect(nodeName, clusterAddr, clusterPort+1+int(crypto.RandInt64(1000)))
+	}
+	if err != nil {
+		log.Fatal(err)
+	}
+	defer c.Leave()
+
+	shutdownCh := make(chan struct{})
+	var shutdownWg sync.WaitGroup
+
+	shutdownWg.Add(1)
+	go func() {
+		defer shutdownWg.Done()
+		for {
+			select {
+			case event := <-eventCh:
+				spew.Dump(event)
+				switch event.EventType() {
+				case serf.EventMemberJoin, serf.EventMemberFailed, serf.EventMemberLeave:
+					memberEvent := event.(serf.MemberEvent)
+					if event.EventType() == serf.EventMemberJoin && len(memberEvent.Members) == 1 && memberEvent.Members[0].Name == nodeName {
+						// ignore event from my own joining of the cluster
+					} else {
+						//spew.Dump(c.Members())
+						alive := getAliveMembers(c.Members())
+						log.Printf("%s: my hash range is now %d of %d\n", nodeName, getHashRangeStart(nodeName, alive), len(alive))
+						// figure out my new hash range based on the start and the number of alive members
+						// get hashes in that range that need announcing
+						// announce them
+						// if more than one node is announcing each hash, figure out how to deal with last_announced_at so both nodes dont announce the same thing at the same time
+					}
+				}
+			case <-shutdownCh:
+				log.Debugln("shutting down event dumper")
+				return
+			}
+		}
+	}()
+
+	interruptChan := make(chan os.Signal, 1)
+	signal.Notify(interruptChan, os.Interrupt, syscall.SIGTERM)
+	<-interruptChan
+	log.Debugln("received interrupt")
+	close(shutdownCh)
+	log.Debugln("waiting for threads to finish")
+	shutdownWg.Wait()
+	log.Debugln("shutting down main thread")
+}
+
 func Connect(nodeName, addr string, port int) (*serf.Serf, <-chan serf.Event, error) {
 	conf := serf.DefaultConfig()
 	conf.MemberlistConfig.BindPort = port

cmd/upload.go

@@ -98,7 +98,7 @@ func uploadCmd(cmd *cobra.Command, args []string) {
 
 			for {
 				select {
-				case <-stopper.Chan():
+				case <-stopper.Ch():
 					return
 				case filename, ok := <-filenameChan:
 					if !ok {
@@ -113,7 +113,7 @@ func uploadCmd(cmd *cobra.Command, args []string) {
 						log.Errorf("worker %d: filename does not match hash (%s != %s), skipping", i, filename, hash)
 						select {
 						case countChan <- errInc:
-						case <-stopper.Chan():
+						case <-stopper.Ch():
 						}
 						continue
 					}
@@ -123,14 +123,14 @@ func uploadCmd(cmd *cobra.Command, args []string) {
 						blobStore.PutSD(hash, blob)
 						select {
 						case countChan <- sdInc:
-						case <-stopper.Chan():
+						case <-stopper.Ch():
 						}
 					} else {
 						log.Printf("worker %d: putting %s", i, hash)
 						blobStore.Put(hash, blob)
 						select {
 						case countChan <- blobInc:
-						case <-stopper.Chan():
+						case <-stopper.Ch():
 						}
 					}
 				}
@@ -143,7 +143,7 @@ func uploadCmd(cmd *cobra.Command, args []string) {
 		defer counterWG.Done()
 		for {
 			select {
-			case <-stopper.Chan():
+			case <-stopper.Ch():
 				return
 			case countType, ok := <-countChan:
 				if !ok {
@@ -172,7 +172,7 @@ Upload:
 
 		select {
 		case filenameChan <- filename:
-		case <-stopper.Chan():
+		case <-stopper.Ch():
 			log.Warnln("Caught interrupt, quitting at first opportunity...")
 			break Upload
 		}

dht/bitmap.go

@@ -10,6 +10,8 @@ import (
 	"github.com/lyoshenka/bencode"
 )
 
+// TODO: http://roaringbitmap.org/
+
 type Bitmap [nodeIDLength]byte
 
 func (b Bitmap) RawString() string {

dht/bootstrap.go

@@ -62,7 +62,7 @@ func (b *BootstrapNode) Connect(conn UDPConn) error {
 			select {
 			case <-t.C:
 				b.check()
-			case <-b.stop.Chan():
+			case <-b.stop.Ch():
 				return
 			}
 		}
@@ -130,7 +130,7 @@ func (b *BootstrapNode) ping(c Contact) {
 
 	select {
 	case res = <-resCh:
-	case <-b.stop.Chan():
+	case <-b.stop.Ch():
 		cancel()
 		return
 	}

dht/dht.go

@@ -81,8 +81,6 @@ type DHT struct {
 	node *Node
 	// stopper to shut down DHT
 	stop *stopOnce.Stopper
-	// wait group for all the things that need to be stopped when DHT shuts down
-	stopWG *sync.WaitGroup
 	// channel is closed when DHT joins network
 	joined chan struct{}
 	// lock for announced list
@@ -107,7 +105,6 @@ func New(config *Config) (*DHT, error) {
 		contact:   contact,
 		node:      NewNode(contact.ID),
 		stop:      stopOnce.New(),
-		stopWG:    &sync.WaitGroup{},
 		joined:    make(chan struct{}),
 		lock:      &sync.RWMutex{},
 		announced: make(map[Bitmap]bool),
@@ -143,6 +140,9 @@ func (dht *DHT) join() {
 	if err != nil {
 		log.Errorf("[%s] join: %s", dht.node.id.HexShort(), err.Error())
 	}
+
+	// TODO: after joining, refresh all the buckets all buckets further away than our closest neighbor
+	// http://xlattice.sourceforge.net/components/protocol/kademlia/specs.html#join
 }
 
 // Start starts the dht
@@ -176,7 +176,7 @@ func (dht *DHT) WaitUntilJoined() {
 func (dht *DHT) Shutdown() {
 	log.Debugf("[%s] DHT shutting down", dht.node.id.HexShort())
 	dht.stop.Stop()
-	dht.stopWG.Wait()
+	dht.stop.Wait()
 	dht.node.Shutdown()
 	log.Debugf("[%s] DHT stopped", dht.node.id.HexShort())
 }
@@ -244,12 +244,12 @@ func (dht *DHT) Announce(hash Bitmap) error {
 }
 
 func (dht *DHT) startReannouncer() {
-	dht.stopWG.Add(1)
+	dht.stop.Add(1)
-	defer dht.stopWG.Done()
+	defer dht.stop.Done()
 	tick := time.NewTicker(tReannounce)
 	for {
 		select {
-		case <-dht.stop.Chan():
+		case <-dht.stop.Ch():
 			return
 		case <-tick.C:
 			dht.lock.RLock()
@@ -262,8 +262,8 @@ func (dht *DHT) startReannouncer() {
 }
 
 func (dht *DHT) storeOnNode(hash Bitmap, c Contact) {
-	dht.stopWG.Add(1)
+	dht.stop.Add(1)
-	defer dht.stopWG.Done()
+	defer dht.stop.Done()
 
 	// self-store
 	if dht.contact.Equals(c) {
@@ -280,7 +280,7 @@ func (dht *DHT) storeOnNode(hash Bitmap, c Contact) {
 
 	select {
 	case res = <-resCh:
-	case <-dht.stop.Chan():
+	case <-dht.stop.Ch():
 		cancel()
 		return
 	}
@@ -304,7 +304,7 @@ func (dht *DHT) storeOnNode(hash Bitmap, c Contact) {
 	go func() {
 		select {
 		case <-resCh:
-		case <-dht.stop.Chan():
+		case <-dht.stop.Ch():
 			cancel()
 		}
 	}()

dht/message.go

@@ -154,7 +154,7 @@ type storeArgsValue struct {
 type storeArgs struct {
 	BlobHash  Bitmap
 	Value     storeArgsValue
-	NodeID    Bitmap
+	NodeID    Bitmap // original publisher id? I think this is getting fixed in the new dht stuff
 	SelfStore bool   // this is an int on the wire
 }
 

dht/node.go

@@ -87,7 +87,7 @@ func (n *Node) Connect(conn UDPConn) error {
 	//			dht.PrintState()
 	//			select {
 	//			case <-t.C:
-	//			case <-dht.stop.Chan():
+	//			case <-dht.stop.Ch():
 	//				return
 	//			}
 	//		}
@@ -106,7 +106,7 @@ func (n *Node) Connect(conn UDPConn) error {
 
 		for {
 			select {
-			case <-n.stop.Chan():
+			case <-n.stop.Ch():
 				return
 			default:
 			}
@@ -128,7 +128,7 @@ func (n *Node) Connect(conn UDPConn) error {
 
 			select { // needs select here because packet consumer can quit and the packets channel gets filled up and blocks
 			case packets <- packet{data: data, raddr: raddr}:
-			case <-n.stop.Chan():
+			case <-n.stop.Ch():
 			}
 		}
 	}()
@@ -143,7 +143,7 @@ func (n *Node) Connect(conn UDPConn) error {
 			select {
 			case pkt = <-packets:
 				n.handlePacket(pkt)
-			case <-n.stop.Chan():
+			case <-n.stop.Ch():
 				return
 			}
 		}
@@ -434,8 +434,8 @@ func (n *Node) startRoutingTableGrooming() {
 		for {
 			select {
 			case <-refreshTicker.C:
-				RoutingTableRefresh(n, tRefresh, n.stop.Chan())
+				RoutingTableRefresh(n, tRefresh, n.stop.Ch())
-			case <-n.stop.Chan():
+			case <-n.stop.Ch():
 				return
 			}
 		}

dht/node_finder.go

@@ -19,8 +19,7 @@ type contactFinder struct {
 	target    Bitmap
 	node      *Node
 
-	done   *stopOnce.Stopper
+	stop *stopOnce.Stopper
-	doneWG *sync.WaitGroup
 
 	findValueMutex  *sync.Mutex
 	findValueResult []Contact
@@ -50,15 +49,14 @@ func newContactFinder(node *Node, target Bitmap, findValue bool) *contactFinder
 		activeContactsMutex: &sync.Mutex{},
 		shortlistMutex:      &sync.Mutex{},
 		shortlistAdded:      make(map[Bitmap]bool),
-		done:                stopOnce.New(),
+		stop:                stopOnce.New(),
-		doneWG:              &sync.WaitGroup{},
 		outstandingRequestsMutex: &sync.RWMutex{},
 	}
 }
 
 func (cf *contactFinder) Cancel() {
-	cf.done.Stop()
+	cf.stop.Stop()
-	cf.doneWG.Wait()
+	cf.stop.Wait()
 }
 
 func (cf *contactFinder) Find() (findNodeResponse, error) {
@@ -73,14 +71,14 @@ func (cf *contactFinder) Find() (findNodeResponse, error) {
 	}
 
 	for i := 0; i < alpha; i++ {
-		cf.doneWG.Add(1)
+		cf.stop.Add(1)
 		go func(i int) {
-			defer cf.doneWG.Done()
+			defer cf.stop.Done()
 			cf.iterationWorker(i + 1)
 		}(i)
 	}
 
-	cf.doneWG.Wait()
+	cf.stop.Wait()
 
 	// TODO: what to do if we have less than K active contacts, shortlist is empty, but we
 	// TODO: have other contacts in our routing table whom we have not contacted. prolly contact them
@@ -131,7 +129,7 @@ func (cf *contactFinder) iterationWorker(num int) {
 			resCh, cancel := cf.node.SendCancelable(contact, req)
 			select {
 			case res = <-resCh:
-			case <-cf.done.Chan():
+			case <-cf.stop.Ch():
 				log.Debugf("[%s] worker %d: canceled", cf.node.id.HexShort(), num)
 				cancel()
 				return
@@ -145,7 +143,7 @@ func (cf *contactFinder) iterationWorker(num int) {
 				cf.findValueMutex.Lock()
 				cf.findValueResult = res.Contacts
 				cf.findValueMutex.Unlock()
-				cf.done.Stop()
+				cf.stop.Stop()
 				return
 			} else {
 				log.Debugf("[%s] worker %d: got contacts", cf.node.id.HexShort(), num)
@@ -158,7 +156,7 @@ func (cf *contactFinder) iterationWorker(num int) {
 
 		if cf.isSearchFinished() {
 			log.Debugf("[%s] worker %d: search is finished", cf.node.id.HexShort(), num)
-			cf.done.Stop()
+			cf.stop.Stop()
 			return
 		}
 	}
@@ -214,7 +212,7 @@ func (cf *contactFinder) isSearchFinished() bool {
 	}
 
 	select {
-	case <-cf.done.Chan():
+	case <-cf.stop.Ch():
 		return true
 	default:
 	}

dht/token_manager.go

@@ -37,7 +37,7 @@ func (tm *tokenManager) Start(interval time.Duration) {
 			select {
 			case <-tick.C:
 				tm.rotateSecret()
-			case <-tm.done.Chan():
+			case <-tm.done.Ch():
 				return
 			}
 		}