Fix flaky TestSingleTopicDiscoveryWithFailover and others (#992)

Other changes:
* needed to patch that loop implementation in Discover V5 implementation in go-ethereum,
* fixed TestStatusNodeReconnectStaticPeers,
* fixed TestBackendAccountsConcurrently.

_assets/patches/geth/0031-fix-discovery-loop-when-no-seeds.patch

@@ -0,0 +1,24 @@
+diff --git a/p2p/discv5/net.go b/p2p/discv5/net.go
+index d0eae28f..25e75d50 100644
+--- a/p2p/discv5/net.go
++++ b/p2p/discv5/net.go
+@@ -641,7 +641,18 @@ loop:
+ 				}()
+ 			} else {
+ 				refreshDone = make(chan struct{})
+-				net.refresh(refreshDone)
++
++				done := make(chan struct{})
++				net.refresh(done)
++				<-done
++
++				// Refresh again only if there are no seeds.
++				// Also, sleep for some time to prevent from
++				// executing too often.
++				go func() {
++					time.Sleep(time.Millisecond * 100)
++					close(refreshDone)
++				}()
+ 			}
+ 		}
+ 	}

geth/api/backend.go

@@ -354,6 +354,9 @@ func (b *StatusBackend) AppStateChange(state string) {
 
 // Logout clears whisper identities.
 func (b *StatusBackend) Logout() error {
+	b.mu.Lock()
+	defer b.mu.Unlock()
+
 	// FIXME(oleg-raev): This method doesn't make stop, it rather resets its cells to an initial state
 	// and should be properly renamed, for example: ResetCells
 	b.jailManager.Stop()
@@ -398,6 +401,9 @@ func (b *StatusBackend) ReSelectAccount() error {
 // using provided password. Once verification is done, decrypted key is injected into Whisper (as a single identity,
 // all previous identities are removed).
 func (b *StatusBackend) SelectAccount(address, password string) error {
+	b.mu.Lock()
+	defer b.mu.Unlock()
+
 	// FIXME(oleg-raev): This method doesn't make stop, it rather resets its cells to an initial state
 	// and should be properly renamed, for example: ResetCells
 	b.jailManager.Stop()

geth/node/status_node_test.go

@@ -238,7 +238,6 @@ func TestStatusNodeReconnectStaticPeers(t *testing.T) {
 	require.NoError(t, n.ReconnectStaticPeers())
 	// first check if a peer gets disconnected
 	require.NoError(t, <-errCh)
-	require.Equal(t, 0, n.PeerCount())
 	// it takes at least 30 seconds to bring back previously connected peer
 	errCh = waitForPeerAsync(n, peerURL, p2p.PeerEventTypeAdd, time.Second*60)
 	require.NoError(t, <-errCh)

geth/peers/peerpool.go

@@ -113,6 +113,15 @@ func (p *PeerPool) Start(server *p2p.Server) error {
 	p.quit = make(chan struct{})
 	p.setDiscoveryTimeout()
 
+	// subscribe to peer events
+	p.events = make(chan *p2p.PeerEvent, 20)
+	p.serverSubscription = server.SubscribeEvents(p.events)
+	p.wg.Add(1)
+	go func() {
+		p.handleServerPeers(server, p.events)
+		p.wg.Done()
+	}()
+
 	// collect topics and start searching for nodes
 	p.topics = make([]*TopicPool, 0, len(p.config))
 	for topic, limits := range p.config {
@@ -126,15 +135,6 @@ func (p *PeerPool) Start(server *p2p.Server) error {
 	// discovery must be already started when pool is started
 	signal.SendDiscoveryStarted()
 
-	// subscribe to peer events
-	p.events = make(chan *p2p.PeerEvent, 20)
-	p.serverSubscription = server.SubscribeEvents(p.events)
-	p.wg.Add(1)
-	go func() {
-		p.handleServerPeers(server, p.events)
-		p.wg.Done()
-	}()
-
 	return nil
 }
 

geth/peers/peerpool_test.go

@@ -19,6 +19,9 @@ import (
 
 	"github.com/status-im/status-go/geth/params"
 	"github.com/status-im/status-go/signal"
+
+	// to access logs in the test with `-log` flag
+	_ "github.com/status-im/status-go/t/utils"
 )
 
 type PeerPoolSimulationSuite struct {
@@ -26,14 +29,22 @@ type PeerPoolSimulationSuite struct {
 
 	bootnode *p2p.Server
 	peers    []*p2p.Server
+	port     uint16
 }
 
 func TestPeerPoolSimulationSuite(t *testing.T) {
-	suite.Run(t, new(PeerPoolSimulationSuite))
+	s := new(PeerPoolSimulationSuite)
+	s.port = 33731
+	suite.Run(t, s)
+}
+
+func (s *PeerPoolSimulationSuite) nextPort() uint16 {
+	s.port++
+	return s.port
 }
 
 func (s *PeerPoolSimulationSuite) SetupTest() {
-	port := 33731
+	bootnodePort := s.nextPort()
 	key, _ := crypto.GenerateKey()
 	name := common.MakeName("bootnode", "1.0")
 	// 127.0.0.1 is invalidated by discovery v5
@@ -41,15 +52,14 @@ func (s *PeerPoolSimulationSuite) SetupTest() {
 		Config: p2p.Config{
 			MaxPeers:    10,
 			Name:        name,
-			ListenAddr:  fmt.Sprintf("0.0.0.0:%d", 33731),
+			ListenAddr:  fmt.Sprintf("0.0.0.0:%d", bootnodePort),
 			PrivateKey:  key,
 			DiscoveryV5: true,
 			NoDiscovery: true,
 		},
 	}
-	port++
 	s.Require().NoError(s.bootnode.Start())
-	bootnodeV5 := discv5.NewNode(s.bootnode.DiscV5.Self().ID, net.ParseIP("127.0.0.1"), uint16(port), uint16(port))
+	bootnodeV5 := discv5.NewNode(s.bootnode.DiscV5.Self().ID, net.ParseIP("127.0.0.1"), bootnodePort, bootnodePort)
 
 	// 1 peer to initiate connection, 1 peer as a first candidate, 1 peer - for failover
 	s.peers = make([]*p2p.Server, 3)
@@ -60,7 +70,7 @@ func (s *PeerPoolSimulationSuite) SetupTest() {
 			Config: p2p.Config{
 				MaxPeers:         10,
 				Name:             common.MakeName("peer-"+strconv.Itoa(i), "1.0"),
-				ListenAddr:       fmt.Sprintf("0.0.0.0:%d", port),
+				ListenAddr:       fmt.Sprintf("0.0.0.0:%d", s.nextPort()),
 				PrivateKey:       key,
 				DiscoveryV5:      true,
 				NoDiscovery:      true,
@@ -68,7 +78,6 @@ func (s *PeerPoolSimulationSuite) SetupTest() {
 				Protocols:        whisper.Protocols(),
 			},
 		}
-		port++
 		s.NoError(peer.Start())
 		s.peers[i] = peer
 	}
@@ -87,7 +96,8 @@ func (s *PeerPoolSimulationSuite) getPeerFromEvent(events <-chan *p2p.PeerEvent,
 		if ev.Type == etype {
 			return ev.Peer
 		}
-	case <-time.After(5 * time.Second):
+		s.Failf("invalid event", "expected %s but got %s for peer %s", etype, ev.Type, ev.Peer)
+	case <-time.After(10 * time.Second):
 		s.Fail("timed out waiting for a peer")
 		return
 	}
@@ -98,25 +108,51 @@ func (s *PeerPoolSimulationSuite) getPoolEvent(events <-chan string) string {
 	select {
 	case ev := <-events:
 		return ev
-	case <-time.After(time.Second):
-		s.FailNow("timed out waiting for a peer")
+	case <-time.After(10 * time.Second):
+		s.FailNow("timed out waiting a pool event")
 		return ""
 	}
 }
 
+func (s *PeerPoolSimulationSuite) TestPeerPoolCache() {
+	var err error
+
+	topic := discv5.Topic("cap=test")
+	config := map[discv5.Topic]params.Limits{
+		topic: params.NewLimits(1, 1),
+	}
+	peerPoolOpts := &Options{100 * time.Millisecond, 100 * time.Millisecond, 0, true}
+	cache, err := newInMemoryCache()
+	s.Require().NoError(err)
+	peerPool := NewPeerPool(config, cache, peerPoolOpts)
+
+	// start peer pool
+	s.Require().NoError(peerPool.Start(s.peers[1]))
+	defer peerPool.Stop()
+
+	// check if cache is passed to topic pools
+	for _, topicPool := range peerPool.topics {
+		s.Equal(cache, topicPool.cache)
+	}
+}
+
 func (s *PeerPoolSimulationSuite) TestSingleTopicDiscoveryWithFailover() {
-	poolEvents := make(chan string, 1)
-	summaries := make(chan []*p2p.PeerInfo, 1)
+	var err error
+
+	// Buffered channels must be used because we expect the events
+	// to be in the same order. Use a buffer length greater than
+	// the expected number of events to avoid deadlock.
+	poolEvents := make(chan string, 10)
+	summaries := make(chan []*p2p.PeerInfo, 10)
 	signal.SetDefaultNodeNotificationHandler(func(jsonEvent string) {
 		var envelope struct {
 			Type  string
 			Event json.RawMessage
 		}
 		s.NoError(json.Unmarshal([]byte(jsonEvent), &envelope))
-		switch envelope.Type {
-		case signal.EventDiscoveryStarted:
-			poolEvents <- envelope.Type
-		case signal.EventDiscoveryStopped:
+
+		switch typ := envelope.Type; typ {
+		case signal.EventDiscoveryStarted, signal.EventDiscoveryStopped:
 			poolEvents <- envelope.Type
 		case signal.EventDiscoverySummary:
 			poolEvents <- envelope.Type
@@ -124,9 +160,9 @@ func (s *PeerPoolSimulationSuite) TestSingleTopicDiscoveryWithFailover() {
 			s.NoError(json.Unmarshal(envelope.Event, &summary))
 			summaries <- summary
 		}
-
 	})
 	defer signal.ResetDefaultNodeNotificationHandler()
+
 	topic := discv5.Topic("cap=test")
 	// simulation should only rely on fast sync
 	config := map[discv5.Topic]params.Limits{
@@ -136,50 +172,48 @@ func (s *PeerPoolSimulationSuite) TestSingleTopicDiscoveryWithFailover() {
 	cache, err := newInMemoryCache()
 	s.Require().NoError(err)
 	peerPool := NewPeerPool(config, cache, peerPoolOpts)
+
+	// create and start topic registry
 	register := NewRegister(topic)
-	s.Require().NoError(register.Start(s.peers[0]))
-	// need to wait for topic to get registered, discv5 can query same node
-	// for a topic only once a minute
+	err = register.Start(s.peers[0])
+	s.Require().NoError(err)
+
+	// subscribe for peer events before starting the peer pool
 	events := make(chan *p2p.PeerEvent, 20)
 	subscription := s.peers[1].SubscribeEvents(events)
 	defer subscription.Unsubscribe()
-	s.NoError(peerPool.Start(s.peers[1]))
+
+	// start the peer pool
+	s.Require().NoError(peerPool.Start(s.peers[1]))
 	defer peerPool.Stop()
 	s.Equal(signal.EventDiscoveryStarted, s.getPoolEvent(poolEvents))
-	connected := s.getPeerFromEvent(events, p2p.PeerEventTypeAdd)
-	s.Equal(s.peers[0].Self().ID, connected)
+	connectedPeer := s.getPeerFromEvent(events, p2p.PeerEventTypeAdd)
+	s.Equal(s.peers[0].Self().ID, connectedPeer)
+	// as the upper limit was reached, Discovery should be stoped
 	s.Equal(signal.EventDiscoveryStopped, s.getPoolEvent(poolEvents))
-	s.Require().Nil(s.peers[1].DiscV5)
-
-	s.Require().Equal(signal.EventDiscoverySummary, s.getPoolEvent(poolEvents))
-	summary := <-summaries
-	s.Len(summary, 1)
+	s.Equal(signal.EventDiscoverySummary, s.getPoolEvent(poolEvents))
+	s.Len(<-summaries, 1)
 
+	// stop topic register and the connected peer
 	register.Stop()
 	s.peers[0].Stop()
-	disconnected := s.getPeerFromEvent(events, p2p.PeerEventTypeDrop)
-	s.Equal(connected, disconnected)
-
-	s.Require().Equal(signal.EventDiscoverySummary, s.getPoolEvent(poolEvents))
-	summary = <-summaries
-	s.Len(summary, 0)
-
+	disconnectedPeer := s.getPeerFromEvent(events, p2p.PeerEventTypeDrop)
+	s.Equal(connectedPeer, disconnectedPeer)
+	s.Equal(signal.EventDiscoverySummary, s.getPoolEvent(poolEvents))
+	s.Len(<-summaries, 0)
+	// Discovery should be restarted because the number of peers dropped
+	// below the lower limit.
 	s.Equal(signal.EventDiscoveryStarted, s.getPoolEvent(poolEvents))
-	s.Require().NotNil(s.peers[1].DiscV5)
-	s.Require().NoError(register.Start(s.peers[2]))
+
+	// register the second peer
+	err = register.Start(s.peers[2])
+	s.Require().NoError(err)
 	defer register.Stop()
 	s.Equal(s.peers[2].Self().ID, s.getPeerFromEvent(events, p2p.PeerEventTypeAdd))
-
+	// Discovery can be stopped again.
 	s.Equal(signal.EventDiscoveryStopped, s.getPoolEvent(poolEvents))
 	s.Require().Equal(signal.EventDiscoverySummary, s.getPoolEvent(poolEvents))
-
-	summary = <-summaries
-	s.Len(summary, 1)
-
-	// verify that we are actually using cache
-	cachedPeers := peerPool.cache.GetPeersRange(topic, 1)
-	s.Len(cachedPeers, 1)
-	s.Equal(s.peers[2].Self().ID, discover.NodeID(cachedPeers[0].ID))
+	s.Len(<-summaries, 1)
 }
 
 // TestPeerPoolMaxPeersOverflow verifies that following scenario will not occur:
@@ -189,7 +223,7 @@ func (s *PeerPoolSimulationSuite) TestSingleTopicDiscoveryWithFailover() {
 // - process peer B
 // - panic because discv5 is nil!!!
 func TestPeerPoolMaxPeersOverflow(t *testing.T) {
-	signals := make(chan string, 1)
+	signals := make(chan string, 10)
 	signal.SetDefaultNodeNotificationHandler(func(jsonEvent string) {
 		var envelope struct {
 			Type string

geth/peers/topicpool_test.go

@@ -63,6 +63,29 @@ func (s *TopicPoolSuite) AssertConsumed(channel <-chan time.Duration, expected t
 	}
 }
 
+func (s *TopicPoolSuite) TestUsingCache() {
+	s.topicPool.limits = params.NewLimits(1, 1)
+
+	peer1 := discv5.NewNode(discv5.NodeID{1}, s.peer.Self().IP, 32311, 32311)
+	s.topicPool.processFoundNode(s.peer, peer1)
+	s.topicPool.ConfirmAdded(s.peer, discover.NodeID(peer1.ID))
+	s.Equal([]*discv5.Node{peer1}, s.topicPool.cache.GetPeersRange(s.topicPool.topic, 10))
+
+	// Add a new peer which exceeds the upper limit.
+	// It should still be added to the cache and
+	// not removed when dropped.
+	peer2 := discv5.NewNode(discv5.NodeID{2}, s.peer.Self().IP, 32311, 32311)
+	s.topicPool.processFoundNode(s.peer, peer2)
+	s.topicPool.ConfirmAdded(s.peer, discover.NodeID(peer2.ID))
+	s.Equal([]*discv5.Node{peer1, peer2}, s.topicPool.cache.GetPeersRange(s.topicPool.topic, 10))
+	s.topicPool.ConfirmDropped(s.peer, discover.NodeID(peer2.ID))
+	s.Equal([]*discv5.Node{peer1, peer2}, s.topicPool.cache.GetPeersRange(s.topicPool.topic, 10))
+
+	// A peer that drops by itself, should be removed from the cache.
+	s.topicPool.ConfirmDropped(s.peer, discover.NodeID(peer1.ID))
+	s.Equal([]*discv5.Node{peer2}, s.topicPool.cache.GetPeersRange(s.topicPool.topic, 10))
+}
+
 func (s *TopicPoolSuite) TestSyncSwitches() {
 	testPeer := discv5.NewNode(discv5.NodeID{1}, s.peer.Self().IP, 32311, 32311)
 	s.topicPool.processFoundNode(s.peer, testPeer)

vendor/github.com/ethereum/go-ethereum/p2p/discv5/net.go

@@ -641,7 +641,18 @@ loop:
 				}()
 			} else {
 				refreshDone = make(chan struct{})
-				net.refresh(refreshDone)
+
+				done := make(chan struct{})
+				net.refresh(done)
+				<-done
+
+				// Refresh again only if there are no seeds.
+				// Also, sleep for some time to prevent from
+				// executing too often.
+				go func() {
+					time.Sleep(time.Millisecond * 100)
+					close(refreshDone)
+				}()
 			}
 		}
 	}