fix!: retrieve transactions from the mempool async, ignore irelevant rounds, and don't start processing messages until switching from blocksync to consensus (#1695)

## Description

this PR started as a simple fix, but then its scope grew significantly
as we found more bugs and made the testnet work

it:
- makes retrieving the transactions from the mempool async during
propagation.
- ajusts priorities for gossip
- stops automatically broadcasting transactions with CAT, instead only
sends seentx (this stops peers from downloading transactions more than
once)
- minor optimizations around not verifying proofs that the node just
generated.
- NOTE: this PR reverts
#1553 and
#1582 cause those break
CAT and we need a working mempool that doesn't re-gossip everything
- ignores irelevant round
- only processes compact blocks until the consensus reactor starts
- a few other minor bug fixes

Author: evan-forbes

blockchain/v0/pool.go

@@ -32,7 +32,7 @@ const (
 	maxTotalRequesters        = 600
 	maxPendingRequests        = maxTotalRequesters
 	maxPendingRequestsPerPeer = 20
-	requestRetrySeconds       = 30
+	requestRetrySeconds       = 45
 
 	// Minimum recv rate to ensure we're receiving blocks from a peer fast
 	// enough. If a peer is not sending us data at least that rate, we
@@ -46,7 +46,7 @@ const (
 	maxDiffBetweenCurrentAndReceivedBlockHeight = 100
 )
 
-var peerTimeout = 15 * time.Second // not const so we can override with tests
+var peerTimeout = 120 * time.Second // not const so we can override with tests
 
 /*
 	Peers self report their heights when we join the block pool.

blockchain/v0/reactor.go

@@ -145,7 +145,7 @@ func (bcR *BlockchainReactor) GetChannels() []*p2p.ChannelDescriptor {
 	return []*p2p.ChannelDescriptor{
 		{
 			ID:                  BlockchainChannel,
-			Priority:            5,
+			Priority:            30,
 			SendQueueCapacity:   1000,
 			RecvBufferCapacity:  50 * 4096,
 			RecvMessageCapacity: bc.MaxMsgSize,

consensus/propagation/catchup.go

@@ -6,6 +6,7 @@ import (
 	proptypes "github.com/tendermint/tendermint/consensus/propagation/types"
 	"github.com/tendermint/tendermint/libs/bits"
 	"github.com/tendermint/tendermint/p2p"
+	"github.com/tendermint/tendermint/pkg/trace/schema"
 	protoprop "github.com/tendermint/tendermint/proto/tendermint/propagation"
 	"github.com/tendermint/tendermint/types"
 )
@@ -14,13 +15,17 @@ import (
 //
 // todo: add a request limit for each part to avoid downloading the block too
 // many times. atm, this code will request the same part from every peer.
-func (blockProp *Reactor) retryWants(currentHeight int64, currentRound int32) {
-	data := blockProp.dumpAll()
+func (blockProp *Reactor) retryWants(currentHeight int64) {
+	if !blockProp.started.Load() {
+		return
+	}
+	data := blockProp.unfinishedHeights()
 	peers := blockProp.getPeers()
 	for _, prop := range data {
 		height, round := prop.compactBlock.Proposal.Height, prop.compactBlock.Proposal.Round
 
-		if height == currentHeight && round == currentRound {
+		// don't re-request parts for any round on the current height
+		if height == currentHeight {
 			continue
 		}
 
@@ -29,17 +34,20 @@ func (blockProp *Reactor) retryWants(currentHeight int64, currentRound int32) {
 		}
 
 		// only re-request original parts that are missing, not parity parts.
-		missing := prop.block.BitArray().Not()
+		missing := prop.block.MissingOriginal()
 		if missing.IsEmpty() {
 			blockProp.Logger.Error("no missing parts yet block is incomplete", "height", height, "round", round)
 			continue
 		}
 
+		schema.WriteRetries(blockProp.traceClient, height, round, missing.String())
+
 		// make requests from different peers
 		peers = shuffle(peers)
 
 		for _, peer := range peers {
 			mc := missing.Copy()
+
 			reqs, has := peer.GetRequests(height, round)
 			if has {
 				mc = mc.Sub(reqs)
@@ -52,7 +60,7 @@ func (blockProp *Reactor) retryWants(currentHeight int64, currentRound int32) {
 			e := p2p.Envelope{
 				ChannelID: WantChannel,
 				Message: &protoprop.WantParts{
-					Parts:  *missing.ToProto(),
+					Parts:  *mc.ToProto(),
 					Height: height,
 					Round:  round,
 					Prove:  true,
@@ -64,17 +72,23 @@ func (blockProp *Reactor) retryWants(currentHeight int64, currentRound int32) {
 				continue
 			}
 
+			schema.WriteCatchupRequest(blockProp.traceClient, height, round, mc.String(), string(peer.peer.ID()))
+
 			// keep track of which requests we've made this attempt.
-			missing.Sub(mc)
+			missing = missing.Sub(mc)
 			peer.AddRequests(height, round, missing)
 		}
 	}
 }
 
 func (blockProp *Reactor) AddCommitment(height int64, round int32, psh *types.PartSetHeader) {
+	blockProp.Logger.Info("adding commitment", "height", height, "round", round, "psh", psh)
 	blockProp.pmtx.Lock()
 	defer blockProp.pmtx.Unlock()
 
+	blockProp.Logger.Info("added commitment", "height", height, "round", round)
+	schema.WriteGap(blockProp.traceClient, height, round)
+
 	if blockProp.proposals[height] == nil {
 		blockProp.proposals[height] = make(map[int32]*proposalData)
 	}

consensus/propagation/catchup_test.go

@@ -1,75 +1,76 @@
 package propagation
 
-// TODO(rachid): fix test
-// func TestCatchup(t *testing.T) {
-// 	reactors, _ := testBlockPropReactors(3)
-// 	reactor1 := reactors[0]
-// 	reactor2 := reactors[1]
-// 	reactor3 := reactors[2]
-
-// 	// setting the proposal for height 8 round 1
-// 	compactBlock := createCompactBlock(8, 1)
-// 	reactor1.AddProposal(compactBlock)
-// 	reactor2.AddProposal(compactBlock)
-// 	reactor3.AddProposal(compactBlock)
-
-// 	// setting the proposal for height 9 round 0
-// 	compactBlock = createCompactBlock(9, 1)
-// 	reactor1.AddProposal(compactBlock)
-// 	reactor2.AddProposal(compactBlock)
-// 	reactor3.AddProposal(compactBlock)
-
-// 	// setting the proposal for height 10 round 0
-// 	compactBlock = createCompactBlock(10, 0)
-// 	reactor1.AddProposal(compactBlock)
-// 	reactor2.AddProposal(compactBlock)
-// 	reactor3.AddProposal(compactBlock)
-
-// 	// setting the proposal for height 10 round 1
-// 	compactBlock = createCompactBlock(10, 1)
-// 	reactor1.AddProposal(compactBlock)
-// 	reactor2.AddProposal(compactBlock)
-// 	reactor3.AddProposal(compactBlock)
-
-// 	// setting the first reactor current height and round
-// 	reactor1.currentHeight = 8
-// 	reactor1.currentRound = 0
-
-// 	// handle the compact block
-// 	reactor1.handleCompactBlock(compactBlock, reactor1.self)
-
-// 	time.Sleep(200 * time.Millisecond)
-
-// 	// check if reactor 1 sent wants to all the connected peers
-// 	_, has := reactor2.getPeer(reactor1.self).GetWants(9, 1)
-// 	require.True(t, has)
-
-// 	_, has = reactor2.getPeer(reactor1.self).GetWants(10, 0)
-// 	require.True(t, has)
-
-// 	_, has = reactor3.getPeer(reactor1.self).GetWants(9, 1)
-// 	require.True(t, has)
-
-// 	_, has = reactor3.getPeer(reactor1.self).GetWants(10, 0)
-// 	require.True(t, has)
-// }
-
-// func createCompactBlock(height int64, round int32) *proptypes.CompactBlock {
-// 	return &proptypes.CompactBlock{
-// 		BpHash:    cmtrand.Bytes(32),
-// 		Signature: cmtrand.Bytes(64),
-// 		LastLen:   0,
-// 		Blobs: []proptypes.TxMetaData{
-// 			{Hash: cmtrand.Bytes(32)},
-// 			{Hash: cmtrand.Bytes(32)},
-// 		},
-// 		Proposal: types.Proposal{
-// 			BlockID: types.BlockID{
-// 				Hash:          nil,
-// 				PartSetHeader: types.PartSetHeader{Total: 30},
-// 			},
-// 			Height: height,
-// 			Round:  round,
-// 		},
-// 	}
-// }
+import (
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+	cfg "github.com/tendermint/tendermint/config"
+	proptypes "github.com/tendermint/tendermint/consensus/propagation/types"
+	cmtrand "github.com/tendermint/tendermint/libs/rand"
+	"github.com/tendermint/tendermint/state"
+	"github.com/tendermint/tendermint/types"
+)
+
+func TestGapCatchup(t *testing.T) {
+	p2pCfg := cfg.DefaultP2PConfig()
+	p2pCfg.SendRate = 100000000
+	p2pCfg.RecvRate = 100000000
+	nodes := 2
+
+	reactors, _ := createTestReactors(nodes, p2pCfg, false, "/home/evan/data/experiments/celestia/fast-recovery/debug")
+	cleanup, _, sm := state.SetupTestCase(t)
+	t.Cleanup(func() {
+		cleanup(t)
+	})
+
+	prop, ps, _, metaData := createTestProposal(sm, 1, 2, 1000000)
+
+	// set the commitment and the data on the first node so that it can respond
+	// to the catchup request
+	n1 := reactors[0]
+	parityBlock, lastLen, err := types.Encode(ps, types.BlockPartSizeBytes)
+	require.NoError(t, err)
+
+	partHashes := extractHashes(ps, parityBlock)
+	proofs := extractProofs(ps, parityBlock)
+
+	cb := &proptypes.CompactBlock{
+		Proposal:    *prop,
+		LastLen:     uint32(lastLen),
+		Signature:   cmtrand.Bytes(64), // todo: sign the proposal with a real signature
+		BpHash:      parityBlock.Hash(),
+		Blobs:       metaData,
+		PartsHashes: partHashes,
+	}
+
+	cb.SetProofCache(proofs)
+
+	added := n1.AddProposal(cb)
+	require.True(t, added)
+
+	_, parts, _, has := n1.getAllState(prop.Height, prop.Round, true)
+	require.True(t, has)
+
+	parts.SetProposalData(ps, parityBlock)
+
+	// add the partset header to the second node and trigger the call to retry
+	// wants
+	n2 := reactors[1]
+
+	_, _, has = n2.GetProposal(prop.Height, prop.Round)
+	require.False(t, has)
+
+	psh := ps.Header()
+	n2.AddCommitment(prop.Height, prop.Round, &psh)
+
+	// this call simulates getting a commitment for a proposal of a higher
+	// height
+	n2.retryWants(2)
+
+	time.Sleep(800 * time.Millisecond)
+
+	_, caughtUp, has := n2.GetProposal(prop.Height, prop.Round)
+	require.True(t, has)
+	require.True(t, caughtUp.IsComplete())
+}