tendermint/consensus/replay_file.go
Anca Zamfir abc30821f4 blockchain: Reorg reactor (#3561)
* go routines in blockchain reactor

* Added reference to the go routine diagram

* Initial commit

* cleanup

* Undo testing_logger change, committed by mistake

* Fix the test loggers

* pulled some fsm code into pool.go

* added pool tests

* changes to the design

added block requests under peer

moved the request trigger in the reactor poolRoutine, triggered now by a ticker

in general moved everything required for making block requests smarter in the poolRoutine

added a simple map of heights to keep track of what will need to be requested next

added a few more tests

* send errors to FSM in a different channel than blocks

send errors (RemovePeer) from switch on a different channel than the
one receiving blocks
renamed channels
added more pool tests

* more pool tests

* lint errors

* more tests

* more tests

* switch fast sync to new implementation

* fixed data race in tests

* cleanup

* finished fsm tests

* address golangci comments :)

* address golangci comments :)

* Added timeout on next block needed to advance

* updating docs and cleanup

* fix issue in test from previous cleanup

* cleanup

* Added termination scenarios, tests and more cleanup

* small fixes to adr, comments and cleanup

* Fix bug in sendRequest()

If we tried to send a request to a peer not present in the switch, a
missing continue statement caused the request to be blackholed in a peer
that was removed and never retried.

While this bug was manifesting, the reactor kept asking for other
blocks that would be stored and never consumed. Added the number of
unconsumed blocks in the math for requesting blocks ahead of current
processing height so eventually there will be no more blocks requested
until the already received ones are consumed.

* remove bpPeer's didTimeout field

* Use distinct err codes for peer timeout and FSM timeouts

* Don't allow peers to update with lower height

* review comments from Ethan and Zarko

* some cleanup, renaming, comments

* Move block execution in separate goroutine

* Remove pool's numPending

* review comments

* fix lint, remove old blockchain reactor and duplicates in fsm tests

* small reorg around peer after review comments

* add the reactor spec

* verify block only once

* review comments

* change to int for max number of pending requests

* cleanup and godoc

* Add configuration flag fast sync version

* golangci fixes

* fix config template

* move both reactor versions under blockchain

* cleanup, golint, renaming stuff

* updated documentation, fixed more golint warnings

* integrate with behavior package

* sync with master

* gofmt

* add changelog_pending entry

* move to improvments

* suggestion to changelog entry
2019-08-01 10:28:32 -07:00

325 lines
8.4 KiB
Go

package consensus
import (
"bufio"
"context"
"fmt"
"io"
"os"
"strconv"
"strings"
"github.com/pkg/errors"
dbm "github.com/tendermint/tm-cmn/db"
cfg "github.com/tendermint/tendermint/config"
cmn "github.com/tendermint/tendermint/libs/common"
"github.com/tendermint/tendermint/libs/log"
"github.com/tendermint/tendermint/mock"
"github.com/tendermint/tendermint/proxy"
sm "github.com/tendermint/tendermint/state"
"github.com/tendermint/tendermint/store"
"github.com/tendermint/tendermint/types"
)
const (
// event bus subscriber
subscriber = "replay-file"
)
//--------------------------------------------------------
// replay messages interactively or all at once
// replay the wal file
func RunReplayFile(config cfg.BaseConfig, csConfig *cfg.ConsensusConfig, console bool) {
consensusState := newConsensusStateForReplay(config, csConfig)
if err := consensusState.ReplayFile(csConfig.WalFile(), console); err != nil {
cmn.Exit(fmt.Sprintf("Error during consensus replay: %v", err))
}
}
// Replay msgs in file or start the console
func (cs *ConsensusState) ReplayFile(file string, console bool) error {
if cs.IsRunning() {
return errors.New("cs is already running, cannot replay")
}
if cs.wal != nil {
return errors.New("cs wal is open, cannot replay")
}
cs.startForReplay()
// ensure all new step events are regenerated as expected
ctx := context.Background()
newStepSub, err := cs.eventBus.Subscribe(ctx, subscriber, types.EventQueryNewRoundStep)
if err != nil {
return errors.Errorf("failed to subscribe %s to %v", subscriber, types.EventQueryNewRoundStep)
}
defer cs.eventBus.Unsubscribe(ctx, subscriber, types.EventQueryNewRoundStep)
// just open the file for reading, no need to use wal
fp, err := os.OpenFile(file, os.O_RDONLY, 0600)
if err != nil {
return err
}
pb := newPlayback(file, fp, cs, cs.state.Copy())
defer pb.fp.Close() // nolint: errcheck
var nextN int // apply N msgs in a row
var msg *TimedWALMessage
for {
if nextN == 0 && console {
nextN = pb.replayConsoleLoop()
}
msg, err = pb.dec.Decode()
if err == io.EOF {
return nil
} else if err != nil {
return err
}
if err := pb.cs.readReplayMessage(msg, newStepSub); err != nil {
return err
}
if nextN > 0 {
nextN--
}
pb.count++
}
}
//------------------------------------------------
// playback manager
type playback struct {
cs *ConsensusState
fp *os.File
dec *WALDecoder
count int // how many lines/msgs into the file are we
// replays can be reset to beginning
fileName string // so we can close/reopen the file
genesisState sm.State // so the replay session knows where to restart from
}
func newPlayback(fileName string, fp *os.File, cs *ConsensusState, genState sm.State) *playback {
return &playback{
cs: cs,
fp: fp,
fileName: fileName,
genesisState: genState,
dec: NewWALDecoder(fp),
}
}
// go back count steps by resetting the state and running (pb.count - count) steps
func (pb *playback) replayReset(count int, newStepSub types.Subscription) error {
pb.cs.Stop()
pb.cs.Wait()
newCS := NewConsensusState(pb.cs.config, pb.genesisState.Copy(), pb.cs.blockExec,
pb.cs.blockStore, pb.cs.txNotifier, pb.cs.evpool)
newCS.SetEventBus(pb.cs.eventBus)
newCS.startForReplay()
if err := pb.fp.Close(); err != nil {
return err
}
fp, err := os.OpenFile(pb.fileName, os.O_RDONLY, 0600)
if err != nil {
return err
}
pb.fp = fp
pb.dec = NewWALDecoder(fp)
count = pb.count - count
fmt.Printf("Reseting from %d to %d\n", pb.count, count)
pb.count = 0
pb.cs = newCS
var msg *TimedWALMessage
for i := 0; i < count; i++ {
msg, err = pb.dec.Decode()
if err == io.EOF {
return nil
} else if err != nil {
return err
}
if err := pb.cs.readReplayMessage(msg, newStepSub); err != nil {
return err
}
pb.count++
}
return nil
}
func (cs *ConsensusState) startForReplay() {
cs.Logger.Error("Replay commands are disabled until someone updates them and writes tests")
/* TODO:!
// since we replay tocks we just ignore ticks
go func() {
for {
select {
case <-cs.tickChan:
case <-cs.Quit:
return
}
}
}()*/
}
// console function for parsing input and running commands
func (pb *playback) replayConsoleLoop() int {
for {
fmt.Printf("> ")
bufReader := bufio.NewReader(os.Stdin)
line, more, err := bufReader.ReadLine()
if more {
cmn.Exit("input is too long")
} else if err != nil {
cmn.Exit(err.Error())
}
tokens := strings.Split(string(line), " ")
if len(tokens) == 0 {
continue
}
switch tokens[0] {
case "next":
// "next" -> replay next message
// "next N" -> replay next N messages
if len(tokens) == 1 {
return 0
}
i, err := strconv.Atoi(tokens[1])
if err != nil {
fmt.Println("next takes an integer argument")
} else {
return i
}
case "back":
// "back" -> go back one message
// "back N" -> go back N messages
// NOTE: "back" is not supported in the state machine design,
// so we restart and replay up to
ctx := context.Background()
// ensure all new step events are regenerated as expected
newStepSub, err := pb.cs.eventBus.Subscribe(ctx, subscriber, types.EventQueryNewRoundStep)
if err != nil {
cmn.Exit(fmt.Sprintf("failed to subscribe %s to %v", subscriber, types.EventQueryNewRoundStep))
}
defer pb.cs.eventBus.Unsubscribe(ctx, subscriber, types.EventQueryNewRoundStep)
if len(tokens) == 1 {
if err := pb.replayReset(1, newStepSub); err != nil {
pb.cs.Logger.Error("Replay reset error", "err", err)
}
} else {
i, err := strconv.Atoi(tokens[1])
if err != nil {
fmt.Println("back takes an integer argument")
} else if i > pb.count {
fmt.Printf("argument to back must not be larger than the current count (%d)\n", pb.count)
} else {
if err := pb.replayReset(i, newStepSub); err != nil {
pb.cs.Logger.Error("Replay reset error", "err", err)
}
}
}
case "rs":
// "rs" -> print entire round state
// "rs short" -> print height/round/step
// "rs <field>" -> print another field of the round state
rs := pb.cs.RoundState
if len(tokens) == 1 {
fmt.Println(rs)
} else {
switch tokens[1] {
case "short":
fmt.Printf("%v/%v/%v\n", rs.Height, rs.Round, rs.Step)
case "validators":
fmt.Println(rs.Validators)
case "proposal":
fmt.Println(rs.Proposal)
case "proposal_block":
fmt.Printf("%v %v\n", rs.ProposalBlockParts.StringShort(), rs.ProposalBlock.StringShort())
case "locked_round":
fmt.Println(rs.LockedRound)
case "locked_block":
fmt.Printf("%v %v\n", rs.LockedBlockParts.StringShort(), rs.LockedBlock.StringShort())
case "votes":
fmt.Println(rs.Votes.StringIndented(" "))
default:
fmt.Println("Unknown option", tokens[1])
}
}
case "n":
fmt.Println(pb.count)
}
}
}
//--------------------------------------------------------------------------------
// convenience for replay mode
func newConsensusStateForReplay(config cfg.BaseConfig, csConfig *cfg.ConsensusConfig) *ConsensusState {
dbType := dbm.DBBackendType(config.DBBackend)
// Get BlockStore
blockStoreDB := dbm.NewDB("blockstore", dbType, config.DBDir())
blockStore := store.NewBlockStore(blockStoreDB)
// Get State
stateDB := dbm.NewDB("state", dbType, config.DBDir())
gdoc, err := sm.MakeGenesisDocFromFile(config.GenesisFile())
if err != nil {
cmn.Exit(err.Error())
}
state, err := sm.MakeGenesisState(gdoc)
if err != nil {
cmn.Exit(err.Error())
}
// Create proxyAppConn connection (consensus, mempool, query)
clientCreator := proxy.DefaultClientCreator(config.ProxyApp, config.ABCI, config.DBDir())
proxyApp := proxy.NewAppConns(clientCreator)
err = proxyApp.Start()
if err != nil {
cmn.Exit(fmt.Sprintf("Error starting proxy app conns: %v", err))
}
eventBus := types.NewEventBus()
if err := eventBus.Start(); err != nil {
cmn.Exit(fmt.Sprintf("Failed to start event bus: %v", err))
}
handshaker := NewHandshaker(stateDB, state, blockStore, gdoc)
handshaker.SetEventBus(eventBus)
err = handshaker.Handshake(proxyApp)
if err != nil {
cmn.Exit(fmt.Sprintf("Error on handshake: %v", err))
}
mempool, evpool := mock.Mempool{}, sm.MockEvidencePool{}
blockExec := sm.NewBlockExecutor(stateDB, log.TestingLogger(), proxyApp.Consensus(), mempool, evpool)
consensusState := NewConsensusState(csConfig, state.Copy(), blockExec,
blockStore, mempool, evpool)
consensusState.SetEventBus(eventBus)
return consensusState
}