mirror of
https://github.com/fluencelabs/tendermint
synced 2025-04-25 06:42:16 +00:00
comment out failing consensus tests for now rewrite rpc httpclient to use new pubsub package import pubsub as tmpubsub, query as tmquery make event IDs constants EventKey -> EventTypeKey rename EventsPubsub to PubSub mempool does not use pubsub rename eventsSub to pubsub new subscribe API fix channel size issues and consensus tests bugs refactor rpc client add missing discardFromChan method add mutex rename pubsub to eventBus remove IsRunning from WSRPCConnection interface (not needed) add a comment in broadcastNewRoundStepsAndVotes rename registerEventCallbacks to broadcastNewRoundStepsAndVotes See https://dave.cheney.net/2014/03/19/channel-axioms stop eventBuses after reactor tests remove unnecessary Unsubscribe return subscribe helper function move discardFromChan to where it is used subscribe now returns an err this gives us ability to refuse to subscribe if pubsub is at its max capacity. use context for control overflow cache queries handle err when subscribing in replay_test rename testClientID to testSubscriber extract var set channel buffer capacity to 1 in replay_file fix byzantine_test unsubscribe from single event, not all events refactor httpclient to return events to appropriate channels return failing testReplayCrashBeforeWriteVote test fix TestValidatorSetChanges refactor code a bit fix testReplayCrashBeforeWriteVote add comment fix TestValidatorSetChanges fixes from Bucky's review update comment [ci skip] test TxEventBuffer update changelog fix TestValidatorSetChanges (2nd attempt) only do wg.Done when no errors benchmark event bus create pubsub server inside NewEventBus only expose config params (later if needed) set buffer capacity to 0 so we are not testing cache new tx event format: key = "Tx" plus a tag {"tx.hash": XYZ} This should allow to subscribe to all transactions! or a specific one using a query: "tm.events.type = Tx and tx.hash = '013ABF99434...'" use TimeoutCommit instead of afterPublishEventNewBlockTimeout TimeoutCommit is the time a node waits after committing a block, before it goes into the next height. So it will finish everything from the last block, but then wait a bit. The idea is this gives it time to hear more votes from other validators, to strengthen the commit it includes in the next block. But it also gives it time to hear about new transactions. waitForBlockWithUpdatedVals rewrite WAL crash tests Task: test that we can recover from any WAL crash. Solution: the old tests were relying on event hub being run in the same thread (we were injecting the private validator's last signature). when considering a rewrite, we considered two possible solutions: write a "fuzzy" testing system where WAL is crashing upon receiving a new message, or inject failures and trigger them in tests using something like https://github.com/coreos/gofail. remove sleep no cs.Lock around wal.Save test different cases (empty block, non-empty block, ...) comments add comments test 4 cases: empty block, non-empty block, non-empty block with smaller part size, many blocks fixes as per Bucky's last review reset subscriptions on UnsubscribeAll use a simple counter to track message for which we panicked also, set a smaller part size for all test cases
299 lines
7.7 KiB
Go
299 lines
7.7 KiB
Go
package consensus
|
|
|
|
import (
|
|
"bufio"
|
|
"context"
|
|
"fmt"
|
|
"io"
|
|
"os"
|
|
"strconv"
|
|
"strings"
|
|
|
|
"github.com/pkg/errors"
|
|
|
|
bc "github.com/tendermint/tendermint/blockchain"
|
|
cfg "github.com/tendermint/tendermint/config"
|
|
"github.com/tendermint/tendermint/proxy"
|
|
sm "github.com/tendermint/tendermint/state"
|
|
"github.com/tendermint/tendermint/types"
|
|
cmn "github.com/tendermint/tmlibs/common"
|
|
dbm "github.com/tendermint/tmlibs/db"
|
|
)
|
|
|
|
//--------------------------------------------------------
|
|
// replay messages interactively or all at once
|
|
|
|
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(cmn.Fmt("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
|
|
newStepCh := make(chan interface{}, 1)
|
|
|
|
ctx := context.Background()
|
|
err := cs.eventBus.Subscribe(ctx, "replay-file", types.EventQueryNewRoundStep, newStepCh)
|
|
if err != nil {
|
|
return errors.Errorf("failed to subscribe replay-file to %v", types.EventQueryNewRoundStep)
|
|
}
|
|
defer cs.eventBus.Unsubscribe(ctx, "replay-file", types.EventQueryNewRoundStep)
|
|
|
|
// just open the file for reading, no need to use wal
|
|
fp, err := os.OpenFile(file, os.O_RDONLY, 0666)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
pb := newPlayback(file, fp, cs, cs.state.Copy())
|
|
defer pb.fp.Close()
|
|
|
|
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, newStepCh); err != nil {
|
|
return err
|
|
}
|
|
|
|
if nextN > 0 {
|
|
nextN -= 1
|
|
}
|
|
pb.count += 1
|
|
}
|
|
return nil
|
|
}
|
|
|
|
//------------------------------------------------
|
|
// 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, newStepCh chan interface{}) error {
|
|
pb.cs.Stop()
|
|
pb.cs.Wait()
|
|
|
|
newCS := NewConsensusState(pb.cs.config, pb.genesisState.Copy(), pb.cs.proxyAppConn, pb.cs.blockStore, pb.cs.mempool)
|
|
newCS.SetEventBus(pb.cs.eventBus)
|
|
newCS.startForReplay()
|
|
|
|
pb.fp.Close()
|
|
fp, err := os.OpenFile(pb.fileName, os.O_RDONLY, 0666)
|
|
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, newStepCh); err != nil {
|
|
return err
|
|
}
|
|
pb.count += 1
|
|
}
|
|
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
|
|
} else {
|
|
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
|
|
newStepCh := make(chan interface{}, 1)
|
|
|
|
err := pb.cs.eventBus.Subscribe(ctx, "replay-file", types.EventQueryNewRoundStep, newStepCh)
|
|
if err != nil {
|
|
cmn.Exit(fmt.Sprintf("failed to subscribe replay-file to %v", types.EventQueryNewRoundStep))
|
|
}
|
|
defer pb.cs.eventBus.Unsubscribe(ctx, "replay-file", types.EventQueryNewRoundStep)
|
|
|
|
if len(tokens) == 1 {
|
|
pb.replayReset(1, newStepCh)
|
|
} 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 {
|
|
pb.replayReset(i, newStepCh)
|
|
}
|
|
}
|
|
|
|
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)
|
|
}
|
|
}
|
|
return 0
|
|
}
|
|
|
|
//--------------------------------------------------------------------------------
|
|
|
|
// convenience for replay mode
|
|
func newConsensusStateForReplay(config cfg.BaseConfig, csConfig *cfg.ConsensusConfig) *ConsensusState {
|
|
// Get BlockStore
|
|
blockStoreDB := dbm.NewDB("blockstore", config.DBBackend, config.DBDir())
|
|
blockStore := bc.NewBlockStore(blockStoreDB)
|
|
|
|
// Get State
|
|
stateDB := dbm.NewDB("state", config.DBBackend, config.DBDir())
|
|
state, err := sm.MakeGenesisStateFromFile(stateDB, config.GenesisFile())
|
|
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, NewHandshaker(state, blockStore))
|
|
_, err = proxyApp.Start()
|
|
if err != nil {
|
|
cmn.Exit(cmn.Fmt("Error starting proxy app conns: %v", err))
|
|
}
|
|
|
|
eventBus := types.NewEventBus()
|
|
if _, err := eventBus.Start(); err != nil {
|
|
cmn.Exit(cmn.Fmt("Failed to start event bus: %v", err))
|
|
}
|
|
|
|
consensusState := NewConsensusState(csConfig, state.Copy(), proxyApp.Consensus(), blockStore, types.MockMempool{})
|
|
|
|
consensusState.SetEventBus(eventBus)
|
|
return consensusState
|
|
}
|