plugeth/eth/sync.go
Felix Lange 56ed6152a1 core, eth, miner: improve shutdown synchronisation
Shutting down geth prints hundreds of annoying error messages in some
cases. The errors appear because the Stop method of eth.ProtocolManager,
miner.Miner and core.TxPool is asynchronous. Left over peer sessions
generate events which are processed after Stop even though the database
has already been closed.

The fix is to make Stop synchronous using sync.WaitGroup.

For eth.ProtocolManager, in order to make use of WaitGroup safe, we need
a way to stop new peer sessions from being added while waiting on the
WaitGroup. The eth protocol Run function now selects on a signaling
channel and adds to the WaitGroup only if ProtocolManager is not
shutting down.

For miner.worker and core.TxPool the number of goroutines is static,
WaitGroup can be used in the usual way without additional
synchronisation.
2016-05-09 13:03:08 +02:00

185 lines
5.4 KiB
Go

// Copyright 2015 The go-ethereum Authors
// This file is part of the go-ethereum library.
//
// The go-ethereum library is free software: you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// The go-ethereum library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
package eth
import (
"math/rand"
"time"
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum/go-ethereum/core/types"
"github.com/ethereum/go-ethereum/eth/downloader"
"github.com/ethereum/go-ethereum/logger"
"github.com/ethereum/go-ethereum/logger/glog"
"github.com/ethereum/go-ethereum/p2p/discover"
)
const (
forceSyncCycle = 10 * time.Second // Time interval to force syncs, even if few peers are available
minDesiredPeerCount = 5 // Amount of peers desired to start syncing
// This is the target size for the packs of transactions sent by txsyncLoop.
// A pack can get larger than this if a single transactions exceeds this size.
txsyncPackSize = 100 * 1024
)
type txsync struct {
p *peer
txs []*types.Transaction
}
// syncTransactions starts sending all currently pending transactions to the given peer.
func (pm *ProtocolManager) syncTransactions(p *peer) {
txs := pm.txpool.GetTransactions()
if len(txs) == 0 {
return
}
select {
case pm.txsyncCh <- &txsync{p, txs}:
case <-pm.quitSync:
}
}
// txsyncLoop takes care of the initial transaction sync for each new
// connection. When a new peer appears, we relay all currently pending
// transactions. In order to minimise egress bandwidth usage, we send
// the transactions in small packs to one peer at a time.
func (pm *ProtocolManager) txsyncLoop() {
var (
pending = make(map[discover.NodeID]*txsync)
sending = false // whether a send is active
pack = new(txsync) // the pack that is being sent
done = make(chan error, 1) // result of the send
)
// send starts a sending a pack of transactions from the sync.
send := func(s *txsync) {
// Fill pack with transactions up to the target size.
size := common.StorageSize(0)
pack.p = s.p
pack.txs = pack.txs[:0]
for i := 0; i < len(s.txs) && size < txsyncPackSize; i++ {
pack.txs = append(pack.txs, s.txs[i])
size += s.txs[i].Size()
}
// Remove the transactions that will be sent.
s.txs = s.txs[:copy(s.txs, s.txs[len(pack.txs):])]
if len(s.txs) == 0 {
delete(pending, s.p.ID())
}
// Send the pack in the background.
glog.V(logger.Detail).Infof("%v: sending %d transactions (%v)", s.p.Peer, len(pack.txs), size)
sending = true
go func() { done <- pack.p.SendTransactions(pack.txs) }()
}
// pick chooses the next pending sync.
pick := func() *txsync {
if len(pending) == 0 {
return nil
}
n := rand.Intn(len(pending)) + 1
for _, s := range pending {
if n--; n == 0 {
return s
}
}
return nil
}
for {
select {
case s := <-pm.txsyncCh:
pending[s.p.ID()] = s
if !sending {
send(s)
}
case err := <-done:
sending = false
// Stop tracking peers that cause send failures.
if err != nil {
glog.V(logger.Debug).Infof("%v: tx send failed: %v", pack.p.Peer, err)
delete(pending, pack.p.ID())
}
// Schedule the next send.
if s := pick(); s != nil {
send(s)
}
case <-pm.quitSync:
return
}
}
}
// syncer is responsible for periodically synchronising with the network, both
// downloading hashes and blocks as well as handling the announcement handler.
func (pm *ProtocolManager) syncer() {
// Start and ensure cleanup of sync mechanisms
pm.fetcher.Start()
defer pm.fetcher.Stop()
defer pm.downloader.Terminate()
// Wait for different events to fire synchronisation operations
forceSync := time.Tick(forceSyncCycle)
for {
select {
case <-pm.newPeerCh:
// Make sure we have peers to select from, then sync
if pm.peers.Len() < minDesiredPeerCount {
break
}
go pm.synchronise(pm.peers.BestPeer())
case <-forceSync:
// Force a sync even if not enough peers are present
go pm.synchronise(pm.peers.BestPeer())
case <-pm.noMorePeers:
return
}
}
}
// synchronise tries to sync up our local block chain with a remote peer.
func (pm *ProtocolManager) synchronise(peer *peer) {
// Short circuit if no peers are available
if peer == nil {
return
}
// Make sure the peer's TD is higher than our own. If not drop.
td := pm.blockchain.GetTd(pm.blockchain.CurrentBlock().Hash())
if peer.Td().Cmp(td) <= 0 {
return
}
// Otherwise try to sync with the downloader
mode := downloader.FullSync
if pm.fastSync {
mode = downloader.FastSync
}
if err := pm.downloader.Synchronise(peer.id, peer.Head(), peer.Td(), mode); err != nil {
return
}
// If fast sync was enabled, and we synced up, disable it
if pm.fastSync {
// Disable fast sync if we indeed have something in our chain
if pm.blockchain.CurrentBlock().NumberU64() > 0 {
glog.V(logger.Info).Infof("fast sync complete, auto disabling")
pm.fastSync = false
}
}
}