review fixes

This commit is contained in:
Ian Norden
2019-12-02 11:26:44 -06:00
parent a834e55b9f
commit db0f024088
15 changed files with 313 additions and 204 deletions
+51 -67
View File
@@ -17,10 +17,10 @@
package storage
import (
"errors"
"fmt"
"sync/atomic"
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum/go-ethereum/rlp"
"github.com/ethereum/go-ethereum/statediff"
"github.com/sirupsen/logrus"
@@ -30,13 +30,13 @@ import (
)
const (
DefaultMaxBatchSize uint64 = 1000
DefaultMaxBatchSize uint64 = 100
defaultMaxBatchNumber int64 = 10
)
// BackFiller is the backfilling interface
type BackFiller interface {
BackFill(endingBlock uint64, backFill chan utils.StorageDiff, errChan chan error, done chan bool) error
BackFill(startingBlock, endingBlock uint64, backFill chan utils.StorageDiff, errChan chan error, done chan bool) error
}
// backFiller is the backfilling struct
@@ -47,52 +47,33 @@ type backFiller struct {
}
// NewStorageBackFiller returns a BackFiller
func NewStorageBackFiller(fetcher fetcher.StateDiffFetcher, startingBlock, batchSize uint64) BackFiller {
func NewStorageBackFiller(fetcher fetcher.StateDiffFetcher, batchSize uint64) BackFiller {
if batchSize == 0 {
batchSize = DefaultMaxBatchSize
}
return &backFiller{
fetcher: fetcher,
batchSize: batchSize,
startingBlock: startingBlock,
fetcher: fetcher,
batchSize: batchSize,
}
}
// BackFill fetches, processes, and returns utils.StorageDiffs over a range of blocks
// It splits a large range up into smaller chunks, batch fetching and processing those chunks concurrently
func (bf *backFiller) BackFill(endingBlock uint64, backFill chan utils.StorageDiff, errChan chan error, done chan bool) error {
if endingBlock < bf.startingBlock {
return errors.New("backfill: ending block number needs to be greater than starting block number")
}
logrus.Infof("going to fill in gap from %d to %d", bf.startingBlock, endingBlock)
// break the range up into bins of smaller ranges
length := endingBlock - bf.startingBlock + 1
numberOfBins := length / bf.batchSize
remainder := length % bf.batchSize
if remainder != 0 {
numberOfBins++
}
blockRangeBins := make([][]uint64, numberOfBins)
for i := range blockRangeBins {
nextBinStart := bf.startingBlock + uint64(bf.batchSize)
if nextBinStart > endingBlock {
nextBinStart = endingBlock + 1
}
blockRange := make([]uint64, 0, nextBinStart-bf.startingBlock+1)
for j := bf.startingBlock; j < nextBinStart; j++ {
blockRange = append(blockRange, j)
}
bf.startingBlock = nextBinStart
blockRangeBins[i] = blockRange
}
func (bf *backFiller) BackFill(startingBlock, endingBlock uint64, backFill chan utils.StorageDiff, errChan chan error, done chan bool) error {
logrus.Infof("going to fill in gap from %d to %d", startingBlock, endingBlock)
// break the range up into bins of smaller ranges
blockRangeBins, err := utils.GetBlockHeightBins(startingBlock, endingBlock, bf.batchSize)
if err != nil {
return err
}
// int64 for atomic incrementing and decrementing to track the number of active processing goroutines we have
var activeCount int64
// channel for processing goroutines to signal when they are done
processingDone := make(chan [2]uint64)
forwardDone := make(chan bool)
// for each block range bin spin up a goroutine to batch fetch and process state diffs for that range
// for each block range bin spin up a goroutine to batch fetch and process state diffs in that range
go func() {
for _, blockHeights := range blockRangeBins {
// if we have reached our limit of active goroutines
@@ -101,39 +82,7 @@ func (bf *backFiller) BackFill(endingBlock uint64, backFill chan utils.StorageDi
// this blocks until a process signals it has finished
<-forwardDone
}
go func(blockHeights []uint64) {
payloads, fetchErr := bf.fetcher.FetchStateDiffsAt(blockHeights)
if fetchErr != nil {
errChan <- fetchErr
}
for _, payload := range payloads {
stateDiff := new(statediff.StateDiff)
stateDiffDecodeErr := rlp.DecodeBytes(payload.StateDiffRlp, stateDiff)
if stateDiffDecodeErr != nil {
errChan <- stateDiffDecodeErr
continue
}
accounts := utils.GetAccountsFromDiff(*stateDiff)
for _, account := range accounts {
logrus.Trace(fmt.Sprintf("iterating through %d Storage values on account", len(account.Storage)))
for _, storage := range account.Storage {
diff, formatErr := utils.FromGethStateDiff(account, stateDiff, storage)
if formatErr != nil {
errChan <- formatErr
continue
}
logrus.Trace("adding storage diff to results",
"keccak of address: ", diff.HashedAddress.Hex(),
"block height: ", diff.BlockHeight,
"storage key: ", diff.StorageKey.Hex(),
"storage value: ", diff.StorageValue.Hex())
backFill <- diff
}
}
}
// when this goroutine is done, send out a signal
processingDone <- [2]uint64{blockHeights[0], blockHeights[len(blockHeights)-1]}
}(blockHeights)
go bf.backFillRange(blockHeights, backFill, errChan, processingDone)
}
}()
@@ -153,7 +102,7 @@ func (bf *backFiller) BackFill(endingBlock uint64, backFill chan utils.StorageDi
}
logrus.Infof("finished fetching gap sub-bin from %d to %d", doneWithHeights[0], doneWithHeights[1])
goroutinesFinished++
if goroutinesFinished == int(numberOfBins) {
if goroutinesFinished >= len(blockRangeBins) {
done <- true
return
}
@@ -163,3 +112,38 @@ func (bf *backFiller) BackFill(endingBlock uint64, backFill chan utils.StorageDi
return nil
}
func (bf *backFiller) backFillRange(blockHeights []uint64, diffChan chan utils.StorageDiff, errChan chan error, doneChan chan [2]uint64) {
payloads, fetchErr := bf.fetcher.FetchStateDiffsAt(blockHeights)
if fetchErr != nil {
errChan <- fetchErr
}
for _, payload := range payloads {
stateDiff := new(statediff.StateDiff)
stateDiffDecodeErr := rlp.DecodeBytes(payload.StateDiffRlp, stateDiff)
if stateDiffDecodeErr != nil {
errChan <- stateDiffDecodeErr
continue
}
accounts := utils.GetAccountsFromDiff(*stateDiff)
for _, account := range accounts {
logrus.Trace(fmt.Sprintf("iterating through %d Storage values on account with key %s", len(account.Storage), common.BytesToHash(account.Key).Hex()))
for _, storage := range account.Storage {
diff, formatErr := utils.FromGethStateDiff(account, stateDiff, storage)
if formatErr != nil {
logrus.Error("failed to format utils.StorageDiff from storage with key: ", common.BytesToHash(storage.Key), "from account with key: ", common.BytesToHash(account.Key))
errChan <- formatErr
continue
}
logrus.Trace("adding storage diff to results",
"keccak of address: ", diff.HashedAddress.Hex(),
"block height: ", diff.BlockHeight,
"storage key: ", diff.StorageKey.Hex(),
"storage value: ", diff.StorageValue.Hex())
diffChan <- diff
}
}
}
// when this is done, send out a signal
doneChan <- [2]uint64{blockHeights[0], blockHeights[len(blockHeights)-1]}
}
+10 -5
View File
@@ -44,11 +44,12 @@ var _ = Describe("BackFiller", func() {
})
It("batch calls statediff_stateDiffAt", func() {
backFiller = storage.NewStorageBackFiller(mockFetcher, test_data.BlockNumber.Uint64(), 100)
backFiller = storage.NewStorageBackFiller(mockFetcher, 100)
backFill := make(chan utils.StorageDiff)
done := make(chan bool)
errChan := make(chan error)
backFillInitErr := backFiller.BackFill(
test_data.BlockNumber.Uint64(),
test_data.BlockNumber2.Uint64(),
backFill,
errChan,
@@ -77,11 +78,12 @@ var _ = Describe("BackFiller", func() {
})
It("has a configurable batch size", func() {
backFiller = storage.NewStorageBackFiller(mockFetcher, test_data.BlockNumber.Uint64(), 1)
backFiller = storage.NewStorageBackFiller(mockFetcher, 1)
backFill := make(chan utils.StorageDiff)
done := make(chan bool)
errChan := make(chan error)
backFillInitErr := backFiller.BackFill(
test_data.BlockNumber.Uint64(),
test_data.BlockNumber2.Uint64(),
backFill,
errChan,
@@ -116,11 +118,12 @@ var _ = Describe("BackFiller", func() {
}
mockFetcher.PayloadsToReturn = payloadsToReturn
// batch size of 2 with 1001 block range => 501 bins
backFiller = storage.NewStorageBackFiller(mockFetcher, test_data.BlockNumber.Uint64(), 2)
backFiller = storage.NewStorageBackFiller(mockFetcher, 2)
backFill := make(chan utils.StorageDiff)
done := make(chan bool)
errChan := make(chan error)
backFillInitErr := backFiller.BackFill(
test_data.BlockNumber.Uint64(),
test_data.BlockNumber.Uint64()+1000,
backFill,
errChan,
@@ -151,11 +154,12 @@ var _ = Describe("BackFiller", func() {
mockFetcher.FetchErrs = map[uint64]error{
test_data.BlockNumber.Uint64(): errors.New("mock fetcher error"),
}
backFiller = storage.NewStorageBackFiller(mockFetcher, test_data.BlockNumber.Uint64(), 1)
backFiller = storage.NewStorageBackFiller(mockFetcher, 1)
backFill := make(chan utils.StorageDiff)
done := make(chan bool)
errChan := make(chan error)
backFillInitErr := backFiller.BackFill(
test_data.BlockNumber.Uint64(),
test_data.BlockNumber2.Uint64(),
backFill,
errChan,
@@ -188,11 +192,12 @@ var _ = Describe("BackFiller", func() {
test_data.BlockNumber2.Uint64(): errors.New("mock fetcher error"),
}
mockFetcher.CalledTimes = 0
backFiller = storage.NewStorageBackFiller(mockFetcher, test_data.BlockNumber.Uint64(), 1)
backFiller = storage.NewStorageBackFiller(mockFetcher, 1)
backFill = make(chan utils.StorageDiff)
done = make(chan bool)
errChan = make(chan error)
backFillInitErr = backFiller.BackFill(
test_data.BlockNumber.Uint64(),
test_data.BlockNumber2.Uint64(),
backFill,
errChan,
+45
View File
@@ -0,0 +1,45 @@
// VulcanizeDB
// Copyright © 2019 Vulcanize
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
// This program 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 Affero General Public License for more details.
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package utils
import "errors"
func GetBlockHeightBins(startingBlock, endingBlock, batchSize uint64) ([][]uint64, error) {
if endingBlock < startingBlock {
return nil, errors.New("backfill: ending block number needs to be greater than starting block number")
}
if batchSize == 0 {
return nil, errors.New("backfill: batchsize needs to be greater than zero")
}
length := endingBlock - startingBlock + 1
numberOfBins := length / batchSize
remainder := length % batchSize
if remainder != 0 {
numberOfBins++
}
blockRangeBins := make([][]uint64, numberOfBins)
for i := range blockRangeBins {
nextBinStart := startingBlock + batchSize
blockRange := make([]uint64, 0, nextBinStart-startingBlock+1)
for j := startingBlock; j < nextBinStart && j <= endingBlock; j++ {
blockRange = append(blockRange, j)
}
startingBlock = nextBinStart
blockRangeBins[i] = blockRange
}
return blockRangeBins, nil
}
@@ -0,0 +1,66 @@
// VulcanizeDB
// Copyright © 2019 Vulcanize
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
// This program 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 Affero General Public License for more details.
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package utils_test
import (
. "github.com/onsi/ginkgo"
. "github.com/onsi/gomega"
"github.com/vulcanize/vulcanizedb/libraries/shared/storage/utils"
)
var _ = Describe("GetBlockHeightBins", func() {
It("splits a block range up into bins", func() {
var startingBlock uint64 = 1
var endingBlock uint64 = 10101
var batchSize uint64 = 100
blockRangeBins, err := utils.GetBlockHeightBins(startingBlock, endingBlock, batchSize)
Expect(err).ToNot(HaveOccurred())
Expect(len(blockRangeBins)).To(Equal(102))
Expect(blockRangeBins[101]).To(Equal([]uint64{10101}))
startingBlock = 101
endingBlock = 10100
batchSize = 100
lastBin := make([]uint64, 0)
for i := 10001; i <= 10100; i++ {
lastBin = append(lastBin, uint64(i))
}
blockRangeBins, err = utils.GetBlockHeightBins(startingBlock, endingBlock, batchSize)
Expect(err).ToNot(HaveOccurred())
Expect(len(blockRangeBins)).To(Equal(100))
Expect(blockRangeBins[99]).To(Equal(lastBin))
})
It("throws an error if the starting block is higher than the ending block", func() {
var startingBlock uint64 = 10102
var endingBlock uint64 = 10101
var batchSize uint64 = 100
_, err := utils.GetBlockHeightBins(startingBlock, endingBlock, batchSize)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("ending block number needs to be greater than starting block number"))
})
It("throws an error if the batch size is zero", func() {
var startingBlock uint64 = 1
var endingBlock uint64 = 10101
var batchSize uint64 = 0
_, err := utils.GetBlockHeightBins(startingBlock, endingBlock, batchSize)
Expect(err).To(HaveOccurred())
Expect(err.Error()).To(ContainSubstring("batchsize needs to be greater than zero"))
})
})