go-pulse/trie/sync_test.go
gary rong ecdbb402ee trie: remove node ordering slice in sync batch (#19929)
When we flush a batch of trie nodes into database during the state
sync, we should guarantee that all children should be flushed before
parent.

Actually the trie nodes commit order is strict by: children -> parent.
But when we flush all ready nodes into db, we don't need the order
anymore since

    (1) they are all ready nodes (no more dependency)
    (2) underlying database provides write atomicity
2019-10-28 18:50:11 +01:00

371 lines
12 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 trie
import (
"bytes"
"testing"
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum/go-ethereum/ethdb/memorydb"
)
// makeTestTrie create a sample test trie to test node-wise reconstruction.
func makeTestTrie() (*Database, *Trie, map[string][]byte) {
// Create an empty trie
triedb := NewDatabase(memorydb.New())
trie, _ := New(common.Hash{}, triedb)
// Fill it with some arbitrary data
content := make(map[string][]byte)
for i := byte(0); i < 255; i++ {
// Map the same data under multiple keys
key, val := common.LeftPadBytes([]byte{1, i}, 32), []byte{i}
content[string(key)] = val
trie.Update(key, val)
key, val = common.LeftPadBytes([]byte{2, i}, 32), []byte{i}
content[string(key)] = val
trie.Update(key, val)
// Add some other data to inflate the trie
for j := byte(3); j < 13; j++ {
key, val = common.LeftPadBytes([]byte{j, i}, 32), []byte{j, i}
content[string(key)] = val
trie.Update(key, val)
}
}
trie.Commit(nil)
// Return the generated trie
return triedb, trie, content
}
// checkTrieContents cross references a reconstructed trie with an expected data
// content map.
func checkTrieContents(t *testing.T, db *Database, root []byte, content map[string][]byte) {
// Check root availability and trie contents
trie, err := New(common.BytesToHash(root), db)
if err != nil {
t.Fatalf("failed to create trie at %x: %v", root, err)
}
if err := checkTrieConsistency(db, common.BytesToHash(root)); err != nil {
t.Fatalf("inconsistent trie at %x: %v", root, err)
}
for key, val := range content {
if have := trie.Get([]byte(key)); !bytes.Equal(have, val) {
t.Errorf("entry %x: content mismatch: have %x, want %x", key, have, val)
}
}
}
// checkTrieConsistency checks that all nodes in a trie are indeed present.
func checkTrieConsistency(db *Database, root common.Hash) error {
// Create and iterate a trie rooted in a subnode
trie, err := New(root, db)
if err != nil {
return nil // Consider a non existent state consistent
}
it := trie.NodeIterator(nil)
for it.Next(true) {
}
return it.Error()
}
// Tests that an empty trie is not scheduled for syncing.
func TestEmptySync(t *testing.T) {
dbA := NewDatabase(memorydb.New())
dbB := NewDatabase(memorydb.New())
emptyA, _ := New(common.Hash{}, dbA)
emptyB, _ := New(emptyRoot, dbB)
for i, trie := range []*Trie{emptyA, emptyB} {
if req := NewSync(trie.Hash(), memorydb.New(), nil, NewSyncBloom(1, memorydb.New())).Missing(1); len(req) != 0 {
t.Errorf("test %d: content requested for empty trie: %v", i, req)
}
}
}
// Tests that given a root hash, a trie can sync iteratively on a single thread,
// requesting retrieval tasks and returning all of them in one go.
func TestIterativeSyncIndividual(t *testing.T) { testIterativeSync(t, 1) }
func TestIterativeSyncBatched(t *testing.T) { testIterativeSync(t, 100) }
func testIterativeSync(t *testing.T, count int) {
// Create a random trie to copy
srcDb, srcTrie, srcData := makeTestTrie()
// Create a destination trie and sync with the scheduler
diskdb := memorydb.New()
triedb := NewDatabase(diskdb)
sched := NewSync(srcTrie.Hash(), diskdb, nil, NewSyncBloom(1, diskdb))
queue := append([]common.Hash{}, sched.Missing(count)...)
for len(queue) > 0 {
results := make([]SyncResult, len(queue))
for i, hash := range queue {
data, err := srcDb.Node(hash)
if err != nil {
t.Fatalf("failed to retrieve node data for %x: %v", hash, err)
}
results[i] = SyncResult{hash, data}
}
if _, index, err := sched.Process(results); err != nil {
t.Fatalf("failed to process result #%d: %v", index, err)
}
batch := diskdb.NewBatch()
if err := sched.Commit(batch); err != nil {
t.Fatalf("failed to commit data: %v", err)
}
batch.Write()
queue = append(queue[:0], sched.Missing(count)...)
}
// Cross check that the two tries are in sync
checkTrieContents(t, triedb, srcTrie.Hash().Bytes(), srcData)
}
// Tests that the trie scheduler can correctly reconstruct the state even if only
// partial results are returned, and the others sent only later.
func TestIterativeDelayedSync(t *testing.T) {
// Create a random trie to copy
srcDb, srcTrie, srcData := makeTestTrie()
// Create a destination trie and sync with the scheduler
diskdb := memorydb.New()
triedb := NewDatabase(diskdb)
sched := NewSync(srcTrie.Hash(), diskdb, nil, NewSyncBloom(1, diskdb))
queue := append([]common.Hash{}, sched.Missing(10000)...)
for len(queue) > 0 {
// Sync only half of the scheduled nodes
results := make([]SyncResult, len(queue)/2+1)
for i, hash := range queue[:len(results)] {
data, err := srcDb.Node(hash)
if err != nil {
t.Fatalf("failed to retrieve node data for %x: %v", hash, err)
}
results[i] = SyncResult{hash, data}
}
if _, index, err := sched.Process(results); err != nil {
t.Fatalf("failed to process result #%d: %v", index, err)
}
batch := diskdb.NewBatch()
if err := sched.Commit(batch); err != nil {
t.Fatalf("failed to commit data: %v", err)
}
batch.Write()
queue = append(queue[len(results):], sched.Missing(10000)...)
}
// Cross check that the two tries are in sync
checkTrieContents(t, triedb, srcTrie.Hash().Bytes(), srcData)
}
// Tests that given a root hash, a trie can sync iteratively on a single thread,
// requesting retrieval tasks and returning all of them in one go, however in a
// random order.
func TestIterativeRandomSyncIndividual(t *testing.T) { testIterativeRandomSync(t, 1) }
func TestIterativeRandomSyncBatched(t *testing.T) { testIterativeRandomSync(t, 100) }
func testIterativeRandomSync(t *testing.T, count int) {
// Create a random trie to copy
srcDb, srcTrie, srcData := makeTestTrie()
// Create a destination trie and sync with the scheduler
diskdb := memorydb.New()
triedb := NewDatabase(diskdb)
sched := NewSync(srcTrie.Hash(), diskdb, nil, NewSyncBloom(1, diskdb))
queue := make(map[common.Hash]struct{})
for _, hash := range sched.Missing(count) {
queue[hash] = struct{}{}
}
for len(queue) > 0 {
// Fetch all the queued nodes in a random order
results := make([]SyncResult, 0, len(queue))
for hash := range queue {
data, err := srcDb.Node(hash)
if err != nil {
t.Fatalf("failed to retrieve node data for %x: %v", hash, err)
}
results = append(results, SyncResult{hash, data})
}
// Feed the retrieved results back and queue new tasks
if _, index, err := sched.Process(results); err != nil {
t.Fatalf("failed to process result #%d: %v", index, err)
}
batch := diskdb.NewBatch()
if err := sched.Commit(batch); err != nil {
t.Fatalf("failed to commit data: %v", err)
}
batch.Write()
queue = make(map[common.Hash]struct{})
for _, hash := range sched.Missing(count) {
queue[hash] = struct{}{}
}
}
// Cross check that the two tries are in sync
checkTrieContents(t, triedb, srcTrie.Hash().Bytes(), srcData)
}
// Tests that the trie scheduler can correctly reconstruct the state even if only
// partial results are returned (Even those randomly), others sent only later.
func TestIterativeRandomDelayedSync(t *testing.T) {
// Create a random trie to copy
srcDb, srcTrie, srcData := makeTestTrie()
// Create a destination trie and sync with the scheduler
diskdb := memorydb.New()
triedb := NewDatabase(diskdb)
sched := NewSync(srcTrie.Hash(), diskdb, nil, NewSyncBloom(1, diskdb))
queue := make(map[common.Hash]struct{})
for _, hash := range sched.Missing(10000) {
queue[hash] = struct{}{}
}
for len(queue) > 0 {
// Sync only half of the scheduled nodes, even those in random order
results := make([]SyncResult, 0, len(queue)/2+1)
for hash := range queue {
data, err := srcDb.Node(hash)
if err != nil {
t.Fatalf("failed to retrieve node data for %x: %v", hash, err)
}
results = append(results, SyncResult{hash, data})
if len(results) >= cap(results) {
break
}
}
// Feed the retrieved results back and queue new tasks
if _, index, err := sched.Process(results); err != nil {
t.Fatalf("failed to process result #%d: %v", index, err)
}
batch := diskdb.NewBatch()
if err := sched.Commit(batch); err != nil {
t.Fatalf("failed to commit data: %v", err)
}
batch.Write()
for _, result := range results {
delete(queue, result.Hash)
}
for _, hash := range sched.Missing(10000) {
queue[hash] = struct{}{}
}
}
// Cross check that the two tries are in sync
checkTrieContents(t, triedb, srcTrie.Hash().Bytes(), srcData)
}
// Tests that a trie sync will not request nodes multiple times, even if they
// have such references.
func TestDuplicateAvoidanceSync(t *testing.T) {
// Create a random trie to copy
srcDb, srcTrie, srcData := makeTestTrie()
// Create a destination trie and sync with the scheduler
diskdb := memorydb.New()
triedb := NewDatabase(diskdb)
sched := NewSync(srcTrie.Hash(), diskdb, nil, NewSyncBloom(1, diskdb))
queue := append([]common.Hash{}, sched.Missing(0)...)
requested := make(map[common.Hash]struct{})
for len(queue) > 0 {
results := make([]SyncResult, len(queue))
for i, hash := range queue {
data, err := srcDb.Node(hash)
if err != nil {
t.Fatalf("failed to retrieve node data for %x: %v", hash, err)
}
if _, ok := requested[hash]; ok {
t.Errorf("hash %x already requested once", hash)
}
requested[hash] = struct{}{}
results[i] = SyncResult{hash, data}
}
if _, index, err := sched.Process(results); err != nil {
t.Fatalf("failed to process result #%d: %v", index, err)
}
batch := diskdb.NewBatch()
if err := sched.Commit(batch); err != nil {
t.Fatalf("failed to commit data: %v", err)
}
batch.Write()
queue = append(queue[:0], sched.Missing(0)...)
}
// Cross check that the two tries are in sync
checkTrieContents(t, triedb, srcTrie.Hash().Bytes(), srcData)
}
// Tests that at any point in time during a sync, only complete sub-tries are in
// the database.
func TestIncompleteSync(t *testing.T) {
// Create a random trie to copy
srcDb, srcTrie, _ := makeTestTrie()
// Create a destination trie and sync with the scheduler
diskdb := memorydb.New()
triedb := NewDatabase(diskdb)
sched := NewSync(srcTrie.Hash(), diskdb, nil, NewSyncBloom(1, diskdb))
var added []common.Hash
queue := append([]common.Hash{}, sched.Missing(1)...)
for len(queue) > 0 {
// Fetch a batch of trie nodes
results := make([]SyncResult, len(queue))
for i, hash := range queue {
data, err := srcDb.Node(hash)
if err != nil {
t.Fatalf("failed to retrieve node data for %x: %v", hash, err)
}
results[i] = SyncResult{hash, data}
}
// Process each of the trie nodes
if _, index, err := sched.Process(results); err != nil {
t.Fatalf("failed to process result #%d: %v", index, err)
}
batch := diskdb.NewBatch()
if err := sched.Commit(batch); err != nil {
t.Fatalf("failed to commit data: %v", err)
}
batch.Write()
for _, result := range results {
added = append(added, result.Hash)
}
// Check that all known sub-tries in the synced trie are complete
for _, root := range added {
if err := checkTrieConsistency(triedb, root); err != nil {
t.Fatalf("trie inconsistent: %v", err)
}
}
// Fetch the next batch to retrieve
queue = append(queue[:0], sched.Missing(1)...)
}
// Sanity check that removing any node from the database is detected
for _, node := range added[1:] {
key := node.Bytes()
value, _ := diskdb.Get(key)
diskdb.Delete(key)
if err := checkTrieConsistency(triedb, added[0]); err == nil {
t.Fatalf("trie inconsistency not caught, missing: %x", key)
}
diskdb.Put(key, value)
}
}