This commit is contained in:
jvn 2026-05-21 21:54:47 -07:00 committed by GitHub
commit 05a68e0a26
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 67 additions and 14 deletions

View file

@ -538,10 +538,9 @@ func importHistory(ctx *cli.Context) error {
default:
return fmt.Errorf("unknown --era.format %q (expected 'era1' or 'erae')", format)
}
if err := utils.ImportHistory(chain, dir, network, from); err != nil {
if err := utils.ImportHistory(chain, db, dir, network, from); err != nil {
return err
}
fmt.Printf("Import done in %v\n", time.Since(start))
return nil
}

View file

@ -249,13 +249,10 @@ func readList(filename string) ([]string, error) {
return strings.Split(string(b), "\n"), nil
}
// ImportHistory imports Era1 files containing historical block information,
// ImportHistory imports Era1/Erae files containing historical block information,
// starting from genesis. The assumption is held that the provided chain
// segment in Era1 file should all be canonical and verified.
func ImportHistory(chain *core.BlockChain, dir string, network string, from func(f era.ReadAtSeekCloser) (era.Era, error)) error {
if chain.CurrentSnapBlock().Number.BitLen() != 0 {
return errors.New("history import only supported when starting from genesis")
}
// segment in Era1/Erae file should all be canonical and verified.
func ImportHistory(chain *core.BlockChain, db ethdb.Database, dir string, network string, from func(f era.ReadAtSeekCloser) (era.Era, error)) error {
entries, err := era.ReadDir(dir, network)
if err != nil {
return fmt.Errorf("error reading %s: %w", dir, err)
@ -269,6 +266,13 @@ func ImportHistory(chain *core.BlockChain, dir string, network string, from func
len(checksums), len(entries))
}
// Determine resume point from last successfully imported block.
var resumeBlock uint64
if tail := rawdb.ReadEraImportTail(db); tail != nil {
resumeBlock = *tail
log.Info("Resuming era import", "lastBlock", resumeBlock)
}
var (
start = time.Now()
reported = time.Now()
@ -281,13 +285,33 @@ func ImportHistory(chain *core.BlockChain, dir string, network string, from func
err := func() error {
path := filepath.Join(dir, file)
// Validate against checksum file in directory.
f, err := os.Open(path)
if err != nil {
return fmt.Errorf("open %s: %w", path, err)
}
defer f.Close()
// Peek at era block range to check if we can skip entirely.
e, err := from(f)
if err != nil {
f.Close()
return fmt.Errorf("error opening era: %w", err)
}
eraStart := e.Start()
eraEnd := eraStart + e.Count() - 1
e.Close()
// Skip era files fully behind resume point.
if resumeBlock > 0 && eraEnd <= resumeBlock {
log.Debug("Skipping already imported Era file", "file", file, "eraEnd", eraEnd, "resumeBlock", resumeBlock)
return nil
}
// reopen for checksum + import.
f, err = os.Open(path)
if err != nil {
return fmt.Errorf("open %s: %w", path, err)
}
if _, err := io.Copy(h, f); err != nil {
return fmt.Errorf("checksum %s: %w", path, err)
}
@ -297,8 +321,12 @@ func ImportHistory(chain *core.BlockChain, dir string, network string, from func
if got != checksums[i] {
return fmt.Errorf("%s checksum mismatch: have %s want %s", file, got, checksums[i])
}
// Import all block data from Era1.
e, err := from(f)
// Seek back for import.
if _, err := f.Seek(0, io.SeekStart); err != nil {
return fmt.Errorf("seek %s: %w", path, err)
}
e, err = from(f)
if err != nil {
return fmt.Errorf("error opening era: %w", err)
}
@ -321,10 +349,14 @@ func ImportHistory(chain *core.BlockChain, dir string, network string, from func
return fmt.Errorf("error inserting blocks %d-%d: %w",
blocks[0].NumberU64(), blocks[len(blocks)-1].NumberU64(), err)
}
// Persist tail after each successful batch.
lastBlock := blocks[len(blocks)-1].NumberU64()
rawdb.WriteEraImportTail(db, lastBlock)
resumeBlock = lastBlock
imported += len(blocks)
if time.Since(reported) >= 8*time.Second {
head := blocks[len(blocks)-1].NumberU64()
log.Info("Importing Era files", "head", head, "imported", imported,
log.Info("Importing Era files", "head", lastBlock, "imported", imported,
"elapsed", common.PrettyDuration(time.Since(start)))
imported = 0
reported = time.Now()
@ -334,6 +366,7 @@ func ImportHistory(chain *core.BlockChain, dir string, network string, from func
return nil
}
)
for it.Next() {
block, err := it.Block()
if err != nil {
@ -342,6 +375,10 @@ func ImportHistory(chain *core.BlockChain, dir string, network string, from func
if block.Number().BitLen() == 0 {
continue // skip genesis
}
// Skip blocks already imported (mid-era resume).
if resumeBlock > 0 && block.Number().Uint64() <= resumeBlock {
continue
}
receipts, err := it.Receipts()
if err != nil {
return fmt.Errorf("error reading receipts %d: %w", it.Number(), err)

View file

@ -182,7 +182,7 @@ func TestHistoryImportAndExport(t *testing.T) {
if err != nil {
t.Fatalf("unable to initialize chain: %v", err)
}
if err := ImportHistory(imported, dir, "mainnet", tt.from); err != nil {
if err := ImportHistory(imported, db2, dir, "mainnet", tt.from); err != nil {
t.Fatalf("failed to import chain: %v", err)
}
if have, want := imported.CurrentHeader(), chain.CurrentHeader(); have.Hash() != want.Hash() {

View file

@ -32,6 +32,21 @@ import (
"github.com/ethereum/go-ethereum/rlp"
)
func ReadEraImportTail(db ethdb.KeyValueReader) *uint64 {
data, _ := db.Get(eraImportTailKey)
if len(data) == 0 {
return nil
}
tail := binary.BigEndian.Uint64(data)
return &tail
}
func WriteEraImportTail(db ethdb.KeyValueWriter, blockNumber uint64) {
buf := make([]byte, 8)
binary.BigEndian.PutUint64(buf, blockNumber)
db.Put(eraImportTailKey, buf)
}
// DecodeTxLookupEntry decodes the supplied tx lookup data.
func DecodeTxLookupEntry(data []byte, db ethdb.Reader) *uint64 {
// Database v6 tx lookup just stores the block number

View file

@ -87,6 +87,8 @@ var (
// txIndexTailKey tracks the oldest block whose transactions have been indexed.
txIndexTailKey = []byte("TransactionIndexTail")
eraImportTailKey = []byte("eraImportTail") // eraImportTailKey -> last fully imported block
// fastTxLookupLimitKey tracks the transaction lookup limit during fast sync.
// This flag is deprecated, it's kept to avoid reporting errors when inspect
// database.