cmd/geth: add prune history command (#31384)

This adds a new subcommand 'geth prune-history' that removes the pre-merge history
on supported networks. Geth is not fully ready to work in this mode, please do not run
this command on your production node.

---------

Co-authored-by: Felix Lange <[email protected]>

Author: s1na

cmd/geth/chaincmd.go

@@ -35,6 +35,7 @@ import (
 	"github.com/ethereum/go-ethereum/core/state"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/crypto"
+	"github.com/ethereum/go-ethereum/eth/ethconfig"
 	"github.com/ethereum/go-ethereum/ethdb"
 	"github.com/ethereum/go-ethereum/internal/era"
 	"github.com/ethereum/go-ethereum/log"
@@ -189,6 +190,18 @@ It's deprecated, please use "geth db import" instead.
 This command dumps out the state for a given block (or latest, if none provided).
 `,
 	}
+
+	pruneCommand = &cli.Command{
+		Action:    pruneHistory,
+		Name:      "prune-history",
+		Usage:     "Prune blockchain history (block bodies and receipts) up to the merge block",
+		ArgsUsage: "",
+		Flags:     utils.DatabaseFlags,
+		Description: `
+The prune-history command removes historical block bodies and receipts from the
+blockchain database up to the merge block, while preserving block headers. This
+helps reduce storage requirements for nodes that don't need full historical data.`,
+	}
 )
 
 // initGenesis will initialise the given JSON format genesis file and writes it as
@@ -598,3 +611,51 @@ func hashish(x string) bool {
 	_, err := strconv.Atoi(x)
 	return err != nil
 }
+
+func pruneHistory(ctx *cli.Context) error {
+	stack, _ := makeConfigNode(ctx)
+	defer stack.Close()
+
+	// Open the chain database
+	chain, chaindb := utils.MakeChain(ctx, stack, false)
+	defer chaindb.Close()
+	defer chain.Stop()
+
+	// Determine the prune point. This will be the first PoS block.
+	prunePoint, ok := ethconfig.HistoryPrunePoints[chain.Genesis().Hash()]
+	if !ok || prunePoint == nil {
+		return errors.New("prune point not found")
+	}
+	var (
+		mergeBlock     = prunePoint.BlockNumber
+		mergeBlockHash = prunePoint.BlockHash.Hex()
+	)
+
+	// Check we're far enough past merge to ensure all data is in freezer
+	currentHeader := chain.CurrentHeader()
+	if currentHeader == nil {
+		return errors.New("current header not found")
+	}
+	if currentHeader.Number.Uint64() < mergeBlock+params.FullImmutabilityThreshold {
+		return fmt.Errorf("chain not far enough past merge block, need %d more blocks",
+			mergeBlock+params.FullImmutabilityThreshold-currentHeader.Number.Uint64())
+	}
+
+	// Double-check the prune block in db has the expected hash.
+	hash := rawdb.ReadCanonicalHash(chaindb, mergeBlock)
+	if hash != common.HexToHash(mergeBlockHash) {
+		return fmt.Errorf("merge block hash mismatch: got %s, want %s", hash.Hex(), mergeBlockHash)
+	}
+
+	log.Info("Starting history pruning", "head", currentHeader.Number, "tail", mergeBlock, "tailHash", mergeBlockHash)
+	start := time.Now()
+	rawdb.PruneTransactionIndex(chaindb, mergeBlock)
+	if _, err := chaindb.TruncateTail(mergeBlock); err != nil {
+		return fmt.Errorf("failed to truncate ancient data: %v", err)
+	}
+	log.Info("History pruning completed", "tail", mergeBlock, "elapsed", common.PrettyDuration(time.Since(start)))
+
+	// TODO(s1na): what if there is a crash between the two prune operations?
+
+	return nil
+}

cmd/geth/main.go

@@ -226,6 +226,7 @@ func init() {
 		removedbCommand,
 		dumpCommand,
 		dumpGenesisCommand,
+		pruneCommand,
 		// See accountcmd.go:
 		accountCommand,
 		walletCommand,

core/blockchain.go

@@ -332,7 +332,8 @@ func NewBlockChain(db ethdb.Database, cacheConfig *CacheConfig, genesis *Genesis
 	bc.prefetcher = newStatePrefetcher(chainConfig, bc.hc)
 	bc.processor = NewStateProcessor(chainConfig, bc.hc)
 
-	bc.genesisBlock = bc.GetBlockByNumber(0)
+	genesisHeader := bc.GetHeaderByNumber(0)
+	bc.genesisBlock = types.NewBlockWithHeader(genesisHeader)
 	if bc.genesisBlock == nil {
 		return nil, ErrNoGenesis
 	}

core/rawdb/accessors_indexes.go

@@ -103,13 +103,14 @@ func DeleteTxLookupEntries(db ethdb.KeyValueWriter, hashes []common.Hash) {
 // DeleteAllTxLookupEntries purges all the transaction indexes in the database.
 // If condition is specified, only the entry with condition as True will be
 // removed; If condition is not specified, the entry is deleted.
-func DeleteAllTxLookupEntries(db ethdb.KeyValueStore, condition func([]byte) bool) {
+func DeleteAllTxLookupEntries(db ethdb.KeyValueStore, condition func(common.Hash, []byte) bool) {
 	iter := NewKeyLengthIterator(db.NewIterator(txLookupPrefix, nil), common.HashLength+len(txLookupPrefix))
 	defer iter.Release()
 
 	batch := db.NewBatch()
 	for iter.Next() {
-		if condition == nil || condition(iter.Value()) {
+		txhash := common.Hash(iter.Key()[1:])
+		if condition == nil || condition(txhash, iter.Value()) {
 			batch.Delete(iter.Key())
 		}
 		if batch.ValueSize() >= ethdb.IdealBatchSize {

core/rawdb/chain_iterator.go

@@ -17,6 +17,7 @@
 package rawdb
 
 import (
+	"encoding/binary"
 	"runtime"
 	"sync/atomic"
 	"time"
@@ -361,3 +362,38 @@ func UnindexTransactions(db ethdb.Database, from uint64, to uint64, interrupt ch
 func unindexTransactionsForTesting(db ethdb.Database, from uint64, to uint64, interrupt chan struct{}, hook func(uint64) bool) {
 	unindexTransactions(db, from, to, interrupt, hook, false)
 }
+
+// PruneTransactionIndex removes all tx index entries below a certain block number.
+func PruneTransactionIndex(db ethdb.Database, pruneBlock uint64) {
+	tail := ReadTxIndexTail(db)
+	if tail == nil || *tail > pruneBlock {
+		return // no index, or index ends above pruneBlock
+	}
+	// There are blocks below pruneBlock in the index. Iterate the entire index to remove
+	// their entries. Note if this fails, the index is messed up, but tail still points to
+	// the old tail.
+	var count, removed int
+	DeleteAllTxLookupEntries(db, func(txhash common.Hash, v []byte) bool {
+		count++
+		if count%10000000 == 0 {
+			log.Info("Pruning tx index", "count", count, "removed", removed)
+		}
+		if len(v) > 8 {
+			log.Error("Skipping legacy tx index entry", "hash", txhash)
+			return false
+		}
+		bn := decodeNumber(v)
+		if bn < pruneBlock {
+			removed++
+			return true
+		}
+		return false
+	})
+	WriteTxIndexTail(db, pruneBlock)
+}
+
+func decodeNumber(b []byte) uint64 {
+	var numBuffer [8]byte
+	copy(numBuffer[8-len(b):], b)
+	return binary.BigEndian.Uint64(numBuffer[:])
+}

core/rawdb/chain_iterator_test.go

@@ -25,6 +25,7 @@ import (
 
 	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/core/types"
+	"github.com/ethereum/go-ethereum/ethdb"
 )
 
 func TestChainIterator(t *testing.T) {
@@ -102,19 +103,18 @@ func TestChainIterator(t *testing.T) {
 	}
 }
 
-func TestIndexTransactions(t *testing.T) {
-	// Construct test chain db
-	chainDb := NewMemoryDatabase()
-
-	var block *types.Block
+func initDatabaseWithTransactions(db ethdb.Database) ([]*types.Block, []*types.Transaction) {
+	var blocks []*types.Block
 	var txs []*types.Transaction
 	to := common.BytesToAddress([]byte{0x11})
 
 	// Write empty genesis block
-	block = types.NewBlock(&types.Header{Number: big.NewInt(int64(0))}, nil, nil, newTestHasher())
-	WriteBlock(chainDb, block)
-	WriteCanonicalHash(chainDb, block.Hash(), block.NumberU64())
+	block := types.NewBlock(&types.Header{Number: big.NewInt(int64(0))}, nil, nil, newTestHasher())
+	WriteBlock(db, block)
+	WriteCanonicalHash(db, block.Hash(), block.NumberU64())
+	blocks = append(blocks, block)
 
+	// Create transactions.
 	for i := uint64(1); i <= 10; i++ {
 		var tx *types.Transaction
 		if i%2 == 0 {
@@ -138,48 +138,59 @@ func TestIndexTransactions(t *testing.T) {
 			})
 		}
 		txs = append(txs, tx)
-		block = types.NewBlock(&types.Header{Number: big.NewInt(int64(i))}, &types.Body{Transactions: types.Transactions{tx}}, nil, newTestHasher())
-		WriteBlock(chainDb, block)
-		WriteCanonicalHash(chainDb, block.Hash(), block.NumberU64())
+		block := types.NewBlock(&types.Header{Number: big.NewInt(int64(i))}, &types.Body{Transactions: types.Transactions{tx}}, nil, newTestHasher())
+		WriteBlock(db, block)
+		WriteCanonicalHash(db, block.Hash(), block.NumberU64())
+		blocks = append(blocks, block)
 	}
+
+	return blocks, txs
+}
+
+func TestIndexTransactions(t *testing.T) {
+	// Construct test chain db
+	chainDB := NewMemoryDatabase()
+
+	_, txs := initDatabaseWithTransactions(chainDB)
+
 	// verify checks whether the tx indices in the range [from, to)
 	// is expected.
 	verify := func(from, to int, exist bool, tail uint64) {
 		for i := from; i < to; i++ {
 			if i == 0 {
 				continue
 			}
-			number := ReadTxLookupEntry(chainDb, txs[i-1].Hash())
+			number := ReadTxLookupEntry(chainDB, txs[i-1].Hash())
 			if exist && number == nil {
 				t.Fatalf("Transaction index %d missing", i)
 			}
 			if !exist && number != nil {
 				t.Fatalf("Transaction index %d is not deleted", i)
 			}
 		}
-		number := ReadTxIndexTail(chainDb)
+		number := ReadTxIndexTail(chainDB)
 		if number == nil || *number != tail {
 			t.Fatalf("Transaction tail mismatch")
 		}
 	}
-	IndexTransactions(chainDb, 5, 11, nil, false)
+	IndexTransactions(chainDB, 5, 11, nil, false)
 	verify(5, 11, true, 5)
 	verify(0, 5, false, 5)
 
-	IndexTransactions(chainDb, 0, 5, nil, false)
+	IndexTransactions(chainDB, 0, 5, nil, false)
 	verify(0, 11, true, 0)
 
-	UnindexTransactions(chainDb, 0, 5, nil, false)
+	UnindexTransactions(chainDB, 0, 5, nil, false)
 	verify(5, 11, true, 5)
 	verify(0, 5, false, 5)
 
-	UnindexTransactions(chainDb, 5, 11, nil, false)
+	UnindexTransactions(chainDB, 5, 11, nil, false)
 	verify(0, 11, false, 11)
 
 	// Testing corner cases
 	signal := make(chan struct{})
 	var once sync.Once
-	indexTransactionsForTesting(chainDb, 5, 11, signal, func(n uint64) bool {
+	indexTransactionsForTesting(chainDB, 5, 11, signal, func(n uint64) bool {
 		if n <= 8 {
 			once.Do(func() {
 				close(signal)
@@ -190,11 +201,11 @@ func TestIndexTransactions(t *testing.T) {
 	})
 	verify(9, 11, true, 9)
 	verify(0, 9, false, 9)
-	IndexTransactions(chainDb, 0, 9, nil, false)
+	IndexTransactions(chainDB, 0, 9, nil, false)
 
 	signal = make(chan struct{})
 	var once2 sync.Once
-	unindexTransactionsForTesting(chainDb, 0, 11, signal, func(n uint64) bool {
+	unindexTransactionsForTesting(chainDB, 0, 11, signal, func(n uint64) bool {
 		if n >= 8 {
 			once2.Do(func() {
 				close(signal)
@@ -206,3 +217,37 @@ func TestIndexTransactions(t *testing.T) {
 	verify(8, 11, true, 8)
 	verify(0, 8, false, 8)
 }
+
+func TestPruneTransactionIndex(t *testing.T) {
+	chainDB := NewMemoryDatabase()
+	blocks, _ := initDatabaseWithTransactions(chainDB)
+	lastBlock := blocks[len(blocks)-1].NumberU64()
+	pruneBlock := lastBlock - 3
+
+	IndexTransactions(chainDB, 0, lastBlock+1, nil, false)
+
+	// Check all transactions are in index.
+	for _, block := range blocks {
+		for _, tx := range block.Transactions() {
+			num := ReadTxLookupEntry(chainDB, tx.Hash())
+			if num == nil || *num != block.NumberU64() {
+				t.Fatalf("wrong TxLookup entry: %x -> %v", tx.Hash(), num)
+			}
+		}
+	}
+
+	PruneTransactionIndex(chainDB, pruneBlock)
+
+	// Check transactions from old blocks not included.
+	for _, block := range blocks {
+		for _, tx := range block.Transactions() {
+			num := ReadTxLookupEntry(chainDB, tx.Hash())
+			if block.NumberU64() < pruneBlock && num != nil {
+				t.Fatalf("TxLookup entry not removed: %x -> %v", tx.Hash(), num)
+			}
+			if block.NumberU64() >= pruneBlock && (num == nil || *num != block.NumberU64()) {
+				t.Fatalf("wrong TxLookup entry after pruning: %x -> %v", tx.Hash(), num)
+			}
+		}
+	}
+}

core/txindexer.go

@@ -20,6 +20,7 @@ import (
 	"errors"
 	"fmt"
 
+	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/ethdb"
 	"github.com/ethereum/go-ethereum/log"
@@ -186,7 +187,7 @@ func (indexer *txIndexer) repair(head uint64) {
 		// potentially leaving dangling indexes in the database.
 		// However, this is considered acceptable.
 		rawdb.WriteTxIndexTail(indexer.db, indexer.cutoff)
-		rawdb.DeleteAllTxLookupEntries(indexer.db, func(blob []byte) bool {
+		rawdb.DeleteAllTxLookupEntries(indexer.db, func(txhash common.Hash, blob []byte) bool {
 			n := rawdb.DecodeTxLookupEntry(blob, indexer.db)
 			return n != nil && *n < indexer.cutoff
 		})