|
| 1 | +// Copyright 2022 The go-ethereum Authors |
| 2 | +// This file is part of the go-ethereum library. |
| 3 | +// |
| 4 | +// The go-ethereum library is free software: you can redistribute it and/or modify |
| 5 | +// it under the terms of the GNU Lesser General Public License as published by |
| 6 | +// the Free Software Foundation, either version 3 of the License, or |
| 7 | +// (at your option) any later version. |
| 8 | +// |
| 9 | +// The go-ethereum library is distributed in the hope that it will be useful, |
| 10 | +// but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 11 | +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 12 | +// GNU Lesser General Public License for more details. |
| 13 | +// |
| 14 | +// You should have received a copy of the GNU Lesser General Public License |
| 15 | +// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>. |
| 16 | + |
| 17 | +package snapshot |
| 18 | + |
| 19 | +import ( |
| 20 | + "bytes" |
| 21 | + "encoding/binary" |
| 22 | + "errors" |
| 23 | + "time" |
| 24 | + |
| 25 | + "github.com/ethereum/go-ethereum/common" |
| 26 | + "github.com/ethereum/go-ethereum/common/math" |
| 27 | + "github.com/ethereum/go-ethereum/core/rawdb" |
| 28 | + "github.com/ethereum/go-ethereum/ethdb" |
| 29 | + "github.com/ethereum/go-ethereum/ethdb/memorydb" |
| 30 | + "github.com/ethereum/go-ethereum/log" |
| 31 | +) |
| 32 | + |
| 33 | +const ( |
| 34 | + snapAccount = "account" // Identifier of account snapshot generation |
| 35 | + snapStorage = "storage" // Identifier of storage snapshot generation |
| 36 | +) |
| 37 | + |
| 38 | +// generatorStats is a collection of statistics gathered by the snapshot generator |
| 39 | +// for logging purposes. |
| 40 | +type generatorStats struct { |
| 41 | + origin uint64 // Origin prefix where generation started |
| 42 | + start time.Time // Timestamp when generation started |
| 43 | + accounts uint64 // Number of accounts indexed(generated or recovered) |
| 44 | + slots uint64 // Number of storage slots indexed(generated or recovered) |
| 45 | + dangling uint64 // Number of dangling storage slots |
| 46 | + storage common.StorageSize // Total account and storage slot size(generation or recovery) |
| 47 | +} |
| 48 | + |
| 49 | +// Log creates an contextual log with the given message and the context pulled |
| 50 | +// from the internally maintained statistics. |
| 51 | +func (gs *generatorStats) Log(msg string, root common.Hash, marker []byte) { |
| 52 | + var ctx []interface{} |
| 53 | + if root != (common.Hash{}) { |
| 54 | + ctx = append(ctx, []interface{}{"root", root}...) |
| 55 | + } |
| 56 | + // Figure out whether we're after or within an account |
| 57 | + switch len(marker) { |
| 58 | + case common.HashLength: |
| 59 | + ctx = append(ctx, []interface{}{"at", common.BytesToHash(marker)}...) |
| 60 | + case 2 * common.HashLength: |
| 61 | + ctx = append(ctx, []interface{}{ |
| 62 | + "in", common.BytesToHash(marker[:common.HashLength]), |
| 63 | + "at", common.BytesToHash(marker[common.HashLength:]), |
| 64 | + }...) |
| 65 | + } |
| 66 | + // Add the usual measurements |
| 67 | + ctx = append(ctx, []interface{}{ |
| 68 | + "accounts", gs.accounts, |
| 69 | + "slots", gs.slots, |
| 70 | + "storage", gs.storage, |
| 71 | + "dangling", gs.dangling, |
| 72 | + "elapsed", common.PrettyDuration(time.Since(gs.start)), |
| 73 | + }...) |
| 74 | + // Calculate the estimated indexing time based on current stats |
| 75 | + if len(marker) > 0 { |
| 76 | + if done := binary.BigEndian.Uint64(marker[:8]) - gs.origin; done > 0 { |
| 77 | + left := math.MaxUint64 - binary.BigEndian.Uint64(marker[:8]) |
| 78 | + |
| 79 | + speed := done/uint64(time.Since(gs.start)/time.Millisecond+1) + 1 // +1s to avoid division by zero |
| 80 | + ctx = append(ctx, []interface{}{ |
| 81 | + "eta", common.PrettyDuration(time.Duration(left/speed) * time.Millisecond), |
| 82 | + }...) |
| 83 | + } |
| 84 | + } |
| 85 | + log.Info(msg, ctx...) |
| 86 | +} |
| 87 | + |
| 88 | +// generatorContext carries a few global values to be shared by all generation functions. |
| 89 | +type generatorContext struct { |
| 90 | + stats *generatorStats // Generation statistic collection |
| 91 | + db ethdb.KeyValueStore // Key-value store containing the snapshot data |
| 92 | + account *holdableIterator // Iterator of account snapshot data |
| 93 | + storage *holdableIterator // Iterator of storage snapshot data |
| 94 | + batch ethdb.Batch // Database batch for writing batch data atomically |
| 95 | + logged time.Time // The timestamp when last generation progress was displayed |
| 96 | +} |
| 97 | + |
| 98 | +// newGeneratorContext initializes the context for generation. |
| 99 | +func newGeneratorContext(stats *generatorStats, db ethdb.KeyValueStore, accMarker []byte, storageMarker []byte) *generatorContext { |
| 100 | + ctx := &generatorContext{ |
| 101 | + stats: stats, |
| 102 | + db: db, |
| 103 | + batch: db.NewBatch(), |
| 104 | + logged: time.Now(), |
| 105 | + } |
| 106 | + ctx.openIterator(snapAccount, accMarker) |
| 107 | + ctx.openIterator(snapStorage, storageMarker) |
| 108 | + return ctx |
| 109 | +} |
| 110 | + |
| 111 | +// openIterator constructs global account and storage snapshot iterators |
| 112 | +// at the interrupted position. These iterators should be reopened from time |
| 113 | +// to time to avoid blocking leveldb compaction for a long time. |
| 114 | +func (ctx *generatorContext) openIterator(kind string, start []byte) { |
| 115 | + if kind == snapAccount { |
| 116 | + iter := ctx.db.NewIterator(rawdb.SnapshotAccountPrefix, start) |
| 117 | + ctx.account = newHoldableIterator(rawdb.NewKeyLengthIterator(iter, 1+common.HashLength)) |
| 118 | + return |
| 119 | + } |
| 120 | + iter := ctx.db.NewIterator(rawdb.SnapshotStoragePrefix, start) |
| 121 | + ctx.storage = newHoldableIterator(rawdb.NewKeyLengthIterator(iter, 1+2*common.HashLength)) |
| 122 | +} |
| 123 | + |
| 124 | +// reopenIterator releases the specified snapshot iterator and re-open it |
| 125 | +// in the next position. It's aimed for not blocking leveldb compaction. |
| 126 | +func (ctx *generatorContext) reopenIterator(kind string) { |
| 127 | + // Shift iterator one more step, so that we can reopen |
| 128 | + // the iterator at the right position. |
| 129 | + var iter = ctx.account |
| 130 | + if kind == snapStorage { |
| 131 | + iter = ctx.storage |
| 132 | + } |
| 133 | + hasNext := iter.Next() |
| 134 | + if !hasNext { |
| 135 | + // Iterator exhausted, release forever and create an already exhausted virtual iterator |
| 136 | + iter.Release() |
| 137 | + if kind == snapAccount { |
| 138 | + ctx.account = newHoldableIterator(memorydb.New().NewIterator(nil, nil)) |
| 139 | + return |
| 140 | + } |
| 141 | + ctx.storage = newHoldableIterator(memorydb.New().NewIterator(nil, nil)) |
| 142 | + return |
| 143 | + } |
| 144 | + next := iter.Key() |
| 145 | + iter.Release() |
| 146 | + ctx.openIterator(kind, next[1:]) |
| 147 | +} |
| 148 | + |
| 149 | +// close releases all the held resources. |
| 150 | +func (ctx *generatorContext) close() { |
| 151 | + ctx.account.Release() |
| 152 | + ctx.storage.Release() |
| 153 | +} |
| 154 | + |
| 155 | +// iterator returns the corresponding iterator specified by the kind. |
| 156 | +func (ctx *generatorContext) iterator(kind string) *holdableIterator { |
| 157 | + if kind == snapAccount { |
| 158 | + return ctx.account |
| 159 | + } |
| 160 | + return ctx.storage |
| 161 | +} |
| 162 | + |
| 163 | +// removeStorageBefore deletes all storage entries which are located before |
| 164 | +// the specified account. When the iterator touches the storage entry which |
| 165 | +// is located in or outside the given account, it stops and holds the current |
| 166 | +// iterated element locally. |
| 167 | +func (ctx *generatorContext) removeStorageBefore(account common.Hash) { |
| 168 | + var ( |
| 169 | + count uint64 |
| 170 | + start = time.Now() |
| 171 | + iter = ctx.storage |
| 172 | + ) |
| 173 | + for iter.Next() { |
| 174 | + key := iter.Key() |
| 175 | + if bytes.Compare(key[1:1+common.HashLength], account.Bytes()) >= 0 { |
| 176 | + iter.Hold() |
| 177 | + break |
| 178 | + } |
| 179 | + count++ |
| 180 | + ctx.batch.Delete(key) |
| 181 | + if ctx.batch.ValueSize() > ethdb.IdealBatchSize { |
| 182 | + ctx.batch.Write() |
| 183 | + ctx.batch.Reset() |
| 184 | + } |
| 185 | + } |
| 186 | + ctx.stats.dangling += count |
| 187 | + snapStorageCleanCounter.Inc(time.Since(start).Nanoseconds()) |
| 188 | +} |
| 189 | + |
| 190 | +// removeStorageAt deletes all storage entries which are located in the specified |
| 191 | +// account. When the iterator touches the storage entry which is outside the given |
| 192 | +// account, it stops and holds the current iterated element locally. An error will |
| 193 | +// be returned if the initial position of iterator is not in the given account. |
| 194 | +func (ctx *generatorContext) removeStorageAt(account common.Hash) error { |
| 195 | + var ( |
| 196 | + count int64 |
| 197 | + start = time.Now() |
| 198 | + iter = ctx.storage |
| 199 | + ) |
| 200 | + for iter.Next() { |
| 201 | + key := iter.Key() |
| 202 | + cmp := bytes.Compare(key[1:1+common.HashLength], account.Bytes()) |
| 203 | + if cmp < 0 { |
| 204 | + return errors.New("invalid iterator position") |
| 205 | + } |
| 206 | + if cmp > 0 { |
| 207 | + iter.Hold() |
| 208 | + break |
| 209 | + } |
| 210 | + count++ |
| 211 | + ctx.batch.Delete(key) |
| 212 | + if ctx.batch.ValueSize() > ethdb.IdealBatchSize { |
| 213 | + ctx.batch.Write() |
| 214 | + ctx.batch.Reset() |
| 215 | + } |
| 216 | + } |
| 217 | + snapWipedStorageMeter.Mark(count) |
| 218 | + snapStorageCleanCounter.Inc(time.Since(start).Nanoseconds()) |
| 219 | + return nil |
| 220 | +} |
| 221 | + |
| 222 | +// removeStorageLeft deletes all storage entries which are located after |
| 223 | +// the current iterator position. |
| 224 | +func (ctx *generatorContext) removeStorageLeft() { |
| 225 | + var ( |
| 226 | + count uint64 |
| 227 | + start = time.Now() |
| 228 | + iter = ctx.storage |
| 229 | + ) |
| 230 | + for iter.Next() { |
| 231 | + count++ |
| 232 | + ctx.batch.Delete(iter.Key()) |
| 233 | + if ctx.batch.ValueSize() > ethdb.IdealBatchSize { |
| 234 | + ctx.batch.Write() |
| 235 | + ctx.batch.Reset() |
| 236 | + } |
| 237 | + } |
| 238 | + ctx.stats.dangling += count |
| 239 | + snapDanglingStorageMeter.Mark(int64(count)) |
| 240 | + snapStorageCleanCounter.Inc(time.Since(start).Nanoseconds()) |
| 241 | +} |
0 commit comments