cmd, core, eth, graphql, trie: no persisted clean trie cache file (#27525)

The clean trie cache is persisted periodically, therefore Geth can
quickly warmup the cache in next restart.

However it will reduce the robustness of system. The assumption is
held in Geth that if the parent trie node is present, then the entire
sub-trie associated with the parent are all prensent.

Imagine the scenario that Geth rewinds itself to a past block and
restart, but Geth finds the root node of "future state" in clean
cache then regard this state is present in disk, while is not in fact.

Another example is offline pruning tool. Whenever an offline pruning
is performed, the clean cache file has to be removed to aviod hitting
the root node of "deleted states" in clean cache.

All in all, compare with the minor performance gain, system robustness
is something we care more.
This commit is contained in:
rjl493456442
2023-07-04 10:21:06 +03:00
committed by GitHub
parent 6ca3ef9a7b
commit 59f7b289c3
13 changed files with 59 additions and 186 deletions
+2 -35
View File
@@ -57,7 +57,6 @@ const (
// Config includes all the configurations for pruning.
type Config struct {
Datadir string // The directory of the state database
Cachedir string // The directory of state clean cache
BloomSize uint64 // The Megabytes of memory allocated to bloom-filter
}
@@ -241,7 +240,7 @@ func (p *Pruner) Prune(root common.Hash) error {
return err
}
if stateBloomRoot != (common.Hash{}) {
return RecoverPruning(p.config.Datadir, p.db, p.config.Cachedir)
return RecoverPruning(p.config.Datadir, p.db)
}
// If the target state root is not specified, use the HEAD-127 as the
// target. The reason for picking it is:
@@ -299,12 +298,6 @@ func (p *Pruner) Prune(root common.Hash) error {
log.Info("Selecting user-specified state as the pruning target", "root", root)
}
}
// Before start the pruning, delete the clean trie cache first.
// It's necessary otherwise in the next restart we will hit the
// deleted state root in the "clean cache" so that the incomplete
// state is picked for usage.
deleteCleanTrieCache(p.config.Cachedir)
// All the state roots of the middle layer should be forcibly pruned,
// otherwise the dangling state will be left.
middleRoots := make(map[common.Hash]struct{})
@@ -342,7 +335,7 @@ func (p *Pruner) Prune(root common.Hash) error {
// pruning can be resumed. What's more if the bloom filter is constructed, the
// pruning **has to be resumed**. Otherwise a lot of dangling nodes may be left
// in the disk.
func RecoverPruning(datadir string, db ethdb.Database, trieCachePath string) error {
func RecoverPruning(datadir string, db ethdb.Database) error {
stateBloomPath, stateBloomRoot, err := findBloomFilter(datadir)
if err != nil {
return err
@@ -378,12 +371,6 @@ func RecoverPruning(datadir string, db ethdb.Database, trieCachePath string) err
}
log.Info("Loaded state bloom filter", "path", stateBloomPath)
// Before start the pruning, delete the clean trie cache first.
// It's necessary otherwise in the next restart we will hit the
// deleted state root in the "clean cache" so that the incomplete
// state is picked for usage.
deleteCleanTrieCache(trieCachePath)
// All the state roots of the middle layers should be forcibly pruned,
// otherwise the dangling state will be left.
var (
@@ -497,23 +484,3 @@ func findBloomFilter(datadir string) (string, common.Hash, error) {
}
return stateBloomPath, stateBloomRoot, nil
}
const warningLog = `
WARNING!
The clean trie cache is not found. Please delete it by yourself after the
pruning. Remember don't start the Geth without deleting the clean trie cache
otherwise the entire database may be damaged!
Check the command description "geth snapshot prune-state --help" for more details.
`
func deleteCleanTrieCache(path string) {
if !common.FileExist(path) {
log.Warn(warningLog)
return
}
os.RemoveAll(path)
log.Info("Deleted trie clean cache", "path", path)
}