core, trie: intermediate mempool between trie and database (#15857)

This commit reduces database I/O by not writing every state trie to disk.

cmd/utils/cmd.go

@@ -116,7 +116,6 @@ func ImportChain(chain *core.BlockChain, fn string) error {
 			return err
 		}
 	}
-
 	stream := rlp.NewStream(reader, 0)
 
 	// Run actual the import.
@@ -150,25 +149,34 @@ func ImportChain(chain *core.BlockChain, fn string) error {
 		if checkInterrupt() {
 			return fmt.Errorf("interrupted")
 		}
-		if hasAllBlocks(chain, blocks[:i]) {
+		missing := missingBlocks(chain, blocks[:i])
+		if len(missing) == 0 {
 			log.Info("Skipping batch as all blocks present", "batch", batch, "first", blocks[0].Hash(), "last", blocks[i-1].Hash())
 			continue
 		}
-
-		if _, err := chain.InsertChain(blocks[:i]); err != nil {
+		if _, err := chain.InsertChain(missing); err != nil {
 			return fmt.Errorf("invalid block %d: %v", n, err)
 		}
 	}
 	return nil
 }
 
-func hasAllBlocks(chain *core.BlockChain, bs []*types.Block) bool {
-	for _, b := range bs {
-		if !chain.HasBlock(b.Hash(), b.NumberU64()) {
-			return false
+func missingBlocks(chain *core.BlockChain, blocks []*types.Block) []*types.Block {
+	head := chain.CurrentBlock()
+	for i, block := range blocks {
+		// If we're behind the chain head, only check block, state is available at head
+		if head.NumberU64() > block.NumberU64() {
+			if !chain.HasBlock(block.Hash(), block.NumberU64()) {
+				return blocks[i:]
+			}
+			continue
+		}
+		// If we're above the chain head, state availability is a must
+		if !chain.HasBlockAndState(block.Hash(), block.NumberU64()) {
+			return blocks[i:]
 		}
 	}
-	return true
+	return nil
 }
 
 func ExportChain(blockchain *core.BlockChain, fn string) error {

cmd/utils/flags.go

@@ -170,7 +170,11 @@ var (
 		Usage: `Blockchain sync mode ("fast", "full", or "light")`,
 		Value: &defaultSyncMode,
 	}
-
+	GCModeFlag = cli.StringFlag{
+		Name:  "gcmode",
+		Usage: `Blockchain garbage collection mode ("full", "archive")`,
+		Value: "full",
+	}
 	LightServFlag = cli.IntFlag{
 		Name:  "lightserv",
 		Usage: "Maximum percentage of time allowed for serving LES requests (0-90)",
@@ -293,8 +297,18 @@ var (
 	// Performance tuning settings
 	CacheFlag = cli.IntFlag{
 		Name:  "cache",
-		Usage: "Megabytes of memory allocated to internal caching (min 16MB / database forced)",
-		Value: 128,
+		Usage: "Megabytes of memory allocated to internal caching",
+		Value: 1024,
+	}
+	CacheDatabaseFlag = cli.IntFlag{
+		Name:  "cache.database",
+		Usage: "Percentage of cache memory allowance to use for database io",
+		Value: 75,
+	}
+	CacheGCFlag = cli.IntFlag{
+		Name:  "cache.gc",
+		Usage: "Percentage of cache memory allowance to use for trie pruning",
+		Value: 25,
 	}
 	TrieCacheGenFlag = cli.IntFlag{
 		Name:  "trie-cache-gens",
@@ -1021,11 +1035,19 @@ func SetEthConfig(ctx *cli.Context, stack *node.Node, cfg *eth.Config) {
 		cfg.NetworkId = ctx.GlobalUint64(NetworkIdFlag.Name)
 	}
 
-	if ctx.GlobalIsSet(CacheFlag.Name) {
-		cfg.DatabaseCache = ctx.GlobalInt(CacheFlag.Name)
+	if ctx.GlobalIsSet(CacheFlag.Name) || ctx.GlobalIsSet(CacheDatabaseFlag.Name) {
+		cfg.DatabaseCache = ctx.GlobalInt(CacheFlag.Name) * ctx.GlobalInt(CacheDatabaseFlag.Name) / 100
 	}
 	cfg.DatabaseHandles = makeDatabaseHandles()
 
+	if gcmode := ctx.GlobalString(GCModeFlag.Name); gcmode != "full" && gcmode != "archive" {
+		Fatalf("--%s must be either 'full' or 'archive'", GCModeFlag.Name)
+	}
+	cfg.NoPruning = ctx.GlobalString(GCModeFlag.Name) == "archive"
+
+	if ctx.GlobalIsSet(CacheFlag.Name) || ctx.GlobalIsSet(CacheGCFlag.Name) {
+		cfg.TrieCache = ctx.GlobalInt(CacheFlag.Name) * ctx.GlobalInt(CacheGCFlag.Name) / 100
+	}
 	if ctx.GlobalIsSet(MinerThreadsFlag.Name) {
 		cfg.MinerThreads = ctx.GlobalInt(MinerThreadsFlag.Name)
 	}
@@ -1157,7 +1179,7 @@ func SetupNetwork(ctx *cli.Context) {
 // MakeChainDatabase open an LevelDB using the flags passed to the client and will hard crash if it fails.
 func MakeChainDatabase(ctx *cli.Context, stack *node.Node) ethdb.Database {
 	var (
-		cache   = ctx.GlobalInt(CacheFlag.Name)
+		cache   = ctx.GlobalInt(CacheFlag.Name) * ctx.GlobalInt(CacheDatabaseFlag.Name) / 100
 		handles = makeDatabaseHandles()
 	)
 	name := "chaindata"
@@ -1209,8 +1231,19 @@ func MakeChain(ctx *cli.Context, stack *node.Node) (chain *core.BlockChain, chai
 			})
 		}
 	}
+	if gcmode := ctx.GlobalString(GCModeFlag.Name); gcmode != "full" && gcmode != "archive" {
+		Fatalf("--%s must be either 'full' or 'archive'", GCModeFlag.Name)
+	}
+	cache := &core.CacheConfig{
+		Disabled:      ctx.GlobalString(GCModeFlag.Name) == "archive",
+		TrieNodeLimit: eth.DefaultConfig.TrieCache,
+		TrieTimeLimit: eth.DefaultConfig.TrieTimeout,
+	}
+	if ctx.GlobalIsSet(CacheFlag.Name) || ctx.GlobalIsSet(CacheGCFlag.Name) {
+		cache.TrieNodeLimit = ctx.GlobalInt(CacheFlag.Name) * ctx.GlobalInt(CacheGCFlag.Name) / 100
+	}
 	vmcfg := vm.Config{EnablePreimageRecording: ctx.GlobalBool(VMEnableDebugFlag.Name)}
-	chain, err = core.NewBlockChain(chainDb, config, engine, vmcfg)
+	chain, err = core.NewBlockChain(chainDb, cache, config, engine, vmcfg)
 	if err != nil {
 		Fatalf("Can't create BlockChain: %v", err)
 	}