feat: Fast storage optimization for queries and iterations (cosmos#468)

* Add unbounded key string to KeyFormat * Add test vectors for unbounded length keys * Add some notes * update .gitignore * Add FastNode struct * WIP: make Get work with new FastNode * when retrieving fastnode fails, return errors vs. panic * add comments clarifying what index represents * make the linter happy * Add small tweaks to fast_node.go * add TODO & small linter tweaks * Update comment * update fast node cache in set * add debugging output when falling back to original logic * return error instead of panic * WIP: refactor set ops to work with fast store * update Set of mutable tree, begin unit testing * update GetVersioned to check fast nodes before trying the immutable * check fast node version before nil value check in get of immutable tree * fix small bugs and typos, continue writing unit tests for Set * unit test saveFastNodeVersion * simplify storing unsaved fast nodes * resolve a bug with not writing prefix for fast node to disk * remove fast nodes from disk on save and clear fast cache when version is deleted, fix all tests but random and with index * resolve an issue with randomized tests caused by the fast node cache not being cleared when latest version is saved * split unsaved fast node changes into additions and removals * save fast node removals * move fast node cache clearing to nodedb * use regular logic only when fast node version is greater than immutable tree' * clean up tree_random_test.go * clear unsaved fast node removals on rollback * fix randomized test failures caused by a typo in ndb DeleteVersion for loop * implement GetFast method to preserve Get with correct index return, convert unit tests from Get to GetFast where applicable * ensure Get and GetFast return the same values in tree_random_test.go * test fast node cache is live in random tree tests * improve mutable tree unit tests related to new functionality * clean up tree_test.go * implement GetVersionedFast to preserve the index in GetVersioned * restore accidentally deleted test in mutable tree test * spurious whitespace * refactor mutable tree * fix comment in mutable tree * add benchmark results * avoid redundant full tree search in GetFast of immutable tree when fast node is nil and tree is latest * fix naming for bench test get on random keys * use method for get latestversio in get fast * optimize GetFast, perform a refactor to ensure that fast nodes on disk are matched and better test * add latest bench * Fast Node Iteration (#7) * propagate errors from nodedb and avoid panicking * begin implementing fast node iteration * resolve rebase issue in random tests * fix iteration to deserialize fast node for extracting the correct value * finalzie iteration and let all unit tests pass * add benchmarks * merge GetVersioned and GetVersionedFast * remove fast node deletion from DeleteVersion and DeleteVersionRange and benchmark * fix and unit test iteration on mutable and immutable trees * implement tests for iterator and iterate, begin testing fast iterator * fix and unit test fast iterator * refactor iterate methods of mutable and immutable trees * resolve some warnings * remove old bench results * refactor bench tests for iteration * Fast Cache Migration (#9) * implement nodedb changes to set and get chain version from the database * implement and unit test upgrade to fast storage in mutable tree * refactor for auto upgrade to fast version in mutable tree contructor, load version and lazy load version * use proper functionality for getting latest version * remove unused error * fix comment * use fast version value in tests * spurious tab * fix style problems and remove redundant code in tests * Rename Get to GetWithIndex and GetFast to Get * refactor and clean up bench tests * remove extra byte from fast node length * clean up immutable tree * refactor nil tree or ndb error for the iterators and their tests * avoid exporting methods for getting unsaved additions and removals * refactor fast upgrade to read from tree instead of traversing disk nodes and orphans, update unit tests * remove unneeded comment * refer to storage version consistently across the project * fix more warnings * optimize removal of fast nodes from cache on deletion * small changes in teh mutable tree * correct storage version key * auto set fast version in SaveVersion * avoid exporting new methods of the immutable tree * go fmt * Fix comment in fast iterator Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * add extra comment for domain in fast iterator Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * add comments for moving the iterator before the first element * add comment for describing what mirror is in assertIterator of testutils * fix checking the mirror for descending iterator in tests * Update testutils_test.go with a comment Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * Update benchmarks/bench_test.go with a comment for runKnownQueriesFast Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * Update benchmarks/bench_test.go with a comment for runQueriesFast Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * export IsFastCacheEnabled and add an assert in bench tests * Update comment immutable_tree.go Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * Update comment for migration in mutable_tree.go Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> * simlify Iterate in mutable tree, add debug log for * Fast Cache - Downgrade - reupgrade protection and other improvements (#12) * add leaf hash to fast node and unit test * refactor get with index and get by index, fix migration in load version and lazy load version * use Get in GetVersioned of mutable tree * refactor non membership proof to use fast storage if available * bench non-membership proof * fix bench tests to work with the new changes * add downgrade-reupgrade protection and unit test * remove leaf hash from fast node * resolve multithreading bug related to iterators not being closed * clean up * use correct tree in bench tests * add cache to tree used to bench non membership proofs * add benc tests for GetWithIndex and GetByIndex * revert GetWithIndex and GetByIndex * remove unused import * unit test re-upgrade protection and fix small issues * remove redundant setStorageVersion method * fix bug with appending to live stage version to storage version and nit test * add comment for setFastStorageVersionToBatch * refactor and improve unit tests for reupgrade protection * rename ndb's isFastStorageEnabled to hasUpgradedToFastStorage and add comments * comment out new implementation for GetNonMembershipProof * update comments in nodedb to reflect the difference between hasUpgradedToFastStorage and shouldForceFastStorageUpdate * refactor nodedb tests * downgrade tendermint to 0.34.14 - osmosis's latest cosmos sdk does not support 0.35.0 * fix bug where fast storage was not enabled when version 0 was attempted to be loaded * implement unsaved fast iterator to be used in mutable tree (cosmos#16) * address comments from unsaved fast iterator PR * expose isUpgradeable method on mutable tree and unit test (cosmos#17) * expose isUpgradeable method on mutable tree and unit test * go fmt * resolve problems with rebasing * update CHANGELOG.md * tendermint v0.35.0 * fix String() bench * fix duplication lint in iterator_test.go * fix lint for tree.ndb.DeleteFastNode error return not checked * fix Error return value of `ndb.resetBatch` is not checked * fix Error return value of `ndb.traversePrefix` is not checked * fix Error: struct of size 64 bytes could be of size 56 bytes * fix Error: `comitted` is a misspelling of `committed` * fix issues in basic_test.go * address comments in fast_iterator.go * address comments in immutable tree * address comments in iterator.go * address comments in key_format.go * address remaining comments * fix Error: Error return value of `ndb.batch.Write` is not checked * fix Error: receiver name t should be consistent with previous receiver name tree for MutableTree * fix Error: struct of size 48 bytes could be of size 40 bytes * go fmt * more linter fixes * fix remaining linter problems * upgrade tm-db and comment out broken bencher databases * skip iterations for BenchmarkLevelDBLargeData bench * attempt to fix linter * force GC, no cache during migration, auto heap profile (cosmos#19) * force GC, no cache during migration, auto heap profile * resolve a potential deadlock from racing between reset and stop * fix small lint issue * remove logs and pprof logic * remove unused libraries * add comment explaining the reason for RAM optimizations * sync access to fast node cache to avoid concurrent write fatal error (cosmos#23) * update go.mod and go.sum to match master versions * revert cosmos#23 (sync access to fast node cache), fix bug related to old height export (cosmos#33) * Revert "sync access to fast node cache to avoid concurrent write fatal error (cosmos#23)" This reverts commit 2a1daf4. * return correct iterator in mutable tree * fix concurrent map panic when querying and comittting concurrently * avoid clearing fast node cache during pruning (cosmos#35) * fix data race related to VersionExists (cosmos#36) * fix data race related to VersionExists * use regular lock instead of RW in mutable_tree.go * hardcode fast node cache size to 100k * go fmt * restore proof_ics23.go * fix linter Co-authored-by: ValarDragon <dojha12@gmail.com> Co-authored-by: jtieri <37750742+jtieri@users.noreply.github.com> Co-authored-by: Dev Ojha <ValarDragon@users.noreply.github.com> Co-authored-by: Roman Akhtariev <34196718+akhtariev@users.noreply.github.com> Co-authored-by: Roman <34196718+r0mvn@users.noreply.github.com>
irisnet · Apr 9, 2022 · 0dcb21b · 0dcb21b
1 parent ccfb418
commit 0dcb21b
Show file tree

Hide file tree

Showing 34 changed files with 3,985 additions and 385 deletions.
diff --git a/.gitignore b/.gitignore
@@ -12,3 +12,6 @@ cpu*.out
 mem*.out
 cpu*.pdf
 mem*.pdf
+
+# IDE files
+.idea/*
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -9,6 +9,12 @@
 
 ### Improvements
 
+- [\#468](https://github.com/cosmos/iavl/pull/468) Fast storage optimization for queries and iterations
+
+## 0.17.3 (December 1, 2021)
+
+### Improvements
+
 - [\#445](https://github.com/cosmos/iavl/pull/445) Bump github.com/tendermint/tendermint to v0.35.0
 - [\#452](https://github.com/cosmos/iavl/pull/452) Optimization: remove unnecessary (*bytes.Buffer).Reset right after creating buffer.
 - [\#453](https://github.com/cosmos/iavl/pull/453),[\#456](https://github.com/cosmos/iavl/pull/456) Optimization: buffer reuse

diff --git a/basic_test.go b/basic_test.go
@@ -35,71 +35,126 @@ func TestBasic(t *testing.T) {
 
 	// Test 0x00
 	{
-		idx, val := tree.Get([]byte{0x00})
+		key := []byte{0x00}
+		expected := ""
+
+		idx, val := tree.GetWithIndex(key)
 		if val != nil {
-			t.Errorf("Expected no value to exist")
+			t.Error("Expected no value to exist")
 		}
 		if idx != 0 {
 			t.Errorf("Unexpected idx %x", idx)
 		}
-		if string(val) != "" {
-			t.Errorf("Unexpected value %v", string(val))
+		if string(val) != expected {
+			t.Errorf("Unexpected value %s", val)
+		}
+
+		val = tree.Get(key)
+		if val != nil {
+			t.Error("Fast method - expected no value to exist")
+		}
+		if string(val) != expected {
+			t.Errorf("Fast method - Unexpected value %s", val)
 		}
 	}
 
 	// Test "1"
 	{
-		idx, val := tree.Get([]byte("1"))
+		key := []byte("1")
+		expected := "one"
+
+		idx, val := tree.GetWithIndex(key)
 		if val == nil {
-			t.Errorf("Expected value to exist")
+			t.Error("Expected value to exist")
 		}
 		if idx != 0 {
 			t.Errorf("Unexpected idx %x", idx)
 		}
-		if string(val) != "one" {
-			t.Errorf("Unexpected value %v", string(val))
+		if string(val) != expected {
+			t.Errorf("Unexpected value %s", val)
+		}
+
+		val = tree.Get(key)
+		if val == nil {
+			t.Error("Fast method - expected value to exist")
+		}
+		if string(val) != expected {
+			t.Errorf("Fast method - Unexpected value %s", val)
 		}
 	}
 
 	// Test "2"
 	{
-		idx, val := tree.Get([]byte("2"))
+		key := []byte("2")
+		expected := "TWO"
+
+		idx, val := tree.GetWithIndex(key)
 		if val == nil {
-			t.Errorf("Expected value to exist")
+			t.Error("Expected value to exist")
 		}
 		if idx != 1 {
 			t.Errorf("Unexpected idx %x", idx)
 		}
-		if string(val) != "TWO" {
-			t.Errorf("Unexpected value %v", string(val))
+		if string(val) != expected {
+			t.Errorf("Unexpected value %s", val)
+		}
+
+		val = tree.Get(key)
+		if val == nil {
+			t.Error("Fast method - expected value to exist")
+		}
+		if string(val) != expected {
+			t.Errorf("Fast method - Unexpected value %s", val)
 		}
 	}
 
 	// Test "4"
 	{
-		idx, val := tree.Get([]byte("4"))
+		key := []byte("4")
+		expected := ""
+
+		idx, val := tree.GetWithIndex(key)
 		if val != nil {
-			t.Errorf("Expected no value to exist")
+			t.Error("Expected no value to exist")
 		}
 		if idx != 2 {
 			t.Errorf("Unexpected idx %x", idx)
 		}
-		if string(val) != "" {
-			t.Errorf("Unexpected value %v", string(val))
+		if string(val) != expected {
+			t.Errorf("Unexpected value %s", val)
+		}
+
+		val = tree.Get(key)
+		if val != nil {
+			t.Error("Fast method - expected no value to exist")
+		}
+		if string(val) != expected {
+			t.Errorf("Fast method - Unexpected value %s", val)
 		}
 	}
 
 	// Test "6"
 	{
-		idx, val := tree.Get([]byte("6"))
+		key := []byte("6")
+		expected := ""
+
+		idx, val := tree.GetWithIndex(key)
 		if val != nil {
-			t.Errorf("Expected no value to exist")
+			t.Error("Expected no value to exist")
 		}
 		if idx != 3 {
 			t.Errorf("Unexpected idx %x", idx)
 		}
-		if string(val) != "" {
-			t.Errorf("Unexpected value %v", string(val))
+		if string(val) != expected {
+			t.Errorf("Unexpected value %s", val)
+		}
+
+		val = tree.Get(key)
+		if val != nil {
+			t.Error("Fast method - expected no value to exist")
+		}
+		if string(val) != expected {
+			t.Errorf("Fast method - Unexpected value %s", val)
 		}
 	}
 }
@@ -252,7 +307,7 @@ func TestIntegration(t *testing.T) {
 		if has := tree.Has([]byte(randstr(12))); has {
 			t.Error("Table has extra key")
 		}
-		if _, val := tree.Get([]byte(r.key)); string(val) != r.value {
+		if val := tree.Get([]byte(r.key)); string(val) != r.value {
 			t.Error("wrong value")
 		}
 	}
@@ -270,7 +325,7 @@ func TestIntegration(t *testing.T) {
 			if has := tree.Has([]byte(randstr(12))); has {
 				t.Error("Table has extra key")
 			}
-			_, val := tree.Get([]byte(r.key))
+			val := tree.Get([]byte(r.key))
 			if string(val) != r.value {
 				t.Error("wrong value")
 			}
@@ -388,7 +443,7 @@ func TestPersistence(t *testing.T) {
 	require.NoError(t, err)
 	t2.Load()
 	for key, value := range records {
-		_, t2value := t2.Get([]byte(key))
+		t2value := t2.Get([]byte(key))
 		if string(t2value) != value {
 			t.Fatalf("Invalid value. Expected %v, got %v", value, t2value)
 		}

diff --git a/benchmarks/bench_test.go b/benchmarks/bench_test.go
@@ -19,7 +19,6 @@ func randBytes(length int) []byte {
 	key := make([]byte, length)
 	// math.rand.Read always returns err=nil
 	// we do not need cryptographic randomness for this test:
-	//nolint:gosec
 	rand.Read(key)
 	return key
 }
@@ -57,21 +56,95 @@ func commitTree(b *testing.B, t *iavl.MutableTree) {
 	}
 }
 
-func runQueries(b *testing.B, t *iavl.MutableTree, keyLen int) {
+// queries random keys against live state. Keys are almost certainly not in the tree.
+func runQueriesFast(b *testing.B, t *iavl.MutableTree, keyLen int) {
+	require.True(b, t.IsFastCacheEnabled())
 	for i := 0; i < b.N; i++ {
 		q := randBytes(keyLen)
 		t.Get(q)
 	}
 }
 
-func runKnownQueries(b *testing.B, t *iavl.MutableTree, keys [][]byte) {
+// queries keys that are known to be in state
+func runKnownQueriesFast(b *testing.B, t *iavl.MutableTree, keys [][]byte) {
+	require.True(b, t.IsFastCacheEnabled()) // to ensure fast storage is enabled
 	l := int32(len(keys))
 	for i := 0; i < b.N; i++ {
 		q := keys[rand.Int31n(l)]
 		t.Get(q)
 	}
 }
 
+func runQueriesSlow(b *testing.B, t *iavl.MutableTree, keyLen int) {
+	b.StopTimer()
+	// Save version to get an old immutable tree to query against,
+	// Fast storage is not enabled on old tree versions, allowing us to bench the desired behavior.
+	_, version, err := t.SaveVersion()
+	require.NoError(b, err)
+
+	itree, err := t.GetImmutable(version - 1)
+	require.NoError(b, err)
+	require.False(b, itree.IsFastCacheEnabled()) // to ensure fast storage is not enabled
+
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		q := randBytes(keyLen)
+		itree.GetWithIndex(q)
+	}
+}
+
+func runKnownQueriesSlow(b *testing.B, t *iavl.MutableTree, keys [][]byte) {
+	b.StopTimer()
+	// Save version to get an old immutable tree to query against,
+	// Fast storage is not enabled on old tree versions, allowing us to bench the desired behavior.
+	_, version, err := t.SaveVersion()
+	require.NoError(b, err)
+
+	itree, err := t.GetImmutable(version - 1)
+	require.NoError(b, err)
+	require.False(b, itree.IsFastCacheEnabled()) // to ensure fast storage is not enabled
+	b.StartTimer()
+	l := int32(len(keys))
+	for i := 0; i < b.N; i++ {
+		q := keys[rand.Int31n(l)]
+		index, value := itree.GetWithIndex(q)
+		require.True(b, index >= 0, "the index must not be negative")
+		require.NotNil(b, value, "the value should exist")
+	}
+}
+
+func runIterationFast(b *testing.B, t *iavl.MutableTree, expectedSize int) {
+	require.True(b, t.IsFastCacheEnabled()) // to ensure fast storage is enabled
+	for i := 0; i < b.N; i++ {
+		itr := t.ImmutableTree.Iterator(nil, nil, false)
+		iterate(b, itr, expectedSize)
+		require.Nil(b, itr.Close(), ".Close should not error out")
+	}
+}
+
+func runIterationSlow(b *testing.B, t *iavl.MutableTree, expectedSize int) {
+	for i := 0; i < b.N; i++ {
+		itr := iavl.NewIterator(nil, nil, false, t.ImmutableTree) // create slow iterator directly
+		iterate(b, itr, expectedSize)
+		require.Nil(b, itr.Close(), ".Close should not error out")
+	}
+}
+
+func iterate(b *testing.B, itr db.Iterator, expectedSize int) {
+	b.StartTimer()
+	keyValuePairs := make([][][]byte, 0, expectedSize)
+	for i := 0; i < expectedSize && itr.Valid(); i++ {
+		itr.Next()
+		keyValuePairs = append(keyValuePairs, [][]byte{itr.Key(), itr.Value()})
+	}
+	b.StopTimer()
+	if g, w := len(keyValuePairs), expectedSize; g != w {
+		b.Errorf("iteration count mismatch: got=%d, want=%d", g, w)
+	} else {
+		b.Logf("completed %d iterations", len(keyValuePairs))
+	}
+}
+
 // func runInsert(b *testing.B, t *iavl.MutableTree, keyLen, dataLen, blockSize int) *iavl.MutableTree {
 // 	for i := 1; i <= b.N; i++ {
 // 		t.Set(randBytes(keyLen), randBytes(dataLen))
@@ -132,7 +205,7 @@ func runBlock(b *testing.B, t *iavl.MutableTree, keyLen, dataLen, blockSize int,
 			data := randBytes(dataLen)
 
 			// perform query and write on check and then real
-			// check.Get(key)
+			// check.GetFast(key)
 			// check.Set(key, data)
 			real.Get(key)
 			real.Set(key, data)
@@ -175,11 +248,11 @@ func BenchmarkMedium(b *testing.B) {
 	benchmarks := []benchmark{
 		{"memdb", 100000, 100, 16, 40},
 		{"goleveldb", 100000, 100, 16, 40},
-		{"cleveldb", 100000, 100, 16, 40},
+		// {"cleveldb", 100000, 100, 16, 40},
 		// FIXME: idk why boltdb is too slow !?
 		// {"boltdb", 100000, 100, 16, 40},
-		{"rocksdb", 100000, 100, 16, 40},
-		{"badgerdb", 100000, 100, 16, 40},
+		// {"rocksdb", 100000, 100, 16, 40},
+		// {"badgerdb", 100000, 100, 16, 40},
 	}
 	runBenchmarks(b, benchmarks)
 }
@@ -188,10 +261,10 @@ func BenchmarkSmall(b *testing.B) {
 	benchmarks := []benchmark{
 		{"memdb", 1000, 100, 4, 10},
 		{"goleveldb", 1000, 100, 4, 10},
-		{"cleveldb", 1000, 100, 4, 10},
-		{"boltdb", 1000, 100, 4, 10},
-		{"rocksdb", 1000, 100, 4, 10},
-		{"badgerdb", 1000, 100, 4, 10},
+		// {"cleveldb", 1000, 100, 4, 10},
+		// {"boltdb", 1000, 100, 4, 10},
+		// {"rocksdb", 1000, 100, 4, 10},
+		// {"badgerdb", 1000, 100, 4, 10},
 	}
 	runBenchmarks(b, benchmarks)
 }
@@ -202,8 +275,8 @@ func BenchmarkLarge(b *testing.B) {
 		{"goleveldb", 1000000, 100, 16, 40},
 		// FIXME: idk why boltdb is too slow !?
 		// {"boltdb", 1000000, 100, 16, 40},
-		{"rocksdb", 1000000, 100, 16, 40},
-		{"badgerdb", 1000000, 100, 16, 40},
+		// {"rocksdb", 1000000, 100, 16, 40},
+		// {"badgerdb", 1000000, 100, 16, 40},
 	}
 	runBenchmarks(b, benchmarks)
 }
@@ -287,14 +360,38 @@ func runSuite(b *testing.B, d db.DB, initSize, blockSize, keyLen, dataLen int) {
 
 	b.ResetTimer()
 
-	b.Run("query-miss", func(sub *testing.B) {
+	b.Run("query-no-in-tree-guarantee-fast", func(sub *testing.B) {
+		sub.ReportAllocs()
+		runQueriesFast(sub, t, keyLen)
+	})
+	b.Run("query-no-in-tree-guarantee-slow", func(sub *testing.B) {
+		sub.ReportAllocs()
+		runQueriesSlow(sub, t, keyLen)
+	})
+	//
+	b.Run("query-hits-fast", func(sub *testing.B) {
 		sub.ReportAllocs()
-		runQueries(sub, t, keyLen)
+		runKnownQueriesFast(sub, t, keys)
 	})
-	b.Run("query-hits", func(sub *testing.B) {
+	b.Run("query-hits-slow", func(sub *testing.B) {
 		sub.ReportAllocs()
-		runKnownQueries(sub, t, keys)
+		runKnownQueriesSlow(sub, t, keys)
 	})
+	//
+	// Iterations for BenchmarkLevelDBLargeData timeout bencher in CI so
+	// we must skip them.
+	if b.Name() != "BenchmarkLevelDBLargeData" {
+		b.Run("iteration-fast", func(sub *testing.B) {
+			sub.ReportAllocs()
+			runIterationFast(sub, t, initSize)
+		})
+		b.Run("iteration-slow", func(sub *testing.B) {
+			sub.ReportAllocs()
+			runIterationSlow(sub, t, initSize)
+		})
+	}
+
+	//
 	b.Run("update", func(sub *testing.B) {
 		sub.ReportAllocs()
 		t = runUpdate(sub, t, dataLen, blockSize, keys)