feat(caching): implement negative stat cache to optimize polling of missing files

docs/semantics.md

@@ -209,7 +209,10 @@ The behavior of stat cache is controlled by the following flags/config parameter
    If this config variable is missing, then the value of ```--stat-cache-ttl``` is used.
    * ```--stat-cache-ttl``` commandline flag, which can be set to a value like ```10s``` or ```1.5h```. The default is one minute. This has been deprecated (starting v2.0) and is currently only available for backward compatibility. If ```metadata-cache: ttl-secs``` is set, ```--stat-cache-ttl``` is ignored.
 
-   Positive and negative stat results will be cached for the specified amount of time.
+   Positive stat results (existing files) will be cached for this specified amount of time.
+
+3. **Negative Stat-cache (Non-existent Entry Caching)**: Controls caching of non-existent paths (`ErrNotExist` / 404 results) to optimize workloads that aggressively poll missing files (e.g., JupyterLab).
+   * `metadata-cache: negative-ttl-secs` in the config-file (integer). Sets the TTL in seconds for negative entries. Default is 5 seconds. Use -1 for infinite TTL. Setting this to 0 disables negative caching.
 
 Warnings: 
 - Using stat caching breaks the consistency guarantees discussed in this document. It is safe only in the following situations:

internal/fs/gcs_metrics_test.go

@@ -20,11 +20,16 @@ import (
 	"os"
 	"testing"
 
+	"time"
+
 	"github.com/googlecloudplatform/gcsfuse/v3/cfg"
+	"github.com/googlecloudplatform/gcsfuse/v3/internal/cache/lru"
+	"github.com/googlecloudplatform/gcsfuse/v3/internal/cache/metadata"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/fs"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/fs/wrappers"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/gcsx"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/monitor"
+	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/caching"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/fake"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/gcs"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/storageutil"
@@ -463,3 +468,82 @@ func TestGCSMetrics_RetryCount(t *testing.T) {
 		attribute.NewSet(attribute.String("retry_error_category", "STALLED_READ_REQUEST")),
 		1)
 }
+
+type fakeBucketManagerForShortCircuit struct {
+	bucket gcs.Bucket
+}
+
+func (bm *fakeBucketManagerForShortCircuit) SetUpBucket(ctx context.Context, name string, _ bool, _ metrics.MetricHandle) (gcsx.SyncerBucket, error) {
+	return gcsx.NewSyncerBucket(0, 120, 10, ".gcsfuse_tmp/", gcsx.NewContentTypeBucket(bm.bucket)), nil
+}
+func (bm *fakeBucketManagerForShortCircuit) ShutDown() {}
+
+// TestGCSMetrics_RequestCount_NegativeCachingShortCircuit validates that when negative entry caching is enabled,
+// repeated lookups for non-existent files short-circuit in memory and do not emit redundant backend GCS requests.
+func TestGCSMetrics_RequestCount_NegativeCachingShortCircuit(t *testing.T) {
+	ctx := context.Background()
+	origProvider := otel.GetMeterProvider()
+	t.Cleanup(func() { otel.SetMeterProvider(origProvider) })
+	reader := metric.NewManualReader()
+	provider := metric.NewMeterProvider(metric.WithReader(reader))
+	otel.SetMeterProvider(provider)
+
+	mh, err := metrics.NewOTelMetrics(ctx, 1, 100)
+	require.NoError(t, err, "metrics.NewOTelMetrics")
+
+	bucketName := "test-bucket"
+	rawBucket := fake.NewFakeBucket(timeutil.RealClock(), bucketName, gcs.BucketType{Hierarchical: false})
+
+	// Enforce production wrapping order: Caching layer MUST wrap Monitoring layer.
+	monitoringInnerBucket := monitor.NewMonitoringBucket(rawBucket, mh)
+	lruCache := lru.NewCache(1024 * 1024)
+	statCacheView := metadata.NewStatCacheBucketView(lruCache, "")
+	cachedOuterBucket := caching.NewFastStatBucket(
+		time.Minute,
+		statCacheView,
+		timeutil.RealClock(),
+		monitoringInnerBucket,
+		time.Minute,
+		true, // isTypeCacheDeprecated
+		true, // implicitDir
+	)
+
+	serverCfg := &fs.ServerConfig{
+		NewConfig: &cfg.Config{
+			EnableNewReader: true,
+		},
+		MetricHandle: mh,
+		TraceHandle:  tracing.NewNoopTracer(),
+		CacheClock:   &timeutil.SimulatedClock{},
+		BucketName:   bucketName,
+		BucketManager: &fakeBucketManagerForShortCircuit{
+			bucket: cachedOuterBucket,
+		},
+	}
+
+	server, err := fs.NewFileSystem(ctx, serverCfg)
+	require.NoError(t, err, "NewFileSystem")
+
+	lookupOp := &fuseops.LookUpInodeOp{
+		Parent: fuseops.RootInodeID,
+		Name:   "missing_file",
+	}
+
+	// 1. First Lookup (Cache Miss) -> Emits backend probes through monitoring layer
+	_ = server.LookUpInode(ctx, lookupOp)
+	waitForMetricsProcessing()
+
+	// Probing a missing file checks dir then file -> 2 actual backend network calls.
+	metrics.VerifyCounterMetric(t, ctx, reader, "gcs/request_count",
+		attribute.NewSet(attribute.String("gcs_method", "StatObject")),
+		2)
+
+	// 2. Second Lookup (Negative Cache Hit) -> Intercepted by outer cache layer
+	_ = server.LookUpInode(ctx, lookupOp)
+	waitForMetricsProcessing()
+
+	// Verify backend request count remains exactly 2 (zero new network requests emitted)
+	metrics.VerifyCounterMetric(t, ctx, reader, "gcs/request_count",
+		attribute.NewSet(attribute.String("gcs_method", "StatObject")),
+		2)
+}

internal/fs/hns_bucket_test.go

@@ -172,7 +172,8 @@ func (t *HNSCachedBucketMountTest) SetupSuite() {
 		uncachedHNSBucket,
 		negativeCacheTTL,
 		IsTypeCacheDeprecated,
-		isImplicitDir)
+		isImplicitDir,
+	)
 
 	// Enable directory type caching.
 	t.serverCfg.DirTypeCacheTTL = ttl

internal/fs/inode/dir.go

@@ -686,6 +686,7 @@ func (d *dirInode) LookUpChild(ctx context.Context, name string) (*Core, error)
 		// 1. Try Directory FIRST (since it's the preferred return type)
 		var dirResult *Core
 		var err error
+		var dirCacheMiss bool
 		if d.Bucket().BucketType().Hierarchical {
 			dirResult, err = findExplicitFolder(ctx, d.Bucket(), NewDirName(d.Name(), name), true)
 		} else {
@@ -696,19 +697,30 @@ func (d *dirInode) LookUpChild(ctx context.Context, name string) (*Core, error)
 		if dirResult != nil {
 			return dirResult, nil
 		}
-		// If we hit a real error (not a cache miss), exit early.
-		if err != nil && !errors.As(err, &cacheMissErr) {
+		if errors.As(err, &cacheMissErr) {
+			dirCacheMiss = true
+		} else if err != nil {
 			return nil, err
 		}
 
 		// 2. Try File ONLY if directory wasn't found
+		var fileCacheMiss bool
 		fileResult, err := findExplicitInode(ctx, d.Bucket(), NewFileName(d.Name(), name), true)
-		if err != nil && !errors.As(err, &cacheMissErr) {
+		if fileResult != nil {
+			return fileResult, nil
+		}
+		if errors.As(err, &cacheMissErr) {
+			fileCacheMiss = true
+		} else if err != nil {
 			return nil, err
 		}
 
-		if fileResult != nil {
-			return fileResult, nil
+		// 3. Short-circuit ONLY on confirmed negative hits (tombstones) for BOTH
+		// directory and file paths. If either lookup resulted in a cache miss,
+		// we cannot short-circuit and must fall back to GCS (fetchCoreEntity)
+		// because the un-cached type might still exist.
+		if !dirCacheMiss && !fileCacheMiss {
+			return nil, nil // Definitive ENOENT
 		}
 	}
 

internal/gcsx/bucket_manager.go

@@ -45,9 +45,9 @@ type BucketConfig struct {
 	// Config for TTL of entries for existing file in stat cache
 	StatCacheTTL time.Duration
 	// Config for TTL of entries for non-existing file in stat cache
-	NegativeStatCacheTTL time.Duration
-	EnableMonitoring     bool
-	LogSeverity          cfg.LogSeverity
+	NegativeStatCacheTTL          time.Duration
+	EnableMonitoring              bool
+	LogSeverity                   cfg.LogSeverity
 
 	// Files backed by on object of length at least AppendThreshold that have
 	// only been appended to (i.e. none of the object's contents have been

internal/storage/caching/fast_stat_bucket.go

@@ -124,6 +124,19 @@ func (b *fastStatBucket) insertListing(ctx context.Context, listing *gcs.Listing
 		return
 	}
 
+	// If the directory listing returned from GCS is empty (both MinObjects and CollapsedRuns
+	// are empty), it means the directory is either an implicit directory that no longer
+	// has any children (because they were deleted) or a non-existent directory.
+	// In either case, if negative caching is enabled (negativeCacheTTL > 0), we cache
+	// this directory name with a trailing slash as a negative entry to short-circuit
+	// subsequent lookups.
+	// Note: Explicit empty directories have a placeholder object (e.g., "dir/"),
+	// which will be returned in MinObjects, so they will not trigger this negative caching.
+	if len(listing.MinObjects) == 0 && len(listing.CollapsedRuns) == 0 && b.negativeCacheTTL > 0 && strings.HasSuffix(dirName, "/") {
+		b.cache.AddNegativeEntry(dirName, b.clock.Now().Add(b.negativeCacheTTL))
+		return
+	}
+
 	expiration := b.clock.Now().Add(b.primaryCacheTTL)
 
 	// 1. Parent Directory Inference (Implicit Check)
@@ -244,6 +257,11 @@ func (b *fastStatBucket) addNegativeEntry(name string) {
 	b.mu.Lock()
 	defer b.mu.Unlock()
 
+	if b.negativeCacheTTL <= 0 {
+		b.cache.Erase(name)
+		return
+	}
+
 	expiration := b.clock.Now().Add(b.negativeCacheTTL)
 	b.cache.AddNegativeEntry(name, expiration)
 }
@@ -253,6 +271,11 @@ func (b *fastStatBucket) addNegativeEntryForFolder(name string) {
 	b.mu.Lock()
 	defer b.mu.Unlock()
 
+	if b.negativeCacheTTL <= 0 {
+		b.cache.Erase(name)
+		return
+	}
+
 	expiration := b.clock.Now().Add(b.negativeCacheTTL)
 	b.cache.AddNegativeEntryForFolder(name, expiration)
 }
@@ -271,6 +294,9 @@ func (b *fastStatBucket) lookUp(name string) (hit bool, m *gcs.MinObject) {
 	defer b.mu.Unlock()
 
 	hit, m = b.cache.LookUp(name, b.clock.Now())
+	if hit && m == nil && b.negativeCacheTTL <= 0 {
+		return false, nil
+	}
 	return
 }
 
@@ -279,6 +305,9 @@ func (b *fastStatBucket) lookUpFolder(name string) (bool, *gcs.Folder) {
 	defer b.mu.Unlock()
 
 	hit, f := b.cache.LookUpFolder(name, b.clock.Now())
+	if hit && f == nil && b.negativeCacheTTL <= 0 {
+		return false, nil
+	}
 	return hit, f
 }
 

internal/storage/caching/fast_stat_bucket_test.go

@@ -645,6 +645,15 @@ func (t *StatObjectTest) CacheHit_Negative() {
 	ExpectThat(err, HasSameTypeAs(&gcs.NotFoundError{}))
 }
 
+func (t *StatObjectTest) CacheHit_Negative_Disabled_FetchOnly() {
+	t.bucket = caching.NewFastStatBucket(primaryCacheTTL, t.cache, &t.clock, t.wrapped, 0, true, true)
+	const name = "taco"
+	ExpectCall(t.cache, "LookUp")(Any(), Any()).WillOnce(Return(true, nil))
+	req := &gcs.StatObjectRequest{Name: name, FetchOnlyFromCache: true}
+	_, _, err := t.bucket.StatObject(context.Background(), req) //nolint:govet
+	ExpectThat(err, HasSameTypeAs(&caching.CacheMissError{}))
+}
+
 func (t *StatObjectTest) IgnoresCacheEntryWhenForceFetchFromGcsIsTrue() {
 	const name = "taco"
 
@@ -790,10 +799,21 @@ func (t *StatObjectTest) WrappedSaysNotFound() {
 	}
 
 	_, _, err := t.bucket.StatObject(context.TODO(), req)
-	ExpectThat(err, HasSameTypeAs(&gcs.NotFoundError{}))
 	ExpectThat(err, Error(HasSubstr("burrito")))
 }
 
+func (t *StatObjectTest) WrappedSaysNotFound_NegativeCachingDisabled() {
+	t.bucket = caching.NewFastStatBucket(primaryCacheTTL, t.cache, &t.clock, t.wrapped, 0, true, true)
+	const name = "taco"
+	ExpectCall(t.cache, "LookUp")(Any(), Any()).WillOnce(Return(false, nil))
+	ExpectCall(t.wrapped, "StatObject")(Any(), Any()).WillOnce(Return(nil, nil, &gcs.NotFoundError{Err: errors.New("burrito")}))
+	// Expect Erase call to invalidate any stale entry
+	ExpectCall(t.cache, "Erase")(name)
+	req := &gcs.StatObjectRequest{Name: name}
+	_, _, err := t.bucket.StatObject(context.Background(), req) //nolint:govet
+	ExpectThat(err, HasSameTypeAs(&gcs.NotFoundError{}))
+}
+
 func (t *StatObjectTest) WrappedSucceeds() {
 	const name = "taco"
 
@@ -997,7 +1017,7 @@ func (t *ListObjectsTest_InsertListing) SetUp(ti *TestInfo) {
 		t.cache,
 		&t.clock,
 		t.wrapped,
-		negativeCacheTTL,
+		0,
 		true,
 		true)
 }
@@ -1031,6 +1051,15 @@ func (t *ListObjectsTest_InsertListing) EmptyListing() {
 	t.callAndVerify(context.TODO(), false, listing, "dir/", expectedInserts, expectedImplicitDirs)
 }
 
+func (t *ListObjectsTest_InsertListing) EmptyListing_NegativeCaching() {
+	t.bucket = caching.NewFastStatBucket(primaryCacheTTL, t.cache, &t.clock, t.wrapped, negativeCacheTTL, true, true)
+	listing := &gcs.Listing{}
+	ExpectCall(t.wrapped, "BucketType")().WillOnce(Return(gcs.BucketType{}))
+	ExpectCall(t.wrapped, "ListObjects")(Any(), Any()).WillOnce(Return(listing, nil))
+	ExpectCall(t.cache, "AddNegativeEntry")("dir/", Any())
+	_, _ = t.bucket.ListObjects(context.Background(), &gcs.ListObjectsRequest{Prefix: "dir/"}) //nolint:govet
+}
+
 func (t *ListObjectsTest_InsertListing) ObjectsOnly() {
 	listing := &gcs.Listing{
 		MinObjects: []*gcs.MinObject{
@@ -1134,7 +1163,7 @@ func (t *ListObjectsTest_InsertListing) ImplicitDirFalse_CollapsedRunsNotCached(
 		t.cache,
 		&t.clock,
 		t.wrapped,
-		negativeCacheTTL,
+		0,
 		true,
 		false)
 	listing := &gcs.Listing{

internal/storage/fake/bucket.go

@@ -30,6 +30,7 @@ import (
 	"time"
 	"unicode/utf8"
 
+	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/caching"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/gcs"
 	"github.com/googlecloudplatform/gcsfuse/v3/internal/storage/storageutil"
 	"github.com/jacobsa/syncutil"
@@ -900,6 +901,9 @@ func (b *bucket) ComposeObjects(
 // LOCKS_EXCLUDED(b.mu)
 func (b *bucket) StatObject(ctx context.Context,
 	req *gcs.StatObjectRequest) (m *gcs.MinObject, e *gcs.ExtendedObjectAttributes, err error) {
+	if req.FetchOnlyFromCache {
+		return nil, nil, &caching.CacheMissError{Err: errors.New("fake bucket has no cache")}
+	}
 	// If ExtendedObjectAttributes are requested without fetching from gcs enabled, panic.
 	if !req.ForceFetchFromGcs && req.ReturnExtendedObjectAttributes {
 		panic("invalid StatObjectRequest: ForceFetchFromGcs: false and ReturnExtendedObjectAttributes: true")
@@ -1143,6 +1147,9 @@ func (b *bucket) DeleteFolder(ctx context.Context, folderName string) (err error
 }
 
 func (b *bucket) GetFolder(ctx context.Context, req *gcs.GetFolderRequest) (*gcs.Folder, error) {
+	if req.FetchOnlyFromCache {
+		return nil, &caching.CacheMissError{Err: errors.New("fake bucket has no cache")}
+	}
 	b.mu.Lock()
 	defer b.mu.Unlock()