Merge pull request #4981 from konidev20/fix-gh-4934-cleanup-removed-snaphots-from-cache

cache: clear snapshot files from cache during load index

changelog/unreleased/issue-4934

@@ -0,0 +1,8 @@
+Enhancement: Clear removed snapshots from local cache of the current host
+
+Restic only removed snapshots from the cache on the host that runs the `forget` command.
+On other hosts that use the same repository, the old snapshots remained in the cache.
+Restic now, automatically clears old snapshots from the local cache of the current host.
+
+https://github.com/restic/restic/issues/4934
+https://github.com/restic/restic/pull/4981

internal/backend/cache/backend.go

@@ -2,11 +2,14 @@ package cache
 
 import (
 	"context"
+	"fmt"
 	"io"
+	"os"
 	"sync"
 
 	"github.com/restic/restic/internal/backend"
 	"github.com/restic/restic/internal/debug"
+	"github.com/restic/restic/internal/restic"
 )
 
 // Backend wraps a restic.Backend and adds a cache.
@@ -215,3 +218,44 @@ func (b *Backend) IsNotExist(err error) bool {
 func (b *Backend) Unwrap() backend.Backend {
 	return b.Backend
 }
+
+func (b *Backend) List(ctx context.Context, t backend.FileType, fn func(f backend.FileInfo) error) error {
+	if !b.Cache.canBeCached(t) {
+		return b.Backend.List(ctx, t, fn)
+	}
+
+	// will contain the IDs of the files that are in the repository
+	ids := restic.NewIDSet()
+
+	// wrap the original function to also add the file to the ids set
+	wrapFn := func(f backend.FileInfo) error {
+		id, err := restic.ParseID(f.Name)
+		if err != nil {
+			// returning error here since, if we cannot parse the ID, the file
+			// is invalid and the list must exit.
+			return err
+		}
+
+		ids.Insert(id)
+
+		// execute the original function
+		return fn(f)
+	}
+
+	err := b.Backend.List(ctx, t, wrapFn)
+	if err != nil {
+		return err
+	}
+
+	if ctx.Err() != nil {
+		return ctx.Err()
+	}
+
+	// clear the cache for files that are not in the repo anymore, ignore errors
+	err = b.Cache.Clear(t, ids)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "error clearing %s files in cache: %v\n", t.String(), err)
+	}
+
+	return nil
+}

internal/backend/cache/backend_test.go

@@ -57,6 +57,13 @@ func randomData(n int) (backend.Handle, []byte) {
 	return h, data
 }
 
+func list(t testing.TB, be backend.Backend, fn func(backend.FileInfo) error) {
+	err := be.List(context.TODO(), backend.IndexFile, fn)
+	if err != nil {
+		t.Fatal(err)
+	}
+}
+
 func TestBackend(t *testing.T) {
 	be := mem.New()
 	c := TestNewCache(t)
@@ -238,3 +245,54 @@ func TestErrorBackend(t *testing.T) {
 
 	wg.Wait()
 }
+
+func TestAutomaticCacheClear(t *testing.T) {
+	be := mem.New()
+	c := TestNewCache(t)
+	wbe := c.Wrap(be)
+
+	// add two handles h1 and h2
+	h1, data := randomData(2000)
+	// save h1 directly to the backend
+	save(t, be, h1, data)
+	if c.Has(h1) {
+		t.Errorf("cache has file1 too early")
+	}
+
+	h2, data2 := randomData(3000)
+
+	// save h2 directly to the backend
+	save(t, be, h2, data2)
+	if c.Has(h2) {
+		t.Errorf("cache has file2 too early")
+	}
+
+	loadAndCompare(t, wbe, h1, data)
+	if !c.Has(h1) {
+		t.Errorf("cache doesn't have file1 after load")
+	}
+
+	loadAndCompare(t, wbe, h2, data2)
+	if !c.Has(h2) {
+		t.Errorf("cache doesn't have file2 after load")
+	}
+
+	// remove h1 directly from the backend
+	remove(t, be, h1)
+	if !c.Has(h1) {
+		t.Errorf("file1 not in cache any more, should be removed from cache only after list")
+	}
+
+	// list all files in the backend
+	list(t, wbe, func(_ backend.FileInfo) error { return nil })
+
+	// h1 should be removed from the cache
+	if c.Has(h1) {
+		t.Errorf("cache has file1 after remove")
+	}
+
+	// h2 should still be in the cache
+	if !c.Has(h2) {
+		t.Errorf("cache doesn't have file2 after list")
+	}
+}

internal/repository/repository.go

@@ -706,19 +706,10 @@ func (r *Repository) prepareCache() error {
 		return nil
 	}
 
-	indexIDs := r.idx.IDs()
-	debug.Log("prepare cache with %d index files", len(indexIDs))
-
-	// clear old index files
-	err := r.Cache.Clear(restic.IndexFile, indexIDs)
-	if err != nil {
-		fmt.Fprintf(os.Stderr, "error clearing index files in cache: %v\n", err)
-	}
-
 	packs := r.idx.Packs(restic.NewIDSet())
 
 	// clear old packs
-	err = r.Cache.Clear(restic.PackFile, packs)
+	err := r.Cache.Clear(restic.PackFile, packs)
 	if err != nil {
 		fmt.Fprintf(os.Stderr, "error clearing pack files in cache: %v\n", err)
 	}