housekeeping: Clean up server info datapks-git-prune-server-info

Even though we don't need it, git-repack(1) has been generating server
info via git-update-server-info(1) used for dumb HTTP protocols by
default in the past. The result is that we have lots of those files on
disk, which take away precious disk space without any merit. In some
cases, we have even accumulated many of those files in a single repo
because git-repack(1) didn't correctly clean those files if it got
interrupted, with some accumulating many gigabytes of temporary server
info files.

We have stopped generating those files in the most important places in
the preceding commit. This commit also introduces cleanup logic such
that we automatically remove all server info to make the disk space
available again.

Changelog: performance

2 files changed, 89 insertions, 0 deletions

diff --git a/internal/git/housekeeping/clean_stale_data.go b/internal/git/housekeeping/clean_stale_data.go
index 2cea9164a..931a97553 100644
--- a/internal/git/housekeeping/clean_stale_data.go
+++ b/internal/git/housekeeping/clean_stale_data.go
@@ -59,6 +59,7 @@ func (m *RepositoryManager) CleanStaleData(ctx context.Context, repo *localrepo.
 		"reflocks":       findStaleReferenceLocks,
 		"packedrefslock": findPackedRefsLock,
 		"packedrefsnew":  findPackedRefsNew,
+		"serverinfo":     findServerInfo,
 	} {
 		staleFiles, err := staleFileFinder(ctx, repoPath)
 		if err != nil {
@@ -364,6 +365,43 @@ func findPackedRefsNew(ctx context.Context, repoPath string) ([]string, error) {
 	return findStaleFiles(repoPath, packedRefsNewGracePeriod, "packed-refs.new")
 }
 
+// findServerInfo returns files generated by git-update-server-info(1). These files are only
+// required to serve Git fetches via the dumb HTTP protocol, which we don't serve at all. It's thus
+// safe to remove all of those files without a grace period.
+func findServerInfo(ctx context.Context, repoPath string) ([]string, error) {
+	var serverInfoFiles []string
+
+	for directory, basename := range map[string]string{
+		filepath.Join(repoPath, "info"):            "refs",
+		filepath.Join(repoPath, "objects", "info"): "packs",
+	} {
+		entries, err := os.ReadDir(directory)
+		if err != nil {
+			if errors.Is(err, os.ErrNotExist) {
+				continue
+			}
+
+			return nil, fmt.Errorf("reading info directory: %w", err)
+		}
+
+		for _, entry := range entries {
+			if !entry.Type().IsRegular() {
+				continue
+			}
+
+			// An exact match is the actual file we care about, while the latter pattern
+			// refers to the temporary files Git uses to write those files.
+			if entry.Name() != basename && !strings.HasPrefix(entry.Name(), basename+"_") {
+				continue
+			}
+
+			serverInfoFiles = append(serverInfoFiles, filepath.Join(directory, entry.Name()))
+		}
+	}
+
+	return serverInfoFiles, nil
+}
+
 // FixDirectoryPermissions does a recursive directory walk to look for
 // directories that cannot be accessed by the current user, and tries to
 // fix those with chmod. The motivating problem is that directories with mode
diff --git a/internal/git/housekeeping/clean_stale_data_test.go b/internal/git/housekeeping/clean_stale_data_test.go
index 7d9796417..75ba47119 100644
--- a/internal/git/housekeeping/clean_stale_data_test.go
+++ b/internal/git/housekeeping/clean_stale_data_test.go
@@ -139,6 +139,7 @@ type cleanStaleDataMetrics struct {
 	refsEmptyDir   int
 	packedRefsLock int
 	packedRefsNew  int
+	serverInfo     int
 }
 
 func requireCleanStaleDataMetrics(t *testing.T, m *RepositoryManager, metrics cleanStaleDataMetrics) {
@@ -159,6 +160,7 @@ func requireCleanStaleDataMetrics(t *testing.T, m *RepositoryManager, metrics cl
 		"packedrefslock": metrics.packedRefsLock,
 		"packedrefsnew":  metrics.packedRefsNew,
 		"refsemptydir":   metrics.refsEmptyDir,
+		"serverinfo":     metrics.serverInfo,
 	} {
 		_, err := builder.WriteString(fmt.Sprintf("gitaly_housekeeping_pruned_files_total{filetype=%q} %d\n", metric, expectedValue))
 		require.NoError(t, err)
@@ -618,6 +620,55 @@ func TestRepositoryManager_CleanStaleData_withSpecificFile(t *testing.T) {
 	}
 }
 
+func TestRepositoryManager_CleanStaleData_serverInfo(t *testing.T) {
+	ctx := testhelper.Context(t)
+
+	cfg, repoProto, repoPath := testcfg.BuildWithRepo(t)
+	repo := localrepo.NewTestRepo(t, cfg, repoProto)
+
+	entries := []entry{
+		d("info", 0o755, 0, Keep, []entry{
+			f("ref", 0, 0o644, Keep),
+			f("refs", 0, 0o644, Delete),
+			f("refsx", 0, 0o644, Keep),
+			f("refs_123456", 0, 0o644, Delete),
+		}),
+		d("objects", 0o755, 0, Keep, []entry{
+			d("info", 0o755, 0, Keep, []entry{
+				f("pack", 0, 0o644, Keep),
+				f("packs", 0, 0o644, Delete),
+				f("packsx", 0, 0o644, Keep),
+				f("packs_123456", 0, 0o644, Delete),
+			}),
+		}),
+	}
+
+	for _, entry := range entries {
+		entry.create(t, repoPath)
+	}
+
+	staleFiles, err := findServerInfo(ctx, repoPath)
+	require.NoError(t, err)
+	require.ElementsMatch(t, []string{
+		filepath.Join(repoPath, "info/refs"),
+		filepath.Join(repoPath, "info/refs_123456"),
+		filepath.Join(repoPath, "objects/info/packs"),
+		filepath.Join(repoPath, "objects/info/packs_123456"),
+	}, staleFiles)
+
+	mgr := NewManager(nil)
+
+	require.NoError(t, mgr.CleanStaleData(ctx, repo))
+
+	for _, entry := range entries {
+		entry.validate(t, repoPath)
+	}
+
+	requireCleanStaleDataMetrics(t, mgr, cleanStaleDataMetrics{
+		serverInfo: 4,
+	})
+}
+
 func TestRepositoryManager_CleanStaleData_referenceLocks(t *testing.T) {
 	ctx := testhelper.Context(t)