feat(state): improve clean up when LibreOffice and Chromium are restarted

pkg/gotenberg/gc.go

@@ -5,13 +5,14 @@ import (
 	"os"
 	"path/filepath"
 	"strings"
+	"time"
 
 	"go.uber.org/zap"
 )
 
 // GarbageCollect scans the root path and deletes files or directories with
-// names containing specific substrings.
-func GarbageCollect(logger *zap.Logger, rootPath string, includeSubstr []string) error {
+// names containing specific substrings and before a given experiation time.
+func GarbageCollect(logger *zap.Logger, rootPath string, includeSubstr []string, expirationTime time.Time) error {
 	logger = logger.Named("gc")
 
 	// To make sure that the next Walk method stays on
@@ -36,7 +37,7 @@ func GarbageCollect(logger *zap.Logger, rootPath string, includeSubstr []string)
 		}
 
 		for _, substr := range includeSubstr {
-			if strings.Contains(info.Name(), substr) || path == substr {
+			if (strings.Contains(info.Name(), substr) || path == substr) && info.ModTime().Before(expirationTime) {
 				err := os.RemoveAll(path)
 				if err != nil {
 					return fmt.Errorf("garbage collect '%s': %w", path, err)

pkg/gotenberg/gc_test.go

@@ -4,6 +4,7 @@ import (
 	"fmt"
 	"os"
 	"testing"
+	"time"
 
 	"github.com/google/uuid"
 	"go.uber.org/zap"
@@ -64,7 +65,7 @@ func TestGarbageCollect(t *testing.T) {
 				}
 			}()
 
-			err := GarbageCollect(zap.NewNop(), tc.rootPath, tc.includeSubstr)
+			err := GarbageCollect(zap.NewNop(), tc.rootPath, tc.includeSubstr, time.Now())
 
 			if !tc.expectError && err != nil {
 				t.Fatalf("expected no error but got: %v", err)

pkg/modules/chromium/browser.go

@@ -15,6 +15,7 @@ import (
 	"github.com/chromedp/cdproto/runtime"
 	"github.com/chromedp/chromedp"
 	"github.com/dlclark/regexp2"
+	"github.com/shirou/gopsutil/v4/process"
 	"go.uber.org/zap"
 
 	"github.com/gotenberg/gotenberg/v8/pkg/gotenberg"
@@ -162,21 +163,60 @@ func (b *chromiumBrowser) Stop(logger *zap.Logger) error {
 
 	// Always remove the user profile directory created by Chromium.
 	copyUserProfileDirPath := b.userProfileDirPath
-	defer func(userProfileDirPath string) {
+	expirationTime := time.Now()
+	defer func(userProfileDirPath string, expirationTime time.Time) {
+		// See:
+		// https://github.com/SeleniumHQ/docker-selenium/blob/7216d060d86872afe853ccda62db0dfab5118dc7/NodeChrome/chrome-cleanup.sh
+		// https://github.com/SeleniumHQ/docker-selenium/blob/7216d060d86872afe853ccda62db0dfab5118dc7/NodeChromium/chrome-cleanup.sh
 		go func() {
+			// Clean up stuck processes.
+			ps, err := process.Processes()
+			if err != nil {
+				logger.Error(fmt.Sprintf("list processes: %v", err))
+			} else {
+				for _, p := range ps {
+					func() {
+						cmdline, err := p.Cmdline()
+						if err != nil {
+							return
+						}
+
+						if !strings.Contains(cmdline, "chromium/chromium") && !strings.Contains(cmdline, "chrome/chrome") {
+							return
+						}
+
+						killCtx, cancel := context.WithTimeout(context.Background(), time.Second*5)
+						defer cancel()
+
+						err = p.KillWithContext(killCtx)
+						if err != nil {
+							logger.Error(fmt.Sprintf("kill process: %v", err))
+						} else {
+							logger.Info(fmt.Sprintf("Chromium process %d killed", p.Pid))
+						}
+					}()
+				}
+			}
+
 			// FIXME: Chromium seems to recreate the user profile directory
 			//  right after its deletion if we do not wait a certain amount
 			//  of time before deleting it.
 			<-time.After(10 * time.Second)
 
-			err := os.RemoveAll(userProfileDirPath)
+			err = os.RemoveAll(userProfileDirPath)
 			if err != nil {
 				logger.Error(fmt.Sprintf("remove Chromium's user profile directory: %s", err))
+			} else {
+				logger.Debug(fmt.Sprintf("'%s' Chromium's user profile directory removed", userProfileDirPath))
 			}
 
-			logger.Debug(fmt.Sprintf("'%s' Chromium's user profile directory removed", userProfileDirPath))
+			// Also remove Chromium specific files in the temporary directory.
+			err = gotenberg.GarbageCollect(logger, os.TempDir(), []string{".org.chromium.Chromium", ".com.google.Chrome"}, expirationTime)
+			if err != nil {
+				logger.Error(err.Error())
+			}
 		}()
-	}(copyUserProfileDirPath)
+	}(copyUserProfileDirPath, expirationTime)
 
 	b.ctxMu.Lock()
 	defer b.ctxMu.Unlock()

pkg/modules/libreoffice/api/libreoffice.go

@@ -190,22 +190,23 @@ func (p *libreOfficeProcess) Stop(logger *zap.Logger) error {
 
 	// Always remove the user profile directory created by LibreOffice.
 	copyUserProfileDirPath := p.userProfileDirPath
-	defer func(userProfileDirPath string) {
+	expirationTime := time.Now()
+	defer func(userProfileDirPath string, expirationTime time.Time) {
 		go func() {
 			err := os.RemoveAll(userProfileDirPath)
 			if err != nil {
 				logger.Error(fmt.Sprintf("remove LibreOffice's user profile directory: %v", err))
+			} else {
+				logger.Debug(fmt.Sprintf("'%s' LibreOffice's user profile directory removed", userProfileDirPath))
 			}
 
-			logger.Debug(fmt.Sprintf("'%s' LibreOffice's user profile directory removed", userProfileDirPath))
-
 			// Also remove LibreOffice specific files in the temporary directory.
-			err = gotenberg.GarbageCollect(logger, os.TempDir(), []string{"OSL_PIPE", ".tmp"})
+			err = gotenberg.GarbageCollect(logger, os.TempDir(), []string{"OSL_PIPE", ".tmp"}, expirationTime)
 			if err != nil {
 				logger.Error(err.Error())
 			}
 		}()
-	}(copyUserProfileDirPath)
+	}(copyUserProfileDirPath, expirationTime)
 
 	p.cfgMu.Lock()
 	defer p.cfgMu.Unlock()