fix: Few more improvments on lockfile and broken captures.

bin/background_indexer.py

@@ -38,10 +38,6 @@ class BackgroundIndexer(AbstractManager):
                 # We already have a pickle file
                 self.logger.debug(f'{uuid_path.parent} has a pickle.')
                 continue
-            elif not list(uuid_path.parent.rglob('*.har.gz')) and not list(uuid_path.parent.rglob('*.har')):
-                # No HAR file
-                self.logger.info(f'{uuid_path.parent} has no HAR file.')
-                continue
 
             if is_locked(uuid_path.parent):
                 # it is really locked
@@ -50,6 +46,14 @@ class BackgroundIndexer(AbstractManager):
 
             with uuid_path.open() as f:
                 uuid = f.read()
+
+            if not list(uuid_path.parent.rglob('*.har.gz')) and not list(uuid_path.parent.rglob('*.har')):
+                # No HAR file
+                self.logger.warning(f'{uuid_path.parent} has no HAR file.')
+                self.lookyloo.redis.hdel('lookup_dirs', uuid)
+                shutil.move(str(uuid_path.parent), str(self.discarded_captures_dir / uuid_path.parent.name))
+                continue
+
             if not self.lookyloo.redis.hexists('lookup_dirs', uuid):
                 # The capture with this UUID exists, but it is for some reason missing in lookup_dirs
                 self.lookyloo.redis.hset('lookup_dirs', uuid, str(uuid_path.parent))

lookyloo/capturecache.py

@@ -273,7 +273,11 @@ class CapturesIndex(Mapping):
             # The pickle is being created somewhere else, wait until it's done.
             while is_locked(capture_dir):
                 time.sleep(5)
-            return load_pickle_tree(capture_dir, capture_dir.stat().st_mtime, logger)
+            try:
+                return load_pickle_tree(capture_dir, capture_dir.stat().st_mtime, logger)
+            except TreeNeedsRebuild:
+                # If this exception is raised, the building failed somewhere else, let's give it another shot.
+                pass
 
         if not (har_files := sorted(capture_dir.glob('*.har'))):
             har_files = sorted(capture_dir.glob('*.har.gz'))

lookyloo/helpers.py

@@ -226,7 +226,6 @@ def get_cache_directory(root: Path, identifier: str, namespace: Optional[Union[s
 def is_locked(locked_dir_path: Path, /) -> bool:
     """Check if a capture directory is locked, if the lock is recent enough,
     and if the locking process is still running.
-    Note: if the lock has been set by the same process, we ignore it.
 
     :param locked_dir_path: Path of the directory.
     """
@@ -237,13 +236,19 @@ def is_locked(locked_dir_path: Path, /) -> bool:
 
     try:
         content = ''
-        while not content:
+        max_wait_content = 5
+        while max_wait_content > 0:
             with lock_file.open('r') as f:
                 if content := f.read():
                     break
-                # The file is empty, we're between the creation and setting the content
+            # The file is empty, we're between the creation and setting the content
-                logger.info(f'Lock file empty ({lock_file}), waiting...')
+            logger.info(f'Lock file empty ({lock_file}), waiting...')
-                time.sleep(1)
+            max_wait_content -= 1
+            time.sleep(1)
+        else:
+            logger.warning('Lock file empty for too long, removing it.')
+            lock_file.unlink(missing_ok=True)
+            return False
 
         ts, pid = content.split(';')
         try: