[pull] master from kevoreilly:master

lib/cuckoo/common/gcp.py

@@ -175,7 +175,7 @@ def check_exists(self, analysis_id: int) -> bool:
         GCS_ENABLED = False
 
 
-def download_from_gcs(gcs_uri: str, destination_path: str, logger: Optional[Any] = None, client: Optional[storage.Client] = None) -> bool:
+def download_from_gcs(gcs_uri: str, destination_path: str, logger: Optional[Any] = None, client: Optional["storage.Client"] = None) -> bool:
     """
     Downloads a file from GCS.
     gcs_uri: gs://bucket_name/object_name

tests/test_run_task_adapter.py

@@ -12,6 +12,9 @@ def fake_process(target=None, sample_sha256=None, task=None, report=False, auto=
     monkeypatch.setattr(proc, "process", fake_process)
     monkeypatch.setattr(proc, "db", db)
 
+    # Force path_exists to return True so the analysis path check succeeds
+    monkeypatch.setattr(proc, "path_exists", lambda *args, **kwargs: True)
+
     with db.session.begin():
         tid = db.add_path(temp_pe32)
         task = db.view_task(tid)
@@ -20,3 +23,35 @@ def fake_process(target=None, sample_sha256=None, task=None, report=False, auto=
 
     assert captured["task_id"] == tid
     assert captured["auto"] is True and captured["report"] is True
+
+
+def test_run_task_fails_gracefully_when_analysis_dir_missing(monkeypatch, db, temp_pe32):
+    process_called = False
+
+    def fake_process(*args, **kwargs):
+        nonlocal process_called
+        process_called = True
+
+    monkeypatch.setattr(proc, "process", fake_process)
+    monkeypatch.setattr(proc, "db", db)
+
+    # Force path_exists to return False for the analysis path
+    original_path_exists = proc.path_exists
+    def fake_path_exists(path, *args, **kwargs):
+        if "storage/analyses" in str(path):
+            return False
+        return original_path_exists(path, *args, **kwargs)
+
+    monkeypatch.setattr(proc, "path_exists", fake_path_exists)
+
+    with db.session.begin():
+        tid = db.add_path(temp_pe32)
+        task = db.view_task(tid)
+
+    proc.run_task(task)
+
+    assert not process_called, "process should not have been called when analysis dir is missing"
+
+    with db.session.begin():
+        updated_task = db.view_task(tid)
+        assert updated_task.status == proc.TASK_FAILED_PROCESSING

utils/dist.py

@@ -4,7 +4,6 @@
 # This file is part of Cuckoo Sandbox - http://www.cuckoosandbox.org
 # See the file 'docs/LICENSE' for copying permission.
 
-# https://github.com/cuckoosandbox/cuckoo/pull/1694/files
 import argparse
 import distutils.util
 import hashlib
@@ -1228,9 +1227,21 @@ def submit_tasks(self, node_name, pend_tasks_num, options_like=False, force_push
                     # print(t.category, t.target)
                     if t.category in ("file", "pcap", "static"):
                         if not path_exists(t.target):
-                            log.info("Task id: %d - File doesn't exist: %s", t.id, t.target)
-                            main_db.set_status(t.id, TASK_BANNED)
-                            continue
+                            sample_sha256 = None
+                            try:
+                                if t.sample:
+                                    sample_sha256 = t.sample.sha256
+                            except Exception as e:
+                                log.debug("Failed to lazy load sample relation for task %d: %s", t.id, e)
+
+                            bin_path = os.path.join(CUCKOO_ROOT, "storage", "binaries", sample_sha256) if sample_sha256 else None
+                            if bin_path and path_exists(bin_path):
+                                log.info("Task id: %d - Target file not found at original path, but found in binaries storage: %s. Updating target path.", t.id, bin_path)
+                                t.target = bin_path
+                            else:
+                                log.info("Task id: %d - File doesn't exist: %s", t.id, t.target)
+                                main_db.set_status(t.id, TASK_BANNED)
+                                continue
                         if not web_conf.general.allow_ignore_size and "ignore_size_check" not in t.options:
                             # We can't upload size bigger than X to our workers. In case we extract archive that contains bigger file.
                             file_size = path_get_size(t.target)

utils/process.py

@@ -40,6 +40,7 @@
 from lib.cuckoo.common.utils import get_options, option_dict_enabled
 from lib.cuckoo.core.database import Database, init_database
 from lib.cuckoo.core.data.task import (
+    TASK_FAILED_PROCESSING,
     TASK_FAILED_REPORTING,
     TASK_REPORTED
 )
@@ -190,6 +191,13 @@ def process(
 def run_task(task, memory_debugging=False, debug=False):
     """Run exactly one completed task to completion (processing -> report).
     Extracted from autoprocess so every engine shares identical per-task setup."""
+    analysis_path = os.path.join(CUCKOO_ROOT, "storage", "analyses", str(task.id))
+    if not path_exists(analysis_path):
+        log.error("Analysis directory %s does not exist. Marking task %s as failed.", analysis_path, task.id)
+        with db.session.begin():
+            db.set_status(task.id, TASK_FAILED_PROCESSING)
+        return
+
     sample_hash = ""
     if task.category != "url":
         with db.session.begin():