Don't mask await_lock body exceptions as FileLockError

changelog/70229.fixed.md

@@ -0,0 +1 @@
+Fixed `salt.utils.files.await_lock` so exceptions raised inside its `async with` body are no longer converted into `FileLockError`, and the minion now logs the traceback when a job_queue lock failure occurs.

salt/minion.py

@@ -2537,6 +2537,7 @@ async def _handle_decoded_payload_impl(self, data):
                     log.warning(
                         "Failed to acquire job_queue lock for jid %s, queuing anyway.",
                         data["jid"],
+                        exc_info=True,
                     )
                     # If we can't get the lock, we assume high contention and queue it to be safe.
                     # Or we could just proceed (unsafe). Queuing is safer for stability.

salt/utils/files.py

@@ -411,6 +411,7 @@ async def await_lock(path, lock_fn=None, timeout=5, sleep=0.1, time_start=None):
     if time_start is None:
         time_start = time.time()
     obtained_lock = False
+    in_body = False
 
     def _raise_error(msg, race=False):
         """
@@ -459,6 +460,7 @@ def _raise_error(msg, race=False):
                 log.trace("Write lock %s obtained", lock_fn)
                 obtained_lock = True
                 # Transfer control back to the code inside the with block
+                in_body = True
                 yield
                 # Exit the loop
                 break
@@ -473,6 +475,9 @@ def _raise_error(msg, race=False):
         raise
 
     except Exception as exc:  # pylint: disable=broad-except
+        if in_body:
+            # Not a locking failure; let the caller see the real exception
+            raise
         _raise_error(f"Error encountered obtaining file lock {lock_fn}: {exc}")
 
     finally:

tests/pytests/unit/utils/test_files.py

@@ -474,3 +474,15 @@ def test_set_umask_nests_in_a_single_thread():
         with salt.utils.files.set_umask(0o022):
             pass
     assert salt.utils.files.get_umask() == orig
+
+
+async def test_await_lock_does_not_mask_body_exceptions(tmp_path):
+    """
+    Exceptions raised inside the ``async with`` body must propagate unchanged
+    and must not be reported as a FileLockError. The lock must be released.
+    """
+    lock_fn = str(tmp_path / "test.lock")
+    with pytest.raises(ValueError, match="boom"):
+        async with salt.utils.files.await_lock(lock_fn, lock_fn=lock_fn, timeout=1):
+            raise ValueError("boom")
+    assert not os.path.exists(lock_fn)