diff --git a/ocaml/xcp-rrdd/scripts/rrdd/test_api_wait_until_next_reading.py b/ocaml/xcp-rrdd/scripts/rrdd/test_api_wait_until_next_reading.py
index 5ca9b897fad..196c44804c1 100644
--- a/ocaml/xcp-rrdd/scripts/rrdd/test_api_wait_until_next_reading.py
+++ b/ocaml/xcp-rrdd/scripts/rrdd/test_api_wait_until_next_reading.py
@@ -1,7 +1,11 @@
 # Test: pytest -v -s ocaml/xcp-rrdd/scripts/rrdd/test_api_wait_until_next_reading.py
 """Parametrized test exercising all conditions in rrdd.API.wait_until_next_reading()"""
+import json
 import socket
+from io import BytesIO
+from struct import pack, unpack
 from warnings import catch_warnings as import_without_warnings, simplefilter
+from zlib import crc32
 
 # Dependencies:
 # pip install pytest-mock
@@ -77,3 +81,109 @@ def test_api_getter_functions(api):
     api.path = "path"
     assert api.get_header() == "header"
     assert api.get_path() == "path"
+
+
+class MockDataSource:
+    def __init__(self, name, metadata, packed_data):
+        self.name = name
+        self.metadata = metadata
+        self.packed_data = packed_data
+
+    def pack_data(self):
+        return self.packed_data
+
+
+@pytest.mark.parametrize(
+    "data_sources, expected_metadata",
+    [
+        pytest.param(
+            [
+                MockDataSource("ds1", {"key1": "value1"}, b"\x00\x01"),
+                MockDataSource("ds2", {"key2": "value2"}, b"\x00\x02"),
+            ],
+            {"key1": "value1", "key2": "value2"},
+        ),
+        pytest.param(
+            [MockDataSource("ds1", {"key1": "value1"}, b"\x00\x01")],
+            {"key1": "value1"},
+        ),
+        pytest.param(
+            [],
+            {},
+        ),
+    ],
+)
+def test_update(
+    mocker,
+    data_sources,
+    expected_metadata,
+):
+    # Arrange
+    def checksum(*args):
+        return crc32(*args) & 0xFFFFFFFF
+
+    class MockAPI(rrdd.API):
+        def __init__(self):  # pylint: disable=super-init-not-called
+            self.dest = BytesIO()
+            self.datasources = data_sources
+
+        def pack_data(self, ds: MockDataSource):
+            return ds.pack_data()
+
+    testee = MockAPI()
+    testee.deregister = mocker.Mock()
+    fixed_time = 1234567890
+    mocker.patch("time.time", return_value=fixed_time)
+
+    # Act
+    testee.update()
+
+    # Assert
+
+    # Read and unpack the header
+    testee.dest.seek(0)
+    # The header is 20 bytes long and has the following format:
+    # 0-11: "DATASOURCES" (12 bytes)
+    # 12-15: data_checksum (4 bytes)
+    # 16-19: metadata_checksum (4 bytes)
+    # 20-23: num_datasources (4 bytes)
+    # 24-31: timestamp (8 bytes)
+    header_len = len("DATASOURCES") + 4 + 4 + 4 + 8
+    header = testee.dest.read(header_len)
+    (
+        unpacked_data_checksum,
+        unpacked_metadata_checksum,
+        unpacked_num_datasources,
+        unpacked_timestamp,
+    ) = unpack(">LLLQ", header[11:])
+
+    # Assert the expected unpacked header value
+    assert header.startswith(b"DATASOURCES")
+    assert unpacked_num_datasources == len(data_sources)
+    assert unpacked_timestamp == fixed_time
+
+    #
+    # Assert datasources and the expected data checksum
+    #
+
+    # Initialize the expected checksum with the fixed time
+    expected_checksum = checksum(pack(">Q", fixed_time))
+    # Loop over the datasources and assert the packed data
+    testee.dest.seek(header_len)
+    # sourcery skip: no-loop-in-tests
+    for ds in data_sources:
+        packed_data = testee.dest.read(len(ds.pack_data()))
+        assert packed_data == ds.pack_data()
+        # Update the checksum with the packed data
+        expected_checksum = checksum(packed_data, expected_checksum)
+
+    assert unpacked_data_checksum == expected_checksum
+
+    #
+    # Assert metadata and the expected metadata checksum
+    #
+    metadata_length = unpack(">L", testee.dest.read(4))[0]
+    metadata_json = testee.dest.read(metadata_length)
+
+    assert json.loads(metadata_json) == {"datasources": expected_metadata}
+    assert unpacked_metadata_checksum == checksum(metadata_json)
diff --git a/python3/stubs/xcp/cmd.pyi b/python3/stubs/xcp/cmd.pyi
new file mode 100644
index 00000000000..950a6d28200
--- /dev/null
+++ b/python3/stubs/xcp/cmd.pyi
@@ -0,0 +1,13 @@
+from basedtyping import Untyped
+from typing import Any
+from xcp import logger as logger
+from xcp.compat import open_defaults_for_utf8_text as open_defaults_for_utf8_text
+
+def runCmd(command: bytes | str | list[str], with_stdout: bool = False, with_stderr: bool = False, inputtext: bytes | str | None = None, **kwargs: Any) -> Any: ...
+
+class OutputCache:
+    cache: Untyped
+    def __init__(self): ...
+    def fileContents(self, fn, *args, **kwargs) -> Untyped: ...
+    def runCmd(self, command, with_stdout: bool = False, with_stderr: bool = False, inputtext: Untyped | None = None, **kwargs) -> Untyped: ...
+    def clearCache(self): ...
diff --git a/python3/stubs/xcp/compat.pyi b/python3/stubs/xcp/compat.pyi
new file mode 100644
index 00000000000..bd2c7cfa4a6
--- /dev/null
+++ b/python3/stubs/xcp/compat.pyi
@@ -0,0 +1,9 @@
+from basedtyping import Untyped
+from typing import Any, IO
+
+def open_textfile(filename: str, mode: str, encoding: str = 'utf-8', **kwargs: Any) -> IO[str]: ...
+
+open_utf8: Untyped
+
+def open_with_codec_handling(filename: str, mode: str = 'r', encoding: str = 'utf-8', **kwargs: Any) -> IO[Any]: ...
+def open_defaults_for_utf8_text(args: tuple[Any, ...] | None, kwargs: Any) -> tuple[str, Any]: ...