nixpkgs/nixos/lib/test-driver/test_driver/driver.py
Raito Bezarius d4d7550108 nixos/test-driver: provide a global timeout
Since the debut of the test-driver, we didn't obtain
a race timer with the test execution to ensure that tests doesn't run beyond
a certain amount of time.

This is particularly important when you are running into hanging tests
which cannot be detected by current facilities (requires more pvpanic wiring up, QMP
API stuff, etc.).

Two easy examples:

- Some QEMU tests may get stuck in some situation and run for more than 24 hours → we default to 1 hour max.
- Some QEMU tests may panic in the wrong place, e.g. UEFI firmware or worse → end users can set a "reasonable" amount of time

And then, we should let the retry logic retest them until they succeed and adjust
their global timeouts.

Of course, this does not help with the fact that the timeout may need to be
a function of the actual busyness of the machine running the tests.
This is only one step towards increased reliability.
2023-10-29 12:45:00 +01:00

260 lines
8.8 KiB
Python

import os
import re
import signal
import tempfile
import threading
from contextlib import contextmanager
from pathlib import Path
from typing import Any, Callable, ContextManager, Dict, Iterator, List, Optional, Union
from test_driver.logger import rootlog
from test_driver.machine import Machine, NixStartScript, retry
from test_driver.polling_condition import PollingCondition
from test_driver.vlan import VLan
def get_tmp_dir() -> Path:
"""Returns a temporary directory that is defined by TMPDIR, TEMP, TMP or CWD
Raises an exception in case the retrieved temporary directory is not writeable
See https://docs.python.org/3/library/tempfile.html#tempfile.gettempdir
"""
tmp_dir = Path(tempfile.gettempdir())
tmp_dir.mkdir(mode=0o700, exist_ok=True)
if not tmp_dir.is_dir():
raise NotADirectoryError(
f"The directory defined by TMPDIR, TEMP, TMP or CWD: {tmp_dir} is not a directory"
)
if not os.access(tmp_dir, os.W_OK):
raise PermissionError(
f"The directory defined by TMPDIR, TEMP, TMP, or CWD: {tmp_dir} is not writeable"
)
return tmp_dir
def pythonize_name(name: str) -> str:
return re.sub(r"^[^A-z_]|[^A-z0-9_]", "_", name)
class Driver:
"""A handle to the driver that sets up the environment
and runs the tests"""
tests: str
vlans: List[VLan]
machines: List[Machine]
polling_conditions: List[PollingCondition]
global_timeout: int
race_timer: threading.Timer
def __init__(
self,
start_scripts: List[str],
vlans: List[int],
tests: str,
out_dir: Path,
keep_vm_state: bool = False,
global_timeout: int = 24 * 60 * 60 * 7,
):
self.tests = tests
self.out_dir = out_dir
self.global_timeout = global_timeout
self.race_timer = threading.Timer(global_timeout, self.terminate_test)
tmp_dir = get_tmp_dir()
with rootlog.nested("start all VLans"):
vlans = list(set(vlans))
self.vlans = [VLan(nr, tmp_dir) for nr in vlans]
def cmd(scripts: List[str]) -> Iterator[NixStartScript]:
for s in scripts:
yield NixStartScript(s)
self.polling_conditions = []
self.machines = [
Machine(
start_command=cmd,
keep_vm_state=keep_vm_state,
name=cmd.machine_name,
tmp_dir=tmp_dir,
callbacks=[self.check_polling_conditions],
out_dir=self.out_dir,
)
for cmd in cmd(start_scripts)
]
def __enter__(self) -> "Driver":
return self
def __exit__(self, *_: Any) -> None:
with rootlog.nested("cleanup"):
self.race_timer.cancel()
for machine in self.machines:
machine.release()
def subtest(self, name: str) -> Iterator[None]:
"""Group logs under a given test name"""
with rootlog.nested("subtest: " + name):
try:
yield
return True
except Exception as e:
rootlog.error(f'Test "{name}" failed with error: "{e}"')
raise e
def test_symbols(self) -> Dict[str, Any]:
@contextmanager
def subtest(name: str) -> Iterator[None]:
return self.subtest(name)
general_symbols = dict(
start_all=self.start_all,
test_script=self.test_script,
machines=self.machines,
vlans=self.vlans,
driver=self,
log=rootlog,
os=os,
create_machine=self.create_machine,
subtest=subtest,
run_tests=self.run_tests,
join_all=self.join_all,
retry=retry,
serial_stdout_off=self.serial_stdout_off,
serial_stdout_on=self.serial_stdout_on,
polling_condition=self.polling_condition,
Machine=Machine, # for typing
)
machine_symbols = {pythonize_name(m.name): m for m in self.machines}
# If there's exactly one machine, make it available under the name
# "machine", even if it's not called that.
if len(self.machines) == 1:
(machine_symbols["machine"],) = self.machines
vlan_symbols = {
f"vlan{v.nr}": self.vlans[idx] for idx, v in enumerate(self.vlans)
}
print(
"additionally exposed symbols:\n "
+ ", ".join(map(lambda m: m.name, self.machines))
+ ",\n "
+ ", ".join(map(lambda v: f"vlan{v.nr}", self.vlans))
+ ",\n "
+ ", ".join(list(general_symbols.keys()))
)
return {**general_symbols, **machine_symbols, **vlan_symbols}
def test_script(self) -> None:
"""Run the test script"""
with rootlog.nested("run the VM test script"):
symbols = self.test_symbols() # call eagerly
exec(self.tests, symbols, None)
def run_tests(self) -> None:
"""Run the test script (for non-interactive test runs)"""
rootlog.info(
f"Test will time out and terminate in {self.global_timeout} seconds"
)
self.race_timer.start()
self.test_script()
# TODO: Collect coverage data
for machine in self.machines:
if machine.is_up():
machine.execute("sync")
def start_all(self) -> None:
"""Start all machines"""
with rootlog.nested("start all VMs"):
for machine in self.machines:
machine.start()
def join_all(self) -> None:
"""Wait for all machines to shut down"""
with rootlog.nested("wait for all VMs to finish"):
for machine in self.machines:
machine.wait_for_shutdown()
self.race_timer.cancel()
def terminate_test(self) -> None:
# This will be usually running in another thread than
# the thread actually executing the test script.
with rootlog.nested("timeout reached; test terminating..."):
for machine in self.machines:
machine.release()
# As we cannot `sys.exit` from another thread
# We can at least force the main thread to get SIGTERM'ed.
# This will prevent any user who caught all the exceptions
# to swallow them and prevent itself from terminating.
os.kill(os.getpid(), signal.SIGTERM)
def create_machine(self, args: Dict[str, Any]) -> Machine:
tmp_dir = get_tmp_dir()
if args.get("startCommand"):
start_command: str = args.get("startCommand", "")
cmd = NixStartScript(start_command)
name = args.get("name", cmd.machine_name)
else:
cmd = Machine.create_startcommand(args) # type: ignore
name = args.get("name", "machine")
return Machine(
tmp_dir=tmp_dir,
out_dir=self.out_dir,
start_command=cmd,
name=name,
keep_vm_state=args.get("keep_vm_state", False),
)
def serial_stdout_on(self) -> None:
rootlog._print_serial_logs = True
def serial_stdout_off(self) -> None:
rootlog._print_serial_logs = False
def check_polling_conditions(self) -> None:
for condition in self.polling_conditions:
condition.maybe_raise()
def polling_condition(
self,
fun_: Optional[Callable] = None,
*,
seconds_interval: float = 2.0,
description: Optional[str] = None,
) -> Union[Callable[[Callable], ContextManager], ContextManager]:
driver = self
class Poll:
def __init__(self, fun: Callable):
self.condition = PollingCondition(
fun,
seconds_interval,
description,
)
def __enter__(self) -> None:
driver.polling_conditions.append(self.condition)
def __exit__(self, a, b, c) -> None: # type: ignore
res = driver.polling_conditions.pop()
assert res is self.condition
def wait(self, timeout: int = 900) -> None:
def condition(last: bool) -> bool:
if last:
rootlog.info(f"Last chance for {self.condition.description}")
ret = self.condition.check(force=True)
if not ret and not last:
rootlog.info(
f"({self.condition.description} failure not fatal yet)"
)
return ret
with rootlog.nested(f"waiting for {self.condition.description}"):
retry(condition, timeout=timeout)
if fun_ is None:
return Poll
else:
return Poll(fun_)