Managing worker crashes

[mavann]

Hello,

I have to maange a pool of shared resources that I allocate to workers to run tests. When a worker crashes I would need a callback in the main process to identify which worker crashed so I can properly restart and re-allocate the shared resource it was using. I have tried quite a few things (runtest_logreport, sessionfinish, etc) but I can't seem to find something that works. Any ideas or suggestions?

Thanks!

[RonnyPfannschmidt]

based on

pytest-xdist/src/xdist/dsession.py

Lines 238 to 242 in 84a8544

	def worker_errordown(self, node: WorkerController, error: object | None) -> None:
	"""Emitted by the WorkerController when a node dies."""
	self.config.hook.pytest_testnodedown(node=node, error=error)
	assert self.sched is not None
	try:

i believe the
the pytest_testnodedown(*, node, error) hook should be the one to use

[mavann]

That's exactly what I was looking for, thank you!