Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[autoscaler v2] test ray-installer #35875

Merged
merged 5 commits into from
May 31, 2023
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Next Next commit
add test
  • Loading branch information
scv119 committed May 30, 2023
commit de83534c11398f71ba784c78d62ba747db0387c1
12 changes: 10 additions & 2 deletions python/ray/autoscaler/v2/instance_manager/ray_installer.py
Original file line number Diff line number Diff line change
@@ -1,4 +1,5 @@
import logging
import subprocess

from ray.autoscaler._private.updater import NodeUpdater
from ray.autoscaler._private.util import with_envs, with_head_node_ip
Expand All @@ -18,9 +19,11 @@ def __init__(
self,
provider: NodeProviderV1,
config: NodeProviderConfig,
process_runner=subprocess,
) -> None:
self._provider = provider
self._config = config
self._process_runner = process_runner

def install_ray(self, instance: Instance, head_node_ip: str) -> bool:
"""
Expand Down Expand Up @@ -71,6 +74,11 @@ def install_ray(self, instance: Instance, head_node_ip: str) -> bool:
use_internal_ip=True,
docker_config=docker_config,
node_resources=instance.node_resources,
process_runner=self._process_runner,
)
updater.run()
# TODO: handle failures
try:
updater.run()
except Exception as e:
# Errors has already been handled.
return False
return True
94 changes: 94 additions & 0 deletions python/ray/autoscaler/v2/tests/test_ray_install.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,94 @@
# coding: utf-8
scv119 marked this conversation as resolved.
Show resolved Hide resolved
import os
import sys
import unittest

import pytest # noqa

from ray._private.test_utils import load_test_config
from ray.autoscaler.v2.instance_manager.config import NodeProviderConfig
from ray.autoscaler.v2.instance_manager.ray_installer import RayInstaller
from ray.core.generated.instance_manager_pb2 import Instance
from ray.tests.autoscaler_test_utils import MockProcessRunner, MockProvider


class FakeCounter:
def dec(self, *args, **kwargs):
pass


class RayInstallerTest(unittest.TestCase):
def setUp(self):
self.base_provider = MockProvider()
self.instance_config_provider = NodeProviderConfig(
load_test_config("test_ray_complex.yaml")
)
self.ray_installer = RayInstaller(
self.base_provider, self.instance_config_provider, MockProcessRunner()
)

def test_node_providers_pass_through(self):
nodes = self.node_provider.create_nodes("worker_nodes1", 1)
assert len(nodes) == 1
assert nodes[0] == Instance(
instance_type="worker_nodes1",
cloud_instance_id="0",
internal_ip="172.0.0.0",
external_ip="1.2.3.4",
status=Instance.INSTANCE_STATUS_UNSPECIFIED,
)
self.assertEqual(len(self.base_provider.mock_nodes), 1)
self.assertEqual(self.node_provider.get_non_terminated_nodes(), {"0": nodes[0]})
nodes1 = self.node_provider.create_nodes("worker_nodes", 2)
assert len(nodes1) == 2
assert nodes1[0] == Instance(
instance_type="worker_nodes",
cloud_instance_id="1",
internal_ip="172.0.0.1",
external_ip="1.2.3.4",
status=Instance.INSTANCE_STATUS_UNSPECIFIED,
)
assert nodes1[1] == Instance(
instance_type="worker_nodes",
cloud_instance_id="2",
internal_ip="172.0.0.2",
external_ip="1.2.3.4",
status=Instance.INSTANCE_STATUS_UNSPECIFIED,
)
self.assertEqual(
self.node_provider.get_non_terminated_nodes(),
{"0": nodes[0], "1": nodes1[0], "2": nodes1[1]},
)
self.assertEqual(
self.node_provider.get_nodes_by_cloud_instance_id(["0"]),
{
"0": nodes[0],
},
)
self.node_provider.terminate_node("0")
self.assertEqual(
self.node_provider.get_non_terminated_nodes(),
{"1": nodes1[0], "2": nodes1[1]},
)
self.assertFalse(self.node_provider.is_readonly())

def test_create_node_failure(self):
self.base_provider.error_creates = NodeLaunchException(
"hello", "failed to create node", src_exc_info=None
)
self.assertEqual(self.node_provider.create_nodes("worker_nodes1", 1), [])
self.assertEqual(len(self.base_provider.mock_nodes), 0)
self.assertTrue(
"worker_nodes1" in self.availability_tracker.summary().node_availabilities
)
self.assertEqual(
self.node_provider.get_non_terminated_nodes(),
{},
)


if __name__ == "__main__":
if os.environ.get("PARALLEL_CI"):
sys.exit(pytest.main(["-n", "auto", "--boxed", "-vs", __file__]))
else:
sys.exit(pytest.main(["-sv", __file__]))