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.

Sign up for GitHub

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

Jump to bottom

[autoscaler] Small fixes for local cluster usability #4864

Merged
merged 7 commits into from
Jul 7, 2019
Merged

[autoscaler] Small fixes for local cluster usability #4864

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
edc5ab1
A couple small fixes
richardliaw May 25, 2019
190ad7a
update verbose
richardliaw May 27, 2019
cbafab9
flake
richardliaw Jun 26, 2019
8e4b8fb
Merge branch 'master' into fix_autoscaler
richardliaw Jun 26, 2019
90db92f
verbose
richardliaw Jun 27, 2019
d5da604
Merge branch 'master' into fix_autoscaler
richardliaw Jul 2, 2019
fd6016d
redirect
richardliaw Jul 4, 2019
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
19 changes: 15 additions & 4 deletions python/ray/autoscaler/local/node_provider.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,10 +28,15 @@ def __init__(self, lock_path, save_path, provider_config):
with self.file_lock:
if os.path.exists(self.save_path):
workers = json.loads(open(self.save_path).read())
head_config = workers.get(provider_config["head_ip"])
if not head_config or head_config.get(
"tags", {}).get(TAG_RAY_NODE_TYPE) != "head":
workers = {}
logger.info("Head IP changed - recreating cluster.")
else:
workers = {}
logger.info("ClusterState: "
"Loaded cluster state: {}".format(workers))
"Loaded cluster state: {}".format(list(workers)))
for worker_ip in provider_config["worker_ips"]:
if worker_ip not in workers:
workers[worker_ip] = {
Expand All @@ -55,8 +60,8 @@ def __init__(self, lock_path, save_path, provider_config):
TAG_RAY_NODE_TYPE] == "head"
assert len(workers) == len(provider_config["worker_ips"]) + 1
with open(self.save_path, "w") as f:
logger.info("ClusterState: "
"Writing cluster state: {}".format(workers))
logger.debug("ClusterState: "
"Writing cluster state: {}".format(workers))
f.write(json.dumps(workers))

def get(self):
Expand All @@ -74,11 +79,17 @@ def put(self, worker_id, info):
workers[worker_id] = info
with open(self.save_path, "w") as f:
logger.info("ClusterState: "
"Writing cluster state: {}".format(workers))
"Writing cluster state: {}".format(
list(workers)))
f.write(json.dumps(workers))


class LocalNodeProvider(NodeProvider):
"""NodeProvider for private/local clusters.

`node_id` is overloaded to also be `node_ip` in this class.
"""

def __init__(self, provider_config, cluster_name):
NodeProvider.__init__(self, provider_config, cluster_name)
self.state = ClusterState("/tmp/cluster-{}.lock".format(cluster_name),
Expand Down
15 changes: 7 additions & 8 deletions python/ray/autoscaler/updater.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@
# How long to wait for a node to start, in seconds
NODE_START_WAIT_S = 300
SSH_CHECK_INTERVAL = 5
CONTROL_PATH_MAX_LENGTH = 70


def get_default_ssh_options(private_key, connect_timeout, ssh_control_path):
Expand Down Expand Up @@ -56,7 +57,7 @@ def __init__(self,
use_internal_ip=False):

ssh_control_path = "/tmp/{}_ray_ssh_sockets/{}".format(
getuser(), cluster_name)
getuser(), cluster_name)[:CONTROL_PATH_MAX_LENGTH]

self.daemon = True
self.process_runner = process_runner
Expand Down Expand Up @@ -197,12 +198,11 @@ def sync_file_mounts(self, sync_cmd):
m = "{}: Synced {} to {}".format(self.node_id, local_path,
remote_path)
with LogTimer("NodeUpdater {}".format(m)):
with open("/dev/null", "w") as redirect:
self.ssh_cmd(
"mkdir -p {}".format(os.path.dirname(remote_path)),
redirect=redirect,
)
sync_cmd(local_path, remote_path, redirect=redirect)
self.ssh_cmd(
"mkdir -p {}".format(os.path.dirname(remote_path)),
redirect=None,
)
sync_cmd(local_path, remote_path, redirect=None)

def do_update(self):
self.provider.set_node_tags(self.node_id,
Expand All @@ -223,7 +223,6 @@ def do_update(self):
# Run init commands
self.provider.set_node_tags(self.node_id,
{TAG_RAY_NODE_STATUS: "setting-up"})

m = "{}: Initialization commands completed".format(self.node_id)
with LogTimer("NodeUpdater: {}".format(m)):
for cmd in self.initialization_commands:
Expand Down