Failed to register worker 01000000ffffffffffffffffffffffffffffffffffffffffffffffff to Raylet. IOError: [RayletClient] Unable to register worker with raylet. No such file or directory #100
-
Beta Was this translation helpful? Give feedback.
Replies: 4 comments 7 replies
-
Hi, Can you upgrade to the latest version ( |
Beta Was this translation helpful? Give feedback.
-
Thanks for the response, This is a very simple demo. Following your advice, I upgraded the version, but the problem still persists. from lean_dojo import *
repo = LeanGitRepo(
# mathlib的地址
"https://github.com/leanprover-community/mathlib4",
# 一个具体的版本
"3ce43c18f614b76e161f911b75a3e1ef641620ff",
)
repo.get_config("lean-toolchain")
# A few minutes if the traced repo is in the cache; many hours otherwise.
traced_repo = trace(repo) error /home/rise/anaconda3/envs/chx-py3.10/bin/python /home/rise/learn/leanDojo/extract.py
2023-11-30 11:18:52.309 | INFO | lean_dojo.data_extraction.trace:trace:182 - Loading the traced repo from /home/rise/.cache/lean_dojo/leanprover-community-mathlib4-3ce43c18f614b76e161f911b75a3e1ef641620ff/mathlib4
2023-11-30 11:18:53,092 INFO worker.py:1489 -- Connecting to existing Ray cluster at address: 198.18.0.1:6379...
[2023-11-30 11:18:53,096 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:18:54,098 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:18:55,099 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:18:56,100 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:18:57,102 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:18:58,103 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:18:59,105 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:19:00,106 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:19:01,107 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
[2023-11-30 11:19:02,109 W 108247 108247] global_state_accessor.cc:407: Some processes that the driver needs to connect to have not registered with GCS, so retrying. Have you run 'ray start' on this node?
2023-11-30 11:19:03,111 INFO worker.py:1638 -- Failed to connect to the default Ray cluster address at 198.18.0.1:6379. This is most likely due to a previous Ray instance that has since crashed. To reset the default address to connect to, run `ray stop` or restart Ray with `ray start`.
Traceback (most recent call last):
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/ray/_private/worker.py", line 1629, in init
_global_node = ray._private.node.Node(
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/ray/_private/node.py", line 254, in __init__
node_info = ray._private.services.get_node_to_connect_for_driver(
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/ray/_private/services.py", line 475, in get_node_to_connect_for_driver
return global_state.get_node_to_connect_for_driver(node_ip_address)
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/ray/_private/state.py", line 751, in get_node_to_connect_for_driver
return self.global_state_accessor.get_node_to_connect_for_driver(
File "python/ray/includes/global_state_accessor.pxi", line 218, in ray._raylet.GlobalStateAccessor.get_node_to_connect_for_driver
RuntimeError: b'GCS has started but no raylets have registered yet.'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/rise/learn/leanDojo/extract.py", line 10, in <module>
traced_repo = trace(repo)
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/lean_dojo/data_extraction/trace.py", line 183, in trace
traced_repo = TracedRepo.load_from_disk(cached_path)
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/lean_dojo/data_extraction/traced_data.py", line 1481, in load_from_disk
with ray_actor_pool(_TracedRepoHelper, root_dir, repo) as pool:
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/contextlib.py", line 135, in __enter__
return next(self.gen)
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/lean_dojo/utils.py", line 72, in ray_actor_pool
ray.init()
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/ray/_private/client_mode_hook.py", line 103, in wrapper
return func(*args, **kwargs)
File "/home/rise/anaconda3/envs/chx-py3.10/lib/python3.10/site-packages/ray/_private/worker.py", line 1645, in init
raise ConnectionError
ConnectionError
Process finished with exit code 1
|
Beta Was this translation helpful? Give feedback.
-
What will happen if you first run import ray
ray.init() Is there any error? BTW, if the problem is with Ray, a (slower) workaround is to disable Ray by setting the |
Beta Was this translation helpful? Give feedback.
-
If you cannot do import ray
ray.init() , the problem is specific to your particular system and Ray installation. There is nothing we can do on LeanDojo's side (other than you can set |
Beta Was this translation helpful? Give feedback.
If you cannot do
, the problem is specific to your particular system and Ray installation. There is nothing we can do on LeanDojo's side (other than you can set
NUM_PROCS=1
to disable Ray). I'd suggest you ask in Ray's GitHub repo.