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

[Dashboard] Ray 1.13.1 cannot start dashboard with grpc version 1.29.0 #28019

Closed
jianoaix opened this issue Aug 19, 2022 · 5 comments · Fixed by #28025
Closed

[Dashboard] Ray 1.13.1 cannot start dashboard with grpc version 1.29.0 #28019

jianoaix opened this issue Aug 19, 2022 · 5 comments · Fixed by #28025
Assignees
Labels
dashboard Issues specific to the Ray Dashboard triage Needs triage (eg: priority, bug/not-bug, and owning component)

Comments

@jianoaix
Copy link
Contributor

The 1.13.1 release set the required grpc version to be "grpcio >= 1.28.1, <= 1.43.0". However, A user who's running on 1.13.1 with grpc version 1.29.0 got the following error when they try to start Ray cluster:

2022-08-19 05:51:03,709	ERROR services.py:1488 -- Failed to start the dashboard: Failed to start the dashboard, return code 1
 The last 10 lines of /mnt/mesos/sandbox/ray_log/session_2022-08-19_05-51-01_217114_207/logs/dashboard.log:
  File "/usr/lib/python3.6/site-packages/ray/dashboard/modules/state/state_head.py", line 11, in <module>
    from ray.dashboard.state_aggregator import StateAPIManager
  File "/usr/lib/python3.6/site-packages/ray/dashboard/state_aggregator.py", line 21, in <module>
    from ray.experimental.state.state_manager import StateDataSourceClient
  File "/usr/lib/python3.6/site-packages/ray/experimental/state/state_manager.py", line 67, in <module>
    class StateDataSourceClient:
  File "/usr/lib/python3.6/site-packages/ray/experimental/state/state_manager.py", line 80, in StateDataSourceClient
    def __init__(self, gcs_channel: grpc.aio.Channel):
AttributeError: module 'grpc' has no attribute 'aio'

....

[2m[33m(raylet)[0m /usr/lib/python3.6/site-packages/ray/dashboard/modules/reporter/reporter_agent.py:45: UserWarning: `gpustat` package is not installed. GPU monitoring is not available. To have full functionality of the dashboard please install `pip install ray[default]`.)
[2m[33m(raylet)[0m   "`gpustat` package is not installed. GPU monitoring is "
[2m[33m(raylet)[0m [2022-08-19 05:51:05,137 E 285 333] (raylet) agent_manager.cc:107: The raylet exited immediately because the Ray agent failed. The raylet fate shares with the agent. This can happen because the Ray agent was unexpectedly killed or failed. See `dashboard_agent.log` for the root cause.
[2m[36m(pid=gcs_server)[0m [2022-08-19 05:51:05,306 E 212 212] (gcs_server) gcs_server.cc:283: Failed to get the resource load: GrpcUnavailable: RPC Error message: failed to connect to all addresses; RPC Error details: 
[2m[36m(pid=gcs_server)[0m [2022-08-19 05:51:06,306 E 212 212] (gcs_server) gcs_server.cc:283: Failed to get the resource load: GrpcUnavailable: RPC Error message: failed to connect to all addresses; RPC Error details: 
[2m[36m(pid=gcs_server)[0m [2022-08-19 05:51:07,306 E 212 212] (gcs_server) gcs_server.cc:283: Failed to get the resource load: GrpcUnavailable: RPC Error message: failed to connect to all addresses; RPC Error details: 
[2m[36m(pid=gcs_server)[0m [2022-08-19 05:51:08,306 E 212 212] (gcs_server) gcs_server.cc:283: Failed to get the resource load: GrpcUnavailable: RPC Error message: failed to connect to all addresses; RPC Error details: 
[2m[36m(pid=gcs_server)[0m [2022-08-19 05:51:09,306 E 212 212] (gcs_server) gcs_server.cc:283: Failed to get the resource load: GrpcUnavailable: RPC Error message: failed to connect to all addresses; RPC Error details: 

It appears the Dashboard is requiring a different grpc version than the rest of Ray.

cc @scv119

@scv119
Copy link
Contributor

scv119 commented Aug 19, 2022

@alanwguo seems an issue with dashboard. are you the right owner of the problem? we could either fix the dashboard, or change the grpcio requirement.

@scv119 scv119 added dashboard Issues specific to the Ray Dashboard triage Needs triage (eg: priority, bug/not-bug, and owning component) labels Aug 19, 2022
@alanwguo
Copy link
Contributor

@jianoaix , what version of python are you using? This might be related to #27269

@jianoaix
Copy link
Contributor Author

@alanwguo python 3.6.

@rkooo567
Copy link
Contributor

Hmm maybe we should bump the minimal version of grpcio. Is it possible to do this @jianoaix for the user?

@alanwguo
Copy link
Contributor

@rkooo567 , i have a PR to address this: #28025

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
dashboard Issues specific to the Ray Dashboard triage Needs triage (eg: priority, bug/not-bug, and owning component)
Projects
None yet
Development

Successfully merging a pull request may close this issue.

4 participants