-
Notifications
You must be signed in to change notification settings - Fork 16
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
Cannot receive data from the simulator. The connection is blocked or the simulator is not running. #15
Comments
It seems like the client can't connect to the simulator for some reason. This usually happens when the Could you check the network connection by running the below commands. This uses the low level interface, >>> from tesse.msgs import DataRequest; from tesse.env import Env
>>> print(Env().request(DataRequest()).metadata)
<TESSE_Agent_Metadata_v0.5>
<position x='-6.649457' y='0.4999968' z='-5.790709'/>
<quaternion x='0' y='0.9426415' z='0' w='-0.3338069'/>
<velocity x_dot='0' y_dot='2.233302E-06' z_dot='0'/>
<angular_velocity x_ang_dot='0' y_ang_dot='0' z_ang_dot='0'/>
<acceleration x_ddot='0' y_ddot='0' z_ddot='0'/>
<angular_acceleration x_ang_ddot='0' y_ang_ddot='0' z_ang_ddot='0'/>
<time>295.8927</time>
<collision status='false' name=''/>
<collider status='true'/>
</TESSE_Agent_Metadata_v0.5> Thanks! |
Error:
|
Here is the content of the /home/.config/unity3d/Editor/PLayer.log:
Here is the output of:
And here is the output of:
It seems unity3d player requires OpenGL "core version" as 3.2+.... which as you can see my machine fits that requirement. Also I don't understand how this could be due to outdated OpenGL since its the latest version nvidia driver on a newly installed google cloud instance with a tesla t4 gpu. |
Thanks for providing those diagnostics, that's really helpful. You're right that OpenGL shouldn't be causing the issue. I noticed that the output of
We have a writeup on how to setup a headless machine, which includes running an X server, here. Could you see if any of those instructions help? |
The last command there
I tried running the eval.py command again but the same error comes up. Thank you for helping btw :) |
Contents of
Contents of
|
Hello, I am struggling with pretty the same issue. I am also trying to use goseek on google servers. [ 952.445] |
@joeljosephjin np, glad to help 😃 . Hopefully we're close to getting this working! I have two questions:
@lexavtanke, were you able to get through the linked instructions for setting up a headless server? If so, could you also provide the contents of your |
@ZacRavichandran Thank you for your fast replay. Here is the solution, but not straightforward, they use openGL. but the most interesting thing:
This option doesn't set in your instruction. Here is the working
|
@lexavtanke That's very helpful, thanks for sharing the link! We tested those instructions against some internal servers and AWS, so it's possible that there are additional steps required on Google Cloud or Colab. Please let me know what you find! |
@ZacRavichandran Sad but new config with your instruction doesn't work may be It's because of the versions of driver for video card and kernel. I think so because of this part of code in remocolab
|
@ZacRavichandran faced with same issue. Differently, a X server process was running on my gpu.
|
@lexavtanke ah yes, it looks like there's a bit of configuration required to get a proper virtual display running on colab. I'll see if I can track down a solution. In the meantime, please let me know if you find any useful resources! |
@ZHMA1996 I noticed that there is no
Sometimes this is because the export DISPLAY=:0 Thanks! |
Should that be And to confirm, are you running this remotely? |
Yes, i am running this remotely |
It's odd that the simulator is still not finding the display. Could you try two things to dig into this further?
>>> echo $DISPLAY
>>> cd ~/goseek-challenge/simulator
>>> ./goseek-v0.1.4.x86_64 |
Ok thanks, that's helpful. To confirm the issue is with the display and not Unity, could you try to test via glxgears? This should look like the following
|
The issue should be concerned with the display. Here is the output after testing via
|
so far, GCP,AWS, RemoColab and Oracle Cloud instances show this connection error, no matter what. |
Which AWS instance type are you using? We're using G4 instances with Nvidia driver version 440.64 and CUDA 10.2. On our side, everything works as expected after setting up a virtual display. Could we walk through the steps you used to configure your AWS instance? Hopefully that'll solve the issues you're seeing. |
xvfb-run python eval.py --agent-config baselines/config/random-agent.yaml --episode-config config/check-ground-truth.yaml
gives this
i am running this on my google cloud instance through chrome remote desktop connection. It has an nvidia GPU.
The text was updated successfully, but these errors were encountered: