Search Unity

[Help Needed] UnityTimeOutException when running on Ubuntu 18.04 with Display

Discussion in 'ML-Agents' started by reubenwong97, Apr 11, 2020.

  1. reubenwong97

    reubenwong97

    Joined:
    Jan 13, 2020
    Posts:
    3
    Hi guys,

    I created a custom environment for research purposes and was able to build it for the Windows platform and it worked perfectly fine. However, I had no such luck in Ubuntu and ran into the timeout exception. In fact, I was not even unable to open the x86_64 file, but I am not really sure what is wrong.

    Other details, not sure if relevant:
    I used an asset which made use of the Burst preview package and when I initially built for Linux from Windows, I ran into an error as I did not have Linux compile tools natively. I then disabled it and it successfully built.

    I have been stuck on this for days and looking at github issues, I cannot find any related to my problem. For most, it seems that running in a headless environment is the issue while I do indeed have a display. Moreover, I am not able to open the x86_64 file as it opens and closes immediately after.

    Any help is greatly appreciated. Thank you all very much.
     
  2. TreyK-47

    TreyK-47

    Unity Technologies

    Joined:
    Oct 22, 2019
    Posts:
    1,820
    I'll flag for the team to take a look. Which version of the editor, ML Agents, Python, & C# are you using?
     
  3. reubenwong97

    reubenwong97

    Joined:
    Jan 13, 2020
    Posts:
    3
    Hi, thank you! I am using 2019.2.17f1, using ML Agents 0.13.0 and Python 3.6.5. I would also like to add that I am connecting to my machine through remote desktop connection. Cheers!
     
  4. christophergoy

    christophergoy

    Joined:
    Sep 16, 2015
    Posts:
    735
    Can you find the player log for the executable on linux? It would help greatly in figuring out your issue.

    I also don't think the connection through remote desktop will work since the executable on linux will look for the connection on the local machine. You'd need to run the python process on the same machine.