search

carla-simulator / carla

Open-source simulator for autonomous driving research.
http://carla.org
MIT License
10.66k stars 3.42k forks source link

carla docker runs wrong #3431

Closed yangyueren closed 3 years ago

yangyueren commented 3 years ago

docker run -p 2000-2002:2000-2002 --runtime=nvidia --gpus all carlasim/carla:0.9.10.1 sh: 1: xdg-user-dir: not found

yangyueren commented 3 years ago

When I try the dockerfile, it goes wrong too. https://github.com/carla-simulator/carla/tree/master/Util/Docker

Step 8/9 : RUN ./GenerateProjectFiles.sh && make ---> Running in 63c6a77889e8 Attempting to set up UE4 pretty printers for gdb (existing UE4Printers.py, if any, will be overwritten)... updated UE4Printers.py no ~/.gdbinit file found - creating a new one.

Setting up Unreal Engine 4 project files...

Fixing inconsistent case in filenames. Setting up Mono ERROR: Unable to instantiate instance of 'BenchmarkToolTarget' object type from compiled assembly 'UE4ProgramRules'. Unreal Build Tool creates an instance of your module's 'Rules' object in order to find out about your module's requirements. The CLR exception details may provide more information: System.Reflection.TargetInvocationException: Exception has been thrown by the target of an invocation. ---> UnrealBuildTool.BuildException: GetBuildPlatform: No BuildPlatform found for Linux at UnrealBuildTool.UEBuildPlatform.GetBuildPlatform (UnrealBuildTool.UnrealTargetPlatform InPlatform, System.Boolean bInAllowFailure) [0x0003e] in <0f62df6941b7421da985bf66bc90ab8a>:0 at UnrealBuildTool.TargetRules..ctor (UnrealBuildTool.TargetInfo Target) [0x002f2] in <0f62df6941b7421da985bf66bc90ab8a>:0 at BenchmarkToolTarget..ctor (UnrealBuildTool.TargetInfo Target) [0x00000] in :0 at (wrapper managed-to-native) System.Reflection.MonoCMethod.InternalInvoke(System.Reflection.MonoCMethod,object,object[],System.Exception&) at System.Reflection.MonoCMethod.InternalInvoke (System.Object obj, System.Object[] parameters, System.Boolean wrapExceptions) [0x00005] in <98fac219bd4e453693d76fda7bd96ab0>:0 --- End of inner exception stack trace --- at System.Reflection.MonoCMethod.InternalInvoke (System.Object obj, System.Object[] parameters, System.Boolean wrapExceptions) [0x00017] in <98fac219bd4e453693d76fda7bd96ab0>:0 at System.Reflection.MonoCMethod.DoInvoke (System.Object obj, System.Reflection.BindingFlags invokeAttr, System.Reflection.Binder binder, System.Object[] parameters, System.Globalization.CultureInfo culture) [0x0007a] in <98fac219bd4e453693d76fda7bd96ab0>:0 at System.Reflection.MonoCMethod.Invoke (System.Object obj, System.Reflection.BindingFlags invokeAttr, System.Reflection.Binder binder, System.Object[] parameters, System.Globalization.CultureInfo culture) [0x0002f] in <98fac219bd4e453693d76fda7bd96ab0>:0 at System.Reflection.MethodBase.Invoke (System.Object obj, System.Object[] parameters) [0x00000] in <98fac219bd4e453693d76fda7bd96ab0>:0 at UnrealBuildTool.RulesAssembly.CreateTargetRulesInstance (System.String TypeName, UnrealBuildTool.TargetInfo TargetInfo) [0x00093] in <0f62df6941b7421da985bf66bc90ab8a>:0 The command '/bin/sh -c ./GenerateProjectFiles.sh && make' returned a non-zero code: 5

jHaselberger commented 3 years ago

docker run -p 2000-2002:2000-2002 --runtime=nvidia --gpus all carlasim/carla:0.9.10.1 sh: 1: xdg-user-dir: not found

I had the same issue. Starting the container with a shared X context works:

docker run \
  -e SDL_VIDEODRIVER=x11 \
  -e DISPLAY=$DISPLAY\
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -p 2000-2002:2000-2002 \
  -it \
  --gpus all \
  carlasim/carla:0.9.10.1 ./CarlaUE4.sh -opengl

btw you don't have to pass --runtime=nvidia.

zchrissirhcz commented 3 years ago

docker run -p 2000-2002:2000-2002 --runtime=nvidia --gpus all carlasim/carla:0.9.10.1 sh: 1: xdg-user-dir: not found

I had the same issue. Starting the container with a shared X context works:

docker run \
  -e SDL_VIDEODRIVER=x11 \
  -e DISPLAY=$DISPLAY\
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -p 2000-2002:2000-2002 \
  -it \
  --gpus all \
  carlasim/carla:0.9.10.1 ./CarlaUE4.sh -opengl

btw you don't have to pass --runtime=nvidia.

I'm using Ubuntu16.04, execute the pasted command but got error:

4.24.3-0+++UE4+Release-4.24 518 0
Disabling core dumps.
sh: 1: xdg-user-dir: not found
No protocol specified
No protocol specified
No protocol specified

@gismo07

jHaselberger commented 3 years ago

Ok maybee we should check if your docker setup is working properly. Can you 'see' your gpus when issuing:

sudo docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi

now you should see something like this:

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 450.51.06    Driver Version: 450.51.06    CUDA Version: 11.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:00:1E.0 Off |                    0 |
| N/A   34C    P8     9W /  70W |      0MiB / 15109MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
zchrissirhcz commented 3 years ago

@gismo07 Yes, I can correct run and get result for command sudo docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi. (Now I'm not with my computer but I remember I can run that)

zchrissirhcz commented 3 years ago

On my ubuntu16.04, it is also nvidia 450 driver which support cuda 11.0

zchrissirhcz commented 3 years ago

On my ubuntu16.04, first run this:

xhost local:root

then run this:

sudo docker run \
  -e SDL_VIDEODRIVER=x11 \
  -e DISPLAY=$DISPLAY\
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -p 2000-2002:2000-2002 \
  -it \
  --gpus all \
  carlasim/carla:0.9.10 ./CarlaUE4.sh -opengl

Then the CARLA window appears.

lz1159435992 commented 3 years ago

(base) [root@localhost ~]# sudo docker run --runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=1 -e SDL_VIDEODRIVER=x11 -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix -p 2000-2002:2000-2002 -it --gpus all carlasim/carla:0.9.10 ./CarlaUE4.sh -opengl 4.24.3-0+++UE4+Release-4.24 518 0 Disabling core dumps. sh: 1: xdg-user-dir: not found

Thank you, but it's not work for me. I got the same error. I guess that maybe path has some problems.

zchrissirhcz commented 3 years ago

@lz1159435992 I am not using root user, but a sudo user. From your pasted content, you are logged in as root user. Do you have a sudo user? If that case, may try it.

zchrissirhcz commented 3 years ago

@lz1159435992 Btw, I'm using a ubuntu desktop, instead of a SSH connected terminal.

lz1159435992 commented 3 years ago

Maybe this is why I have so many problems. I have faced uncountable errors since I began to configure Carla on the server without display.

zchrissirhcz commented 3 years ago

Well I think SSH to a remote linux server, which is with NVIDIA GPU, is a common required stuff.

I'm on windows10, and with Xming and putty, I can get a DISPLAY. With this, commands like xclock can be displayed in my Windows.

However I still don't know how to let CARLA window display in Windows.

jHaselberger commented 3 years ago

Maybe this is why I have so many problems. I have faced uncountable errors since I began to configure Carla on the server without display.

Hmm I think, that running carla requires a display. I was also struggling at that point.

Have you tried this: https://github.com/carla-simulator/carla/issues/225

lz1159435992 commented 3 years ago

Maybe this is why I have so many problems. I have faced uncountable errors since I began to configure Carla on the server without display.

Hmm I think, that running carla requires a display. I was also struggling at that point.

Have you tried this: #225

OK，I will try this, if you fix this please let me know how.

thytless commented 3 years ago

docker run -d -it --runtime=nvidia --name carla-server -e NVIDIA_VISIBLE_DEVICES=1 carlasim/carla:0.9.10.1 /bin/bash docker exec -it carla-server bash SDL_VIDEODRIVER=offscreen ./CarlaUE4.sh -opengl

Above works for me.

lz1159435992 commented 3 years ago

Well I think SSH to a remote linux server, which is with NVIDIA GPU, is a common required stuff.

I'm on windows10, and with Xming and putty, I can get a DISPLAY. With this, commands like xclock can be displayed in my Windows.

However I still don't know how to let CARLA window display in Windows.

There are little tutorials for doing this. I have spend two weeks but still can‘t fix it.

lz1159435992 commented 3 years ago

SDL_VIDEODRIVER=offscreen ./CarlaUE4.sh -opengl

carla@bed76db766e3:~$ SDL_VIDEODRIVER=offscreen ./CarlaUE4.sh -opengl 4.24.3-0+++UE4+Release-4.24 518 0 Disabling core dumps. sh: 1: xdg-user-dir: not found ALSA lib confmisc.c:767:(parse_card) cannot find card '0' ALSA lib conf.c:4528:(_snd_config_evaluate) function snd_func_card_driver returned error: No such file or directory ALSA lib confmisc.c:392:(snd_func_concat) error evaluating strings ALSA lib conf.c:4528:(_snd_config_evaluate) function snd_func_concat returned error: No such file or directory ALSA lib confmisc.c:1246:(snd_func_refer) error evaluating name ALSA lib conf.c:4528:(_snd_config_evaluate) function snd_func_refer returned error: No such file or directory ALSA lib conf.c:5007:(snd_config_expand) Evaluate error: No such file or directory ALSA lib pcm.c:2495:(snd_pcm_open_noupdate) Unknown PCM default ALSA lib confmisc.c:767:(parse_card) cannot find card '0' ALSA lib conf.c:4528:(_snd_config_evaluate) function snd_func_card_driver returned error: No such file or directory ALSA lib confmisc.c:392:(snd_func_concat) error evaluating strings ALSA lib conf.c:4528:(_snd_config_evaluate) function snd_func_concat returned error: No such file or directory ALSA lib confmisc.c:1246:(snd_func_refer) error evaluating name ALSA lib conf.c:4528:(_snd_config_evaluate) function snd_func_refer returned error: No such file or directory ALSA lib conf.c:5007:(snd_config_expand) Evaluate error: No such file or directory ALSA lib pcm.c:2495:(snd_pcm_open_noupdate) Unknown PCM default

I got this error. It's new for me.

zchrissirhcz commented 3 years ago

Well I think SSH to a remote linux server, which is with NVIDIA GPU, is a common required stuff. I'm on windows10, and with Xming and putty, I can get a DISPLAY. With this, commands like xclock can be displayed in my Windows. However I still don't know how to let CARLA window display in Windows.

There are little tutorials for doing this. I have spend two weeks but still can‘t fix it.

I think opengl's indirect rendering caused the failure of CARLA while xclock can start.

To be precise, I'm using VcXsrv software (Xlaunch) for X11 Service,

image

image

And glxinfo | grep ':' gives: image

which means I cannot start a program that using opengl4.6

Related links:

I've tried adding my linux user to video group, as said in the pasted reference page, re-login shell, but glxinfo's output remains.

update

Indirect rendering can't show opengl4's stuff, and indirect rendering may bring more latency:

Direct Rendering:

Application --> LibGL --> Driver --> Hardware

Indirect Rendering

Application --> LibGL --> Xorg --> Driver --> Hardware

https://superuser.com/a/984901/1194649

Thus I would not consider X11Forward+NVidia Docker solution any more.

thytless commented 3 years ago

@lz1159435992

I got this error. It's new for me.

docker exec -u root -it ${CARLA_CONTAINER_NAME} bash -c 'echo pcm.\!default { type plug slave.pcm "null" } > /etc/asound.conf' see #2820

lz1159435992 commented 3 years ago

I can't fix this. I have tried all the methods mentioned above. I try to install carla on windows10 with screen.

zchrissirhcz commented 3 years ago

@lz1159435992 Yes, I think install and run CARLA on Windows 10 with screen is the best choice.

stale[bot] commented 3 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

alaamohamed2 commented 3 years ago

@lz1159435992 Have you solved this issue?

dbersan commented 3 years ago

Maybe this is why I have so many problems. I have faced uncountable errors since I began to configure Carla on the server without display.

Hmm I think, that running carla requires a display. I was also struggling at that point.

Have you tried this: #225

I thought whole point of using Docker is not to need a display, if you are running on a remote server for instance:

This tutorial is designed for:

  • People who don't need to render the full simulation (the server is headless).

https://carla.readthedocs.io/en/0.9.11/build_docker/

dbersan commented 3 years ago

Doing:

sudo nvidia-docker run -it --rm carlasim/carla:0.9.10 /bin/bash

and then:

SDL_VIDEODRIVER=offscreen ./CarlaUE4.sh -opengl

I get the same results as @lz1159435992:

4.24.3-0+++UE4+Release-4.24 518 0 Disabling core dumps. sh: 1: xdg-user-dir: not found ALSA lib confmisc.c:767:(parse_card) cannot find card '0' ...

But maybe this is normal? When checking nvidia-smi it does appear to be using the card. However, I am unable to connect to the carla-server using a local client...

Traceback (most recent call last): File "set_environment.py", line 100, in main world = client.load_world(args.town) RuntimeError: time-out of 5000ms while waiting for the simulator, make sure the simulator is ready and connected to 127.0.0.1:2000