Ytdyklly

I am currently setting up a GCP compute instance for some reinforcement learning which requires the ability to render an image to a display. For this I am making use of pyvirtualdisplay which appears to wrap xvfb-run under the hood.

Currently I am struggling to grasp the details of glx and where it necessary comes from.

running xvfb-run -a -s +extension glx -- glxinfo,

I get /usr/bin/xvfb-run: 183: /usr/bin/xvfb-run: glx: not found

Looking at the nvidia drivers which were installed due to the use of install-nvidia-driver=True within the metadata of the GCP instance create command.

running nvidia-smi I get

+-----------------------------------------------------------------------------+

| NVIDIA-SMI 410.72       Driver Version: 410.72       CUDA Version: 10.0     |

|-------------------------------+----------------------+----------------------+

| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |

| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |

|===============================+======================+======================|

|   0  Tesla K80           Off  | 00000000:00:04.0 Off |                    0 |

| N/A   33C    P0    70W / 149W |      0MiB / 11441MiB |    100%      Default |

+-------------------------------+----------------------+----------------------+



+-----------------------------------------------------------------------------+

| Processes:                                                       GPU Memory |

|  GPU       PID   Type   Process name                             Usage      |

|=============================================================================|

|  No running processes found                                                 |

+-----------------------------------------------------------------------------+

What could be missing to cause glx to be missing from my instance? The instance was created with this command:

export IMAGE_FAMILY="tf-latest-gpu"

export ZONE="europe-west1-b"

export INSTANCE_NAME="INSTANCE_1"

export INSTANCE_TYPE="n1-standard-2"

gcloud compute instances create $INSTANCE_NAME 

        --zone=$ZONE 

        --image-family=tf-latest-gpu 

        --image-project="deeplearning-platform-release" 

        --maintenance-policy=TERMINATE 

        --accelerator="type=nvidia-tesla-k80,count=1" 

        --machine-type=$INSTANCE_TYPE 

        --boot-disk-size=120GB 

        --metadata="install-nvidia-driver=True"