Configuration step to use GPU fails under proxy environment

[#1]

Hi,

I am configuring CDSW to use NVIDIA GPU in the environment where the Proxy server exists based on the procedures here:

https://www.cloudera.com/documentation/data-science-workbench/latest/topics/cdsw_gpu.html

However, I got the following error message when I tried to pull the docker image:

$sudo nvidia-docker run --rm nvidia/cuda:9.2-devel-ubuntu16.04 nvidia-smi
nvidia-docker | 2018/10/08 13:14:19 Error: Error response from daemon: Get https://registry-1.docker.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)

I guess the root cause may be related to improper DNS or Proxy setting and I reviewed them many times, but I can still not solve the problem.

One weird thing I found is that 'sudo cdsw status -v' resulted in the following error when I configured 'HTTP Proxy' from 'Cloudera Manager > CDSW > Configuration > Cloudera Data Science Workbench (Service-Wide)':

{metric: http_port_check, status: FAILURE, report_to_cm: True, timestampMs: 1538973062774, message: HTTP port check timed out after 60 seconds.}

The error was cleared when I deleted the HTTP Proxy configuration, but I could not do the 'docker run' command.

I am using **bleep**.io for the wildcard DNS subdomain requirement.

Any suggestion would be appreciated.

[#2 · Accepted solution]

Hi,

We are tracking an issue where the docker daemon is not picking up the proxy configuration. The easiest workaround is to download the 'nvidia/cuda:9.2-devel-ubuntu16.04' image to your local computer and move the image with the docker save; scp; docker load commands to the CDSW host. After doing this the 'nvidia-docker run' should work.

Regards,

Peter