Started docker container with Proxy configuration as shown below
docker run --runtime=nvidia -it --rm -e HTTP_PROXY=http://xxx:zzz@abc.corpnet2.com:800 -e HTTPS_PROXY=http://xxx:zzz@abc.corpnet2.com:800 -v /data01/tlExp:/workspace/tlt-experiments --shm-size=1g --ulimit memlock=-1 --ulimit stack=67108864 nvcr.io/nvtltea/med/tlt-annotate:v0.1-py3 /bin/bash
Inside the container ENV shows proxy information as below
TF_ADJUST_HUE_FUSED=1
CUDA_TOOLKIT_PATH=/usr/local/cuda
HOROVOD_GPU_ALLREDUCE=NCCL
PWD=/opt/nvidia
HOROVOD_WITHOUT_PYTORCH=1
CUDA_VERSION=10.0.130
HTTPS_PROXY=http://xxx:zzz@abc.corpnet2.com:800
OMPI_MCA_btl_vader_single_copy_mechanism=none
HOME=/root
SHLVL=1
NVIDIA_TENSORFLOW_VERSION=19.01
TF_ADJUST_SATURATION_FUSED=1
HTTP_PROXY=http://xxx:zzz@abc.corpnet2.com:800
NCCL_VERSION=2.3.7
OPENMPI_VERSION=3.1.3
BASH_ENV=/etc/bash.bashrc
PYTHONPATH=:/opt/nvidia
LESSOPEN=| /usr/bin/lesspipe %s
NCCL_HDR_PATH=/usr/include
TENSORFLOW_VERSION=1.12.0+
HOROVOD_NCCL_LINK=SHARED
LESSCLOSE=/usr/bin/lesspipe %s %s
_=/usr/bin/env
Actual ttl-pull listing models giving errors
root@407e224d5c77:/opt/nvidia# tlt-pull -k $API_KEY -lm -o nvtltea -t med
Traceback (most recent call last):
File “/usr/lib/python2.7/runpy.py”, line 174, in _run_module_as_main
“main”, fname, loader, pkg_name)
File “/usr/lib/python2.7/runpy.py”, line 72, in _run_code
exec code in run_globals
File “medical/ngc/pull.py”, line 78, in
File “medical/ngc/pull.py”, line 64, in main
File “/home/isaacy/workspace/repos/ai-infra/medical/ngc/controller.py”, line 52, in list_models
File “/usr/local/lib/python2.7/dist-packages/ngccli/controller/model.py”, line 36, in list
model_list_resp = self.model_api.list(org_name, team_name)
File “/usr/local/lib/python2.7/dist-packages/ngccli/api/models.py”, line 90, in list
auth_team=team_name)
File “/usr/local/lib/python2.7/dist-packages/ngccli/api/connection.py”, line 72, in make_api_request
team=auth_team)
File “/usr/local/lib/python2.7/dist-packages/ngccli/api/authentication.py”, line 110, in get_token
Authentication._get_token, base_url, app_key, org, team, scopes
File “/usr/local/lib/python2.7/dist-packages/ngccli/expiring_cache.py”, line 27, in setitem
OrderedDict.setitem(self, key, self._CacheValue(value))
File “/usr/local/lib/python2.7/dist-packages/ngccli/expiring_cache.py”, line 12, in init
self.value = callback()
File “/usr/local/lib/python2.7/dist-packages/ngccli/util/ngc_utils.py”, line 104, in call
return self._fn(*self._args, **self._kwargs)
File “/usr/local/lib/python2.7/dist-packages/ngccli/api/authentication.py”, line 91, in _get_token
scopes=scopes)
File “/usr/local/lib/python2.7/dist-packages/ngccli/api/authentication.py”, line 80, in __make_token_request
raise NgcException(msg)
ngccli.errors.NgcException: Url: https://authn.nvidia.com/token?scope=group/ngc:nvtltea&scope=group/ngc:nvtltea/med is not reachable.