ERROR while pulling images

[SolenoidWGT]

Hi，everyone I encountered some problems when running the pyxis test. It reported an error when I pulled the pytorch 20.02 image. The reason for the error was "curl: (22) The requested URL returned error: 401 Unauthorized". However I can see that “Authentication succeeded”. Why does this problem occur, am I missing some configuration?

Here is the error log:

✗ nvcr.io PyTorch 20.02 (from function run_srun' in file tests/./common.bash, line 32, in test file tests/docker_image.bats, line 22) run_srun --container-image=nvcr.io#nvidia/pytorch:20.02-py3 sh -c 'echo $PYTORCH_VERSION'' failed

• srun -N1 --oversubscribe --container-image=nvcr.io#nvidia/pytorch:20.02-py3 sh -c echo $PYTORCH_VERSION pyxis: importing docker image: nvcr.io#nvidia/pytorch:20.02-py3 slurmstepd: error: pyxis: child 129103 failed with error code: 1 slurmstepd: error: pyxis: failed to import docker image slurmstepd: error: pyxis: printing enroot log file: slurmstepd: error: pyxis: [INFO] Querying registry for permission grant slurmstepd: error: pyxis: [INFO] Authenticating with user: slurmstepd: error: pyxis: [INFO] Authentication succeeded slurmstepd: error: pyxis: [INFO] Fetching image manifest list slurmstepd: error: pyxis: [INFO] Fetching image manifest slurmstepd: error: pyxis: [INFO] Downloading 59 missing layers... slurmstepd: error: pyxis: curl: (22) The requested URL returned error: 401 Unauthorized slurmstepd: error: pyxis: couldn't start container slurmstepd: error: pyxis: if the image has an unusual entrypoint, try using --no-container-entrypoint slurmstepd: error: spank: required plugin spank_pyxis.so: task_init() failed with rc=-1 slurmstepd: error: Failed to invoke spank plugin stack srun: error: XXXXXXX: task 0: Exited with exit code 1
• exit status: 1

Here is my system info: centos 7 enroot 3.4.0 slurm 20.11.8-6 pyxis 0.12.0 Docker 20.10.12

Thank you all

[flx42]

Hello,

It might have been a temporary issue with the NVIDIA container registry at nvcr.io, is it still happening with 100% frequency?

You could also check if you have more luck with the very latest pytorch image:

$ enroot import docker://nvcr.io#nvidia/pytorch:22.01-py3

[SolenoidWGT]

Hello,

It might have been a temporary issue with the NVIDIA container registry at nvcr.io, is it still happening with 100% frequency?

You could also check if you have more luck with the very latest pytorch image:

$ enroot import docker://nvcr.io#nvidia/pytorch:22.01-py3

Thank you very much, the problem should be related to the network configuration on my system, I will deal with it later. And I can use enroot to download the container mirror directly, and then use the absolute path to point out the container mirror location in pyxis.