add devcontainer prebuild action

[pstjohn]

No description provided.

[jstjohn]

Should this be: nvcr.io/nvidian/cvai_bnmo_trng/bionemo-devcontainer?

[jstjohn]

I don't know the right answer to be clear, I am curious about this vs the target defined in prebuild.

[pstjohn]

Right, so this isn't quite ready to get merged, mainly testing CI at this point. But eventually, the hope would be that the .devcontainer/devcontainer.json (i.e., the root one) would point to the pre-built image, while this one inside the .github folder would build from the included top-level dockerfile. So everytime a new commit got pushed to main, we'd build a new latest tag, and it would update what folks could quickly pull when they launched the devcontainer setup

[pstjohn]

Ah ok -- the NGC connectivity might be a red herring: from the detailed job jobs:

2024-08-02T20:03:44.3417339Z About to run devcontainer build --workspace-folder /home/runner/work/bionemo-fw-ea/bionemo-fw-ea --config /home/runner/work/bionemo-fw-ea/bionemo-fw-ea/.github/.devcontainer/devcontainer.json --image-name nvcr.io/nvidian/cvai_bnmo_trng/bionemo-devcontainer:latest --cache-from nvcr.io/nvidian/cvai_bnmo_trng/bionemo-devcontainer:latest --cache-from nvcr.io/nvidian/cvai_bnmo_trng/bionemo-devcontainer
2024-08-02T20:03:44.5093633Z [2024-08-02T20:03:44.507Z] @devcontainers/cli 0.67.0. Node.js v18.20.4. linux 6.5.0-1025-azure x64.
2024-08-02T20:03:44.5094786Z 
2024-08-02T20:03:44.9533971Z [2024-08-02T20:03:44.952Z] Request 'https://nvcr.io/v2/nvidia/pytorch/manifests/24.02-py3' failed
2024-08-02T20:03:44.9535634Z 
2024-08-02T20:03:45.1555435Z [2024-08-02T20:03:45.154Z] Request 'https://nvcr.io/v2/nvidia/pytorch/manifests/24.02-py3' failed
2024-08-02T20:03:45.1556954Z 
2024-08-02T20:07:21.1697193Z ##[warning]You are running out of disk space. The runner will stop working when the machine runs out of disk space. Free space left: 0 MB
2024-08-02T20:07:21.9450465Z ##[error]Dev container build failed: Command failed: docker inspect --type image nvcr.io/nvidia/pytorch:24.02-py3 (exit code: undefined)
An error occurred building the container.
2024-08-02T20:07:21.9455273Z ##[error]Command failed: docker inspect --type image nvcr.io/nvidia/pytorch:24.02-py3
2024-08-02T20:07:21.9457720Z ##[endgroup]
2024-08-02T20:07:21.9618213Z Post job cleanup.

we might just be running out of disk space in the github runner

[pstjohn]

@ohadmo, is this something we could replicate in blossom?