Add nix devcontainer Helm chart#3
Draft
hinriksnaer wants to merge 3 commits into
Draft
Conversation
hinriksnaer
added a commit
that referenced
this pull request
Jun 9, 2026
Helm chart for deploying nix-managed GPU dev environments on OpenShift. Uses a minimal Fedora + Nix container image with home-manager for declarative tool management. CUDA/PyTorch dev dependencies are installed at runtime via nixtorch, not baked into the image. - Parameterized deployment with --set flags (username, GPU type, settings repo) - GPU profiles: MIG 1g.18gb, MIG 2g.35gb, MIG 1g.18gb x2, full GPU, CPU-only - Persistent home directory (100Gi) and nix binary cache (50Gi) on NFS - SSH server (dropbear) for VS Code Remote access - values.schema.json for OpenShift console form rendering - Input validation (required username, supported track/GPU/pullPolicy enums) - PVCs preserved on helm uninstall via resource-policy annotation - README references existing create_dev_user.sh for secret creation stack-info: PR: #3, branch: hinriksnaer/stack/3
725e115 to
543a189
Compare
Helm chart for deploying nix-managed GPU dev environments on OpenShift. Uses a minimal Fedora + Nix container image with home-manager for declarative tool management. CUDA/PyTorch dev dependencies are installed at runtime via nixtorch, not baked into the image. - Parameterized deployment with --set flags (username, GPU type, settings repo) - GPU profiles: MIG 1g.18gb, MIG 2g.35gb, MIG 1g.18gb x2, full GPU, CPU-only - Persistent home directory (100Gi) and nix binary cache (50Gi) on NFS - SSH server (dropbear) for VS Code Remote access - values.schema.json for OpenShift console form rendering - Input validation (required username, supported track/GPU/pullPolicy enums) - PVCs preserved on helm uninstall via resource-policy annotation - README references existing create_dev_user.sh for secret creation stack-info: PR: #3, branch: hinriksnaer/stack/3
- Default storage class from nfs-rwx to ocs-storagecluster-cephfs; templatize PVC access mode instead of hardcoding ReadWriteMany - Default GPU from MIG 2g.35gb to full GPU (nvidia.com/gpu) with matching resource profile (4/16 CPU, 32Gi/32Gi memory) - Add hostNetwork toggle and RDMA device passthrough support - Add --track flag to create_dev_admin.sh and create_dev_user.sh; nix track skips workspace PVC, ConfigMaps, and static deployments - NOTES.txt documents cluster-reader role and hostnetwork SCC prereqs
8c75632 to
fa9d2a4
Compare
Storage class and access mode are no longer hardcoded in chart defaults. Instead, create_dev_user.sh mirrors the same detection order as create_dev_admin.sh (CephFS > LVMS > NFS) and passes the result to helm install. Chart values default to empty strings, requiring the provisioning script or explicit --set overrides.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Helm chart for deploying nix-managed GPU dev environments on OpenShift.
Uses a minimal Fedora + Nix container image with home-manager for
declarative tool management. CUDA/PyTorch dev dependencies are installed
at runtime via nixtorch, not baked into the image.