Skip to content

Automate filesystem CSI driver installation, tests, and cleanup#879

Merged
aaronbfagan merged 2 commits intomainfrom
feature/automate-fs-csi-driver-install
Mar 23, 2026
Merged

Automate filesystem CSI driver installation, tests, and cleanup#879
aaronbfagan merged 2 commits intomainfrom
feature/automate-fs-csi-driver-install

Conversation

@aaronbfagan
Copy link
Copy Markdown
Collaborator

@aaronbfagan aaronbfagan commented Mar 17, 2026

Release Notes (Mandatory Description)

Automates Nebius Shared Filesystem CSI installation in Terraform whenever a shared filesystem is present, whether the filesystem is created by Terraform or provided externally.

Also converts the helper workflow into a validation-only flow under k8s-training/filesystem-csi-validation:

  1. verify the node mount
  2. run a default-StorageClass smoke test
  3. run a cross-node RWX test
  4. clean up validation resources

What changed

  • installs the Nebius Shared Filesystem CSI driver automatically from Terraform
  • promotes csi-mounted-fs-path-sc to the default StorageClass by default
  • uses a single shared filesystem mount path for both cloud-init and CSI
  • improves plan-time validation for non-MIG platforms and unsupported GPU cluster presets
  • renames filestore-csi-setup to filesystem-csi-validation
  • removes manual CSI install from the helper workflow
  • updates the validation tests to verify default StorageClass inheritance

Validation

Tested successfully with:

  • a shared filesystem created by Terraform
  • a preexisting shared filesystem provided by ID

- install Nebius Shared Filesystem CSI automatically when a shared filesystem is present
- support both Terraform-created and externally provided filesystem IDs
- patch the CSI StorageClass as default by default
- unify the shared filesystem mount path across cloud-init and CSI
- improve plan-time validation for MIG and GPU cluster compatibility
- simplify the helper workflow into validation-only scripts
- rename filestore-csi-setup to filesystem-csi-validation
@aaronbfagan aaronbfagan force-pushed the feature/automate-fs-csi-driver-install branch from ab1a9cb to 5f494c1 Compare March 19, 2026 17:22
@aaronbfagan aaronbfagan had a problem deploying to project-e00pjzzrtk1fs3yavy March 19, 2026 17:23 — with GitHub Actions Error
@aaronbfagan aaronbfagan had a problem deploying to project-e00pjzzrtk1fs3yavy March 21, 2026 12:10 — with GitHub Actions Error
@aaronbfagan aaronbfagan had a problem deploying to project-e00pjzzrtk1fs3yavy March 21, 2026 12:10 — with GitHub Actions Error
Copy link
Copy Markdown
Collaborator

@dasbatta dasbatta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed the outdated edits on tfvars and rest of the PR looks good. Discussed it over the zoom and Thank you for automating the FS mount.

Copy link
Copy Markdown
Collaborator

@aaroniscode aaroniscode left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Did a quick review and talked to @aaronbfagan , I think this looks good

@aaronbfagan aaronbfagan merged commit 7dbbdc6 into main Mar 23, 2026
3 of 5 checks passed
@aaronbfagan aaronbfagan deleted the feature/automate-fs-csi-driver-install branch March 23, 2026 16:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants