feat(runtimes): add validation for reserved MPI environment variables

[adiprathapa]

Picking up #3145 per @andreyvelich's ask. The original PR was approved in approach but went stale waiting on a rebase, so this is that rebase.

The MPI plugin's Validate function already catches a few things; this adds a check that rejects any TrainJob.spec.trainer.env entry whose name is in a new MPIReservedEnvNames set. The set lives in pkg/constants/constants.go and covers the four OpenMPI variables the operator manages internally (OMPI_MCA_orte_default_hostfile, OMPI_MCA_plm_rsh_args, OMPI_MCA_orte_keep_fqdn_hostnames, OMPI_MCA_orte_set_default_slots). Approach mirrors the Torch plugin's TorchRunReservedEnvNames check.

Tests in mpi_test.go cover one reserved var, two reserved vars, and a custom var that still passes.

Rebase notes: minor conflicts in constants.go (master added XGBoostReservedEnvNames at the same insertion point) and mpi.go (master dropped an unused intstr import). go vet, gofmt, and the package tests all pass locally.

Fixes #3126

Co-authored-by: Vishal Painjane painjanevishal2204@gmail.com

[github-actions]

🎉 Welcome to the Kubeflow Trainer! 🎉

Thanks for opening your first PR! We're happy to have you as part of our community 🚀

Here's what happens next:

• If you haven't already, please check out our Contributing Guide for repo-specific guidelines and the Kubeflow Contributor Guide for general community standards.
• Our team will review your PR soon! cc @kubeflow/kubeflow-trainer-team

Join the community:

• Slack: Join our #kubeflow-trainer Slack channel.
• Meetings: Attend the Kubeflow AutoML and Training Working Group bi-weekly meetings.

Feel free to ask questions in the comments if you need any help or clarification!
Thanks again for contributing to Kubeflow! 🙏

[Copilot]

Pull request overview

Adds server-side validation in the MPI runtime plugin to prevent users from setting OpenMPI environment variables that the operator manages internally, avoiding misconfiguration that can break MPI worker discovery.

Changes:

• Introduces constants.MPIReservedEnvNames (set of OpenMPI-reserved env var names).
• Extends MPI plugin Validate to reject TrainJob.spec.trainer.env entries that use any reserved name.
• Adds unit tests covering single reserved env, multiple reserved envs, and a non-reserved env case.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
pkg/runtime/framework/plugins/mpi/mpi.go	Rejects TrainJobs that set operator-managed OpenMPI env vars in spec.trainer.env.
pkg/runtime/framework/plugins/mpi/mpi_test.go	Adds validation test cases for reserved and non-reserved MPI env vars.
pkg/constants/constants.go	Defines the MPIReservedEnvNames set used by validation.

[andreyvelich]

Thanks for this @adiprathapa!
/ok-to-test
/lgtm
/approve

[google-oss-prow]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: andreyvelich

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [andreyvelich]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment