Repository navigation
✨ feat(sandbox): scheduler_name for the Job runtime pod template - #14662
Conversation
Accelerator clusters commonly gate device-requesting pods behind a custom scheduler via a validating admission webhook. On the Spyre dev cluster the Job create is denied outright: admission webhook "spyre-job-validator.k8s.io" denied the request: A pod must use "schedulerName: spyre-scheduler". SandboxJobConfig had no way to set it, so every runtime: job kick that requests ibm.com/spyre_pf failed before a pod existed. Add agents.<name>.sandbox.job.scheduler_name, passed through verbatim as the pod's schedulerName (empty = cluster default), with launcher and config round-trip tests and a doc line. Signed-off-by: Andrew Anderson <andy@clubanderson.com>
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
|
@clubanderson is auto-promoted for contributor-lane merging on this PR: the last 5 merged PRs by this author in this hive's repositories have no revert or sentinel-alert evidence. CI, reviewer verdict, diff fences, soak, holds and branch protection still apply. |
|
@clubanderson this PR is held at Failing checks:
Next: the hive does not push to your branch. Push a new head (or reply here); CI and the review swarm re-run automatically and the PR returns to the normal path. |
Settings parity requires every hive.yaml leaf to be reachable from the Settings dialog; add the Job scheduler name field next to the other sandbox Job fields and allowlist the per-agent override like its siblings. Signed-off-by: Andrew Anderson <andy@clubanderson.com>
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
Why
Accelerator clusters often require device-requesting pods to use a custom scheduler, enforced by a validating webhook. On the IBM Spyre dev cluster (
a1-vllm-spyre) everyruntime: jobkick that requestsibm.com/spyre_pfis denied at Job create:SandboxJobConfighad no way to setschedulerName, so the sandbox kick fails before a pod exists (visible only inaudit.jsonlassandbox_failed).What
agents.<name>.sandbox.job.scheduler_name(optional) →kubejob.Options.SchedulerName→ podschedulerName. Empty leaves the cluster default.kubejob_test.go(manifest) andsandbox_job_runtime_test.go(config→options).docs/sandbox-isolation.md, changelog.d entry.Additive; no behaviour change when unset.
Test
cd src/crux && go test ./kubejob/ ./config/ ./agent/— ok.