30268547324
68%
master: 68%

Ran 27 Jul 2026 01:16PM UTC

Jobs 1

Files 42

Run time 1min

Badge

Embed ▾

Committed 27 Jul 2026 01:02PM UTC coverage: 68.256%. Remained the same

Build # 30268547324

Build Type

Pull #3809

github

Committed by

sanskar-singh-2403

Commit Message

fix(api): reject zero and negative numNodes and numProcPerNode

TrainJob.Spec.Trainer.NumNodes and NumProcPerNode had no lower-bound
validation, so values like numNodes: 0 or numProcPerNode: -5 were
accepted at every layer (CRD schema, validating webhook, runtime
plugins). The bad values then propagated into the pod spec, e.g. the
Torch plugin rendered PET_NPROC_PER_NODE=-5 and set the trainer PodSet
count to 0, failing at pod runtime instead of at admission.

Add +kubebuilder:validation:Minimum=1 to both fields and regenerate the
CRDs so the API server rejects non-positive values at admission. This
mirrors the FluxMLPolicySource.NumProcPerNode contract (self >= 1) and
covers all runtimes at once (Torch, PlainML, JAX, XGBoost, MPI).

A pre-existing Flux numProcPerNode < 1 integration case expected a
Forbidden error from the plugin webhook; the value is now rejected
earlier by the CRD schema, so its matcher is updated to Invalid.

Closes #3805

Signed-off-by: Sanskar Singh <sanskarsinghty1234@gmail.com>

Pull Request Pull Request #3809: fix(api): reject zero and negative numNodes and numProcPerNode

Coverage Stats

2563 of 3755 relevant lines covered (68.26%)

0.8 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
1	30268547324.1	27 Jul 2026 01:16PM UTC	42	68.26	GitHub Action Run

kubeflow / trainer / 30268547324
68%
master: 68%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 30268547324

kubeflow / trainer / 30268547324 68% master: 68%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 30268547324

kubeflow / trainer / 30268547324
68%
master: 68%

README BADGES
x