Restrict use of pod spec.nodeName to prevent bypassing the scheduler (!3184) · Merge requests · Sylva-projects / sylva-core

Description

This merge request adds a native Kubernetes ValidatingAdmissionPolicy (VAP) to restrict Pod scheduling by preventing the use of spec.nodeName

Context

Allowing users to define nodeName can bypass the Kubernetes scheduler logic and lead to undesirable scheduling behavior.

This ValidatingAdmissionPolicy enforces the same scheduling constraints at the Kubernetes API level, using CEL expressions, and provides an additional layer of control independent of Kyverno.

Changes introduced

Added a ValidatingAdmissionPolicy targeting Pod creation
Added a ValidatingAdmissionPolicyBinding configured in Deny mode
Implemented CEL validations to detect spec.nodeName
Excluded Sylva system namespaces from validation
Scoped the policy to CREATE operations only

Excluded Namespaces

The sylva-managed namespaces are exempt from the restrict-node-selection policy. These may require explicit nodeName to ensure correct scheduling on dedicated infrastructure nodes, access node-local resources, or comply with constraints (imposed by Rancher itself) which manages their lifecycle independently from tenant workloads.

These namespaces are identified by this label:

sylva.org/sylva-managed

These namespaces are not subject to tenant-level placement restrictions and are trusted to manage their own scheduling constraints.

Behavior

Scenario	Result
Pod without `nodeName`	✅ Allowed
Pod with `nodeName`	❌ Rejected

References

#1731

cc @satyawanj @marc.bailly1

CI configuration

Below you can choose test deployment variants to run in this MR's CI.

Click to open to CI configuration

Legend:

Icon	Meaning	Available values
☁️	Infra Provider	`capd`, `capo`, `capm3`
🚀	Bootstrap Provider	`kubeadm` (alias `kadm`), `rke2`, `okd`, `ck8s`
🐧	Node OS	`ubuntu`, `suse`, `na`, `leapmicro`
🛠️	Deployment Options	Deployment option list and description
🎬	Pipeline Scenarios	Available scenario list and description
🟢	Enabled units	Any available units name, by default apply to management and workload cluster. Can be prefixed by `mgmt:` or `wkld:` to be applied only to a specific cluster type
🔴	Disabled units	Any available units name, by default apply to management and workload cluster. Can be prefixed by `mgmt:` or `wkld:` to be applied only to a specific cluster type
🏗️	Target platform	Can be used to select specific deployment environment Available platform list and description

Global config for deployment pipelines

autorun pipelines
allow failure on pipelines
record sylvactl events

Notes:

Enabling autorun will make deployment pipelines to be run automatically without human interaction
Disabling allow failure will make deployment pipelines mandatory for pipeline success.
if both autorun and allow failure are disabled, deployment pipelines will need manual triggering but will be blocking the pipeline

Be aware: after configuration change, pipeline is not triggered automatically. Please run it manually (by clicking the run pipeline button in Pipelines tab) or push new code.

Edited May 29, 2026 by Thomas Morin

Restrict use of pod spec.nodeName to prevent bypassing the scheduler