Argo Workflows Tolerations Quick Fix - Taint Tolerance Config

Argo Workflows tolerations allow pods to schedule on tainted nodes. Incorrect toleration syntax prevents pod scheduling on dedicated or specialized nodes. This guide covers the fix.

Quick Fix

Wrong

tolerations:
- key: dedicated
  operator: Equal
  value: gpu

Missing effect: NoSchedule parameter. The toleration is incomplete and Kubernetes may ignore it. Pods cannot schedule on nodes with the dedicated=gpu:NoSchedule taint.

Right

tolerations:
- key: dedicated
  operator: Equal
  value: gpu
  effect: NoSchedule
- key: nvidia.com/gpu
  operator: Exists
  effect: NoExecute
  tolerationSeconds: 300

# Expected output after applying the fix
# Pod tolerates dedicated=gpu:NoSchedule taint
# Pod tolerates nvidia.com/gpu:NoExecute for 300s
# Pod scheduled on GPU-enabled node
# Status: Running

Prevention

• Always specify effect in tolerations (NoSchedule, PreferNoSchedule, NoExecute)
• Use operator: Exists when the value doesn't matter
• Use tolerationSeconds for NoExecute tolerations
• Verify node taints with kubectl describe nodes | grep Taints
• Test scheduling with kubectl apply --dry-run=server

DodaTech Tools

Doda Browser's toleration checker validates tolerations against actual node taints. DodaZIP archives toleration configurations for cluster documentation. Durga Antivirus Pro identifies overly permissive tolerations.

FAQ