kubectl-uncordon

Mark a node as schedulable

$ kubectl uncordon [node_name]

$ kubectl uncordon [node_name1 node_name2 ...]

$ kubectl uncordon [node_name] --context [context_name]

$ kubectl uncordon [[-l|--selector]] [label_key]=[label_value]

$ kubectl uncordon [node_name] --dry-run=[none|server|client]

kubectl uncordon NODE|-l selector [options]

NODE
    Name of the node(s) to mark as schedulable (positional argument, supports multiple)

--dry-run=server|client|none
    Print objects that would be sent without applying changes

--field-selector='
    Filter nodes by field query (e.g., 'metadata.name=my-node')

--selector|-l=string
    Filter nodes by label selector (e.g., 'disktype=ssd')

--timeout=duration
    Timeout for operation (default 0s)

The kubectl uncordon command marks one or more Kubernetes nodes as schedulable, reversing the effects of kubectl cordon. When a node is cordoned, its spec.unschedulable field is set to true, preventing the scheduler from placing new pods on it. This is useful during maintenance to safely drain pods without disruption.

After completing maintenance tasks like OS upgrades or hardware fixes, run kubectl uncordon to allow workloads to resume on the node. It supports targeting specific nodes by name, label selectors, or field selectors. The command updates the node's status via the Kubernetes API server.

Key use cases include post-drain recovery, scaling clusters dynamically, or recovering from temporary node issues. It does not evict existing pods or remove custom taints—only the cordon unschedulable flag. Always verify node status with kubectl get nodes before and after.

Does not remove custom taints or node selectors; node must exist and be reachable via API; requires cluster-admin permissions.

kubectl uncordon node01
kubectl uncordon -l node-role.kubernetes.io/worker
kubectl uncordon my-node --dry-run=client -o yaml

Use kubectl describe node <name> to confirm Unschedulable: false after uncordon.

Introduced in Kubernetes v1.0 (2014) as core node management tool; enhanced with selectors in v1.2+ for multi-node operations.

kubectl cordon(1), kubectl drain(1), kubectl get nodes(1), kubectl taint(1)