Docs

Volcengine Kubernetes Engine

Pod exception troubleshooting

How do I fix a failure in scheduling a pod in the Pending state?

Issue description

The pod remains in the Pending state for an extended period and is not scheduled.

Cause analysis

View pod event information to preliminarily determine the cause of the pending pod. For details, see Viewing pod information.

Event message example	Reason for scheduling failure	Solution
No nodes available to schedule pods	The pod fails to be scheduled onto a node because the cluster does not have available nodes.	No available nodes in the cluster
0/3 nodes are available: 3 Insufficient cpu. 0/2 nodes are available: 2 Insufficient memory.	The pod fails to be scheduled onto a node due to insufficient node resources (CPU and memory).	Insufficient node resources
0/1 nodes are available: 1 node(s) had taints that the pod didn't tolerate.	Nodes have taints that cannot be tolerated by the pod.	Intolerable node taints for the pod
0/2 nodes are available: 1 node(s) didn't match node selector, 1 node(s) didn't match pod affinity rules, 1 node(s) didn't match pod affinity/anti-affinity.	Node affinity, pod affinity, or pod anti-affinity configured on the workload to which the pod belongs do not match any existing node.	Pod affinity mismatch with nodes

Solution

No available nodes in the cluster

Click the name of the target cluster. In the left navigation bar of the cluster management page, click Node.

On the Node page, check whether there are available nodes (those in the Running state).

If no node exists, add new nodes or existing ECS instances to the cluster. For details, see Adding nodes and Adding existing nodes.

If the cluster has nodes but none of them are in the Running state, fix the abnormal node status. If nodes are available and have sufficient resources, the pod can be scheduled. For information about how to fix abnormal nodes, see Node and node pool FAQs.

Insufficient node resources

Click the name of the target cluster. In the left navigation bar of the cluster management page, click Node.

On the Node page, view the CPU and memory resource usage of available nodes (in the Running state).

If the number of nodes is too small, add new nodes to the cluster to ensure that node resources meet pod scheduling requirements. For details, see Adding nodes.

If the specifications of individual nodes are too small, add nodes with larger specifications to ensure sufficient CPU and memory resources. For details, see Adding nodes and Creating node pools.

Delete unnecessary pods on existing nodes to free up node resources. For details, see Deleting pods.

Adjust the resource quota (request and limit) of the container in the pod so that it can be scheduled onto a node. For details, see Workloads.

Intolerable node taints for the pod

Click the name of the target cluster. In the left navigation bar of the cluster management page, click Node.

On the Node details page, view the taints of the node and process the scheduling issue between the Pod and the node as needed.

Modify the scheduling policy of the workload to which the pod belongs to tolerate the taints. For details, see Workloads.

Batch update node taints by modifying the node pool configuration. For details, see Editing node pools. You can also edit or delete node taints individually.

tip

System taints cannot be deleted.

Pod affinity mismatch with nodes

Click the name of the target cluster. In the left navigation bar of the cluster management page, click Node.

On the Node page, find the target node, and in the Actions column on the right side, click ... > Label management to view node labels.

Based on node labels, adjust the scheduling policy of the workload corresponding to the pod. The following figure uses a Deployment as an example. For details, see Managing workloads. For more information about affinity, see the Kubernetes documentation.

Last updated: 2026.10.16 17:40:20