Docs

Volcengine Kubernetes Engine

FAQs

How do I handle VCI pod scheduling when a cluster contains subnets within multiple availability zones?

Problem description

When no VCI pod subnet is specified separately, the subnet of the VCI pod defaults to the pod CIDR of the VPC-CNI cluster it belongs to. Considering scenarios such as exhaustion of IP addresses in the pod subnet, VCI provides extended support: for VCI pod creation using the Kubernetes approach, additional pod subnets can be specified by using a pod annotation. For details, see Creating instances with specified subnets.

vke.volcengine.com/preferred-subnet-ids vke.volcengine.com/preferred-subnet-ids annotation are in a different availability zone (for example, Subnet-B-1 in availability zone B) from the subnets specified for the VPC-CNI cluster where this VCI pod is located (for example, Subnet-A-1 in availability zone A), the cluster has subnets from two different availability zones. In the VPC-CNI cluster with subnets from two different availability zones, when a new VCI pod is created with no subnets specified for the VCI pod, there is a 50% chance that the pod will be scheduled to Subnet-B-1 in availability zone B, where the VCI pod will fail to have an IP address.

Solution

For the above problem, solutions are provided for two scenarios:

Schedule the new VCI pod to a subnet in the new availability zone (not the one specified when creating the cluster):

The VCI pod must have the subnet specified for it by using the vke.volcengine.com/preferred-subnet-ids annotation. For details, see Creating instances with specified subnets.

Schedule the new VCI pod to the subnet in the availability zone specified when creating the cluster:

When creating the VCI pod, use node affinity to schedule the VCI pod to the specified virtual node (VirtualNode). To do so, follow the steps below:

In the left navigation bar, click Cluster.

On the Cluster page, click the name of the target cluster.

In the left navigation bar of the cluster management page, click Nodes > Virtual node. On the page that appears, find the target virtual node and click its name.

On the Label tab of the Node details page, find the cluster.vke.volcengine.com/node-name key and obtain the corresponding value.

When you create a VCI pod (that is, when you create a workload and choose to deploy it on VCI), configure node affinity with the cluster.vke.volcengine.com/node-name value obtained in the previous step so that the VCI pod is scheduled to the specified virtual node. For details, see Workloads.

tip

VCI does not support DaemonSets. For more information, see Limits.

Last updated: 2026.07.24 16:18:05