You need to enable JavaScript to run this app.
Volcengine Kubernetes Engine

Volcengine Kubernetes Engine

Copy page
Download PDF
Storage FAQs
How do I provision cloud disk PVs when a cluster contains nodes of various ECS instance specifications?
Copy page
Download PDF
How do I provision cloud disk PVs when a cluster contains nodes of various ECS instance specifications?
Issue description​
In a VKE cluster, worker nodes are fully compatible with ECS instances of new specifications. For details, see Product announcement.​
A cluster may contain ECS nodes of various specifications. In such cases, the compatibility between different ECS specifications and cloud disk types is as follows:​
Warning​
  • VKE no longer supports creating StorageClasses, PersistentVolumes, and PersistentVolumeClaims of the Performance SSD (PTSSD) type. Existing PTSSD resources are not affected and can continue to be used until they are deleted.​
  • EBS no longer supports creating PTSSD volumes. Therefore, exercise caution when choosing instance families that are still available for sale and only support mounting PTSSD volumes. Existing PTSSD volumes are not affected.​
  • The following ECS instance families support mounting only PTSSD volumes: General-purpose g1, Compute-optimized c1, Memory-optimized r1, Big Data d1s, GPU-accelerated g1t, and High Performance Computing GPU hpcg1v.​
tip​
  • ECS specifications that support mounting only ESSD volumes: Except for the specifications listed above, all other ECS instance specifications supported by VKE use ESSD volumes.​
tip​
  • You can view the list of ECS instance specifications supported by VKE. To access it, log in to the VKE console and go to the Create node pool page of the target cluster.​
Impacts​
When ECS nodes of multiple specifications are present in a cluster, if cloud disks are not used as described in this topic, the following issues may occur:​
For example, you create a static PersistentVolume (PV) of the ESSD type and associate this PV with a PersistentVolumeClaim (PVC). When you later create a workload (such as a Deployment), you mount and use this PVC. In this case, without any node affinity scheduling configuration, the pod corresponding to the workload may be scheduled to an older-specification ECS node in the cluster. Since the volume cannot be properly mounted to the older-specification ECS instance, the created pod will remain in the Pending state and fail to run normally.​
Solution​
This part demonstrates how to create a Deployment. The procedure for creating other workloads is similar.​
  1. Log in to the VKE console.​
  1. In the left navigation bar, click Cluster. On the page that appears, find the target cluster and click its name.​
  1. In the left navigation bar of the cluster management page, click Nodes > Node to view the specifications of individual nodes.​
  1. Create a Deployment by following the steps described in Creating Deployments.​
  • Configure the parameters below as described:​
  • Basic information​
  • Configure the parameters according to your requirements. For parameter details, see Creating Deployments.​
  • Container configuration​
  • In the Storage configuration section, select and mount the target PVC. For other parameters, see Creating Deployments.​
  • Advanced settings​
  • In the Scheduling policy section, enable Node affinity scheduling to schedule the workload that mounts the PV to a node of the supported specification. For more configuration instructions, see Creating Deployments.​
  • If you select and mount a PV of the ESSD type in the Storage configuration section in the Container configuration step, add the new specifications obtained in Step 3, such as ecs.c1ie.large and ecs.g1ie.8xlarge.​
  • If you select and mount a PV of the PTSSD type in the Storage configuration section in the Container configuration step, add the old specifications obtained in Step 3, such as ecs.c1.large and ecs.g1.large.​
​
Last updated: 2026.07.23 14:31:46