All Products
Search

This Product

    Container Service for Kubernetes:Perform online scale-out for FlexVolume disks in ACK

    Document Center

    Container Service for Kubernetes:Perform online scale-out for FlexVolume disks in ACK

    Last Updated:Mar 26, 2026

    As workloads grow, disk volumes mounted to pods can run out of space. Container Service for Kubernetes (ACK) supports online disk volume expansion using FlexVolume in Kubernetes 1.16 and later — no pod restart is required in most cases. This topic covers both expansion scenarios: expanding while the pod stays Running, and expanding after restarting the pod.

    Prerequisites

    Before you begin, make sure you have:

    • A kubeconfig file for your cluster and kubectl configured to connect to it. See Connect to an ACK cluster by using kubectl.

    • A dynamically provisioned persistent volume (PV) mounted via a persistent volume claim (PVC) that includes the storageClassName parameter.

    • The FlexVolume plugin at version v1.14.8.109-5f88ade-aliyun or later. Check the cluster components page in the ACK console to upgrade.

    Limitations

    Before you expand a disk volume, review the following constraints:

    Constraint Details
    Dynamically provisioned PVs only Only PVs provisioned dynamically with a StorageClass can be expanded online. Inline volumes and volumes backed by basic disks cannot be expanded without service interruption.
    Maximum disk size Disks of 2,000 GiB or smaller can be expanded without service interruption.
    Pod must be Running The pod must remain in Running state during expansion. If it is not Running, use the restart-based method instead.

    StorageClass requirement

    The StorageClass must have allowVolumeExpansion: true. ACK-created StorageClasses set this automatically. For manually created StorageClasses, set it explicitly:

    apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: alicloud-disk-available
provisioner: alicloud/disk
allowVolumeExpansion: true

    Back up your data

    Before expanding, create a disk snapshot to protect against data loss from errors during the expansion process.

    Grant the ResizeDisk permission to the RAM role

    Before expanding a disk volume, grant the ResizeDisk permission to the cluster's Master RAM role. Without this permission, the expansion request returns an authorization error.

    1. Log on to the ACK console.

    2. In the left-side navigation pane, click Clusters.

    3. Find the cluster and click Details in the Actions column.

    4. In the left-side navigation pane, click Cluster Information.

    5. Click the Cluster Resources tab, then click the hyperlink next to Master RAM Role.

    6. In the RAM console, grant the ResizeDisk permission to the role. For details, see Modify the document and description of a custom policy.

      resizedisk

    For ACK dedicated clusters using FlexVolume or ACK managed clusters, follow steps 1–4 above and click the hyperlink next to Master RAM Role.

    Expand a disk volume without restarting the pod

    Use this method when the pod is in Running state. The disk is resized online with no service interruption.

    Check the current state

    1. Verify the pod is Running:

      kubectl get pod

      Expected output:

      web-0   1/1   Running   0   42s

    2. Check the current disk usage inside the pod:

      kubectl exec web-0 -- df /data

      Expected output:

      Filesystem   1K-blocks    Used  Available  Use%  Mounted on
/dev/vdb      20511312   45080   20449848    1%  /data

      Note the current size: 20 GiB.

    3. Confirm the PVC is bound and check its current capacity:

      kubectl get pvc

      Expected output:

      NAME             STATUS   VOLUME                   CAPACITY   ACCESS MODES   STORAGECLASS              AGE
disk-ssd-web-0   Bound    d-wz9hpoifm43yn9zie6gl   20Gi       RWO            alicloud-disk-available   57s

    Expand the volume

    1. Patch the PVC to request the new storage size:

      kubectl patch pvc disk-ssd-web-0 -p '{"spec":{"resources":{"requests":{"storage":"30Gi"}}}}'

      Wait about one minute for the expansion to complete.

    Verify the expansion

    1. Confirm the PV is updated to 30 GiB:

      kubectl get pv d-wz9hpoifm43yn9zie6gl

      Expected output:

      NAME                     CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS   CLAIM                    STORAGECLASS              REASON   AGE
d-wz9hpoifm43yn9zie6gl   30Gi       RWO            Delete           Bound    default/disk-ssd-web-0   alicloud-disk-available            5m23s

    2. Confirm the PVC reflects the new capacity:

      kubectl get pvc

      Expected output:

      NAME             STATUS   VOLUME                   CAPACITY   ACCESS MODES   STORAGECLASS              AGE
disk-ssd-web-0   Bound    d-wz9hpoifm43yn9zie6gl   30Gi       RWO            alicloud-disk-available   5m10s

    3. Verify the new disk size is visible inside the pod:

      kubectl exec web-0 -- df /data

      Expected output:

      Filesystem   1K-blocks    Used  Available  Use%  Mounted on
/dev/vdb      30832548   45036   30771128    1%  /data

      The disk volume has been expanded from 20 GiB to 30 GiB with no downtime.

    Restart the pod to expand the disk volume

    Use this method when the file system does not reflect the new size until the pod is restarted, or when the pod is not in Running state.

    Check the current state

    1. Verify the pod status and note the current disk size:

      kubectl get pod

      Expected output:

      web-0   1/1   Running   0   42s

    2. Check the current disk usage inside the pod:

      kubectl exec web-0 -- df /data

      Expected output:

      Filesystem   1K-blocks    Used  Available  Use%  Mounted on
/dev/vdb      20511312   45080   20449848    1%  /data

    3. Confirm the PVC and PV are bound:

      kubectl get pvc
kubectl get pv

    Prevent the pod from rescheduling

    1. Find the zone label of the PV:

      kubectl get pv d-wz9g2j5qbo37r2lamkg4 -oyaml | grep failure-domain.beta.kubernetes.io/zone

      Expected output:

      failure-domain.beta.kubernetes.io/zone: cn-shenzhen-e

    2. Change the zone label to an invalid value so the pod cannot be scheduled:

      kubectl label pv d-wz9g2j5qbo37r2lamkg4 failure-domain.beta.kubernetes.io/zone=cn-shenzhen-e-nozone --overwrite

      Expected output:

      persistentvolume/d-wz9g2j5qbo37r2lamkg4 labeled

    3. Delete the pod to trigger a restart. The pod enters Pending state because the zone label no longer matches any node:

      kubectl delete pod web-0

      Verify the pod is Pending:

      kubectl get pod

      Expected output:

      web-0   0/1   Pending   0   27s

    Expand the volume

    1. Patch the PVC to request the new storage size:

      kubectl patch pvc disk-ssd-web-0 -p '{"spec":{"resources":{"requests":{"storage":"30Gi"}}}}'

    Restore the pod

    1. Restore the original zone label so the pod can be scheduled again:

      kubectl label pv d-wz9g2j5qbo37r2lamkg4 failure-domain.beta.kubernetes.io/zone=cn-shenzhen-e --overwrite

      Expected output:

      persistentvolume/d-wz9g2j5qbo37r2lamkg4 labeled

      Wait about one minute for the pod to restart and the expansion to complete.

    Verify the expansion

    1. Confirm the pod is Running:

      kubectl get pod

      Expected output:

      web-0   1/1   Running   0   3m23s

    2. Confirm the PVC reflects the new capacity:

      kubectl get pvc

      Expected output:

      disk-ssd-web-0   Bound   d-wz9g2j5qbo37r2lamkg4   30Gi   RWO   alicloud-disk-available   17m

    3. Confirm the PV reflects the new capacity:

      kubectl get pv d-wz9g2j5qbo37r2lamkg4

      Expected output:

      d-wz9g2j5qbo37r2lamkg4   30Gi   RWO   Delete   Bound   default/disk-ssd-web-0   alicloud-disk-available   17m

    4. Verify the new disk size is visible inside the pod:

      kubectl exec web-0 -- df /data

      Expected output:

      /dev/vdb   30832548   45036   30771128   1%   /data

      The disk volume has been expanded from 20 GiB to 30 GiB.

    Related topics