Resize persistent volume to increase size

This article explains how to resize an existing persistent volume to increase its size by editing the PersistentVolumeClaim (PVC) object.

Note

Resizing PVCs using this method only works your StorageClass supports AllowVolumeExpansion=True.

When you deploy an Azure Arc enabled SQL managed instance, you can configure the size of the persistent volume (PV) for data, logs, datalogs, and backups. The deployment creates these volumes based on the values set by parameters --volume-size-data, --volume-size-logs, --volume-size-datalogs, and --volume-size-backups. When these volumes become full, you will need to resize the PersistentVolumes. Azure Arc enabled SQL Managed Instance is deployed as part of a StatefulSet for both General Purpose or Business Critical service tiers. Kubernetes supports automatic resizing for persistent volumes but not for volumes attached to StatefulSet.

Following are the steps to resize persistent volumes attached to StatefulSet:

  1. Scale the StatefulSet replicas to 0
  2. Patch the PVC to the new size
  3. Scale the StatefulSet replicas back to the original size

During the patching of PersistentVolumeClaim, the status of the persistent volume claim will likely change from: Attached to Resizing to FileSystemResizePending to Attached. The exact states will depend on the storage provisioner.

Note

Ensure the managed instance is in a healthy state before you proceed. Run kubectl get sqlmi -n <namespace> and check the status of the managed instance.

1. Scale the StatefulSet replicas to 0

There is one StatefulSet deployed for each Arc SQL MI. The number of replicas in the StatefulSet is equal to the number of replicas in the Arc SQL MI. For General Purpose service tier, this is 1. For Business Critical service tier it could be 1, 2 or 3 depending on how many replicas were specified. Run the below command to get the number of StatefulSet replicas if you have a Business Critical instance.

kubectl get sts --namespace  <namespace>

For example, if the namespace is arc, run:

kubectl get sts --namespace arc

Notice the number of stateful sets under the READY column for the SQL managed instance(s).

Run the below command to scale the StatefulSet replicas to 0:

kubectl scale statefulsets <statefulset> --namespace <namespace> --replicas= <number>

For example:

kubectl scale statefulsets sqlmi1 --namespace arc --replicas=0

2. Patch the PVC to the new size

Run the below command to get the name of the PersistentVolumeClaim which needs to be resized:

kubectl get pvc --namespace <namespace>

For example:

kubectl get pvc --namespace arc

Once the stateful StatefulSet replicas have completed scaling down to 0, patch the StatefulSet. Run the following command:

$newsize='{\"spec\":{\"resources\":{\"requests\":{\"storage\":\"<newsize>Gi\"}}}}'
kubectl patch pvc <name of PVC> --namespace <namespace> --type merge --patch $newsize

For example: the following command will resize the data PVC to 50Gi.

$newsize='{\"spec\":{\"resources\":{\"requests\":{\"storage\":\"50Gi\"}}}}'
kubectl patch pvc data-a6gt3be7mrtq60eao0gmgxgd-sqlmi1-0 --namespace arcns --type merge --patch $newsize

3. Scale the StatefulSet replicas to original size

Once the resize completes, scale the StatefulSet replicas back to its original size by running the below command:

kubectl scale statefulsets <statefulset> --namespace <namespace> --replicas= <number>

For example: The below command sets the StatefulSet replicas to 3.

kubectl scale statefulsets sqlmi1 --namespace arc --replicas=3

Ensure the Arc enabled SQL managed instance is back to ready status by running:

kubectl get sqlmi -A

See also

Sizing Guidance