HOWTO - Check Volume Usage

Enterprise Applicable to Enterprise Edition

Community Applicable to Community Edition

As an operator, I may need to know the usage of a volume. This document describes how to check the usage of user volumes, group volumes, and PV dataset volumes. Before checking the volume usage, please make sure you can access your Kubernetes cluster by kubectl, we use kubectl exec <pod> -- df /path/to/the/volume to check the volume

Usage of User Volume

Because user volume can only be mounted by a notebook pod, we need to check the usage by the user's running notebook pod.

1. Find the user's notebook pod. If the pod is not launched, you cannot check the user volume usage.

   kubectl -n hub get pod
   

   Find the pod jupyter-<username>.

2. The user volume is mounted at /home/jovyan. Use the command df to check the usage.

   kubectl -n hub exec jupyter-<username> -- df -h /home/jovyan
   

   The sample output

   Filesystem                   Size  Used  Avail  Capacity    Mounted on
   /dev/rbd11                    49G   53M   47G   1%          /home/jovyan
   

Usage of Group Volume

Note: This method can only works if the configuration primehub.sharedVolumeStorageClass is not set

If the shared volume of a group is enabled, an NFS server is created as the group volume. To check the volume usage, we can check the usage in the NFS server pod.

1. Find the group NFS server pod

   kubectl -n hub get pod
   

   Find the pod nfs-project-<group>-0.

2. The disk is mounted at /exports

   kubectl -n hub exec <pod-name> --  df /exports
   

   The sample output

   Filesystem                   Size  Used  Avail  Capacity    Mounted on
   /dev/rbd11                    196G   3.9M   182G   3%       /exports
   

Usage of PV Dataset Volume

Note: This method only works if the configuration primehub.sharedVolumeStorageClass is not set

If a PV dataset is configured as auto provisioning, an NFS server is created as the dataset volume. To check the volume usage, we can check the usage in the NFS server pod.

1. Find the pod of the PV dataset NFS server

   kubectl -n hub get pod
   

   Find the pod nfs-dataset-<dataset>-0.

2. The disk is mounted at /exports

   kubectl -n hub exec <pod-name> --  df /exports
   

   The sample output.

   Filesystem                   Size  Used  Avail  Capacity    Mounted on
   /dev/rbd11                    196G   3.9M   182G   3%       /exports