Not Enough Available Disk Space For The Current Import Operation

News Knowledge Base
1

How to increase the storage size of a Data Labeling session when getting an error of "Not enough available disk space for the current import operation. Current import data after processing takes x.x MB, but available space is x.x MB. ?

Checks To Be Performed Before Increasing Storage Size:

  1. Check whether Data Manager (Document Manager) is standalone or if it is in AI Center.
  2. If using DM in AIC, check which version it is, in order to see if the issue was caused because of using a version that is in private preview (which imposes a hard limit on how many docs can be imported). Read more on Data Manager .
  3. Check the the number of documents that are being imported. For AI Center version 21.4 or earlier, it is recommended only for Trial or Demo scenarios involving datasets of less than 500 images. (In the case of 21.4 or earlier versions, either an upgrade to 21.10 or installing Data Manager Standalone is needed)


However, before upgrading to 21.10 is completed or before the standalone version is installation is completed, as a workaround, the storage size of a data labeling session can be increased in order to alleviate the size limit on imports.

Increasing the Storage Size: By default, all labeling sessions are allocated a PVC with 5GB storage capacity. In the exceptional case of needing to increase the storage size, please follow the below:

  1. Identify the namespace under which the labeling session exists. It'll follow the naming convention of data-manager-
  • kubectl get ns | grep "data-manager"
  1. Once the namespace is present, stop the pod running for the labeling session. Get the sessionId (id) of the labeling session and use the below command to stop the pod. The first of the two commands will get the deployment name.
  • kubectl -n get deployment | grep ""
  • kubectl -n edit deployment
  1. Once in Edit mode, change the replicas from 1 to 0. Save the changes (ESC + :wq). This will kill any running pod.

  1. Increase the PVC size. PVC name will be pvc-
  • kubectl -n edit pvc "pvc-"
  1. In the edit mode, increase the storage to the desired value. Save the changes (ESC + :wq). The screenshot below shows an entry edited from 5Gi to 6Gi.

  1. To confirm the size has increased, see the output of the below command to show the current status
  • kubectl -n get pvc -o yaml

The output will contain:

message: Waiting for user to (re-)start a pod to finish file system resize of volume on node.

status: "True"

type: FileSystemResizePending

  1. Now edit the same deployment and bring the replica count from 0 to 1. Save the changes.
  • kubectl -n edit deployment

  1. Once the deployment editing is complete, the pod should start running in about a minute. Once its status shows "Running", start using the labeling session.
  • kubectl -n get pods | grep "" .