Free Up Disk Space When Writes Are Forbidden

This procedure allows you to free up disk space on an ARTESCA cluster where write operations are currently forbidden due to high disk usage.

When data disk usage equals or exceeds 95%, ARTESCA refuses write operations if writeProtectionEnabled is active. This protects system stability by preventing the cluster from reaching full capacity. To resolve the situation, plan a capacity expansion and/or delete data to bring disk usage back below 95%.

Important

Some applications require a write operation to perform a deletion. For example, deleting an object in a versioned S3 bucket requires writing a delete marker. In such cases, write protection must be temporarily disabled to allow the delete operations to proceed.

Re-enable write protection as soon as the delete operations are complete.

Prerequisites

  • SSH access to the ARTESCA bootstrap node is required to perform this procedure.

  • Platform Administrator privileges are required to perform this procedure.

Procedure

  1. Disable write protection to allow the delete operations to proceed.

    Warning

    With write protection disabled, the cluster can accept writes up to full capacity, which may cause system instability. Limit write activity strictly to the delete operations required to free up space.

    From the ARTESCA bootstrap node, run the following command:

    [artesca-os@bootstrap ~]$

    sudo kubectl --kubeconfig=/etc/kubernetes/admin.conf \
    patch storageservice artesca-storage-service \
    --namespace xcore --type=merge \
    --patch='{"spec":{"proxy":{"writeProtectionEnabled":false}}}'

  2. Perform the necessary delete operations using your application.

  3. Ensure that disk usage has dropped below 95% before re-enabling write protection. In the ARTESCA UI, verify that the ClusterCapacityCritical alert is no longer active on the Alerts page. For more information on alerts, refer to Alert Management.

    Note

    Disk usage does not drop immediately after delete operations. The freed space is only reclaimed once a relocation pass has completed. Allow some time before checking the alert status.

  4. Re-enable write protection to restore the stability safeguard.

    From the ARTESCA bootstrap node, run the following command:

    [artesca-os@bootstrap ~]$

    sudo kubectl --kubeconfig=/etc/kubernetes/admin.conf \
    patch storageservice artesca-storage-service \
    --namespace xcore --type=merge \
    --patch='{"spec":{"proxy":{"writeProtectionEnabled":true}}}'