Monitoring storage capacity

You can monitor the total usable space available on Storage Nodes to ensure that the StorageGRID Webscale system does not run out of storage space. This information is available at the deployment level, at the site level, and at the grid node level. You can also monitor the amount of space available for object metadata on object store 0 of each Storage Node.

Before you begin

You must be signed in to the Grid Management Interface using a supported browser.

Steps

  1. Select Dashboard.
  2. Review Available Storage information.

    screenshot of Grid Management Inteface dashboard
  3. Select Grid, and select any Storage Node > LDR > Storage > Overview > Main page.

    screenshot showing LDR > Storage > Overview > Main
  4. Monitor the Total Usable Space (STAS) attribute over time to estimate the rate at which usable storage space is being consumed.
    Usable space is the actual real amount of storage space available to store objects. For more information about storage space, see the Administrator Guide.

    To maintain normal system operations, you have to add Storage Nodes, add storage volumes, or archive object data before the storage disks’ usable space is consumed.

    Example
    In the following example, usable storage space is being consumed at a rate of approximately 4% per month, which means that there are eight months left before this Storage Node runs out of space.
    screenshot showing chart of total usable space over time
  5. Select Grid, and select any Storage Node > DDS > Datastore page.
  6. On the Overview tab, note the current value for the attribute Data Load (Percent) (CDLP).
    By default, 2 TB of space is reserved for the Cassandra database on the first storage volume of each Storage Node. Cassandra requires a certain amount of free storage space to perform essential operations such as compaction and repair. If the metadata load grows above 60% capacity, these Cassandra operations will be impacted. You must expand the StorageGRID Webscale system by adding Storage Nodes. Once storage is added, the system automatically rebalances object metadata across all Storage Nodes, and the alarm clears.