How to recover from a full disk

The FortiAnalyzer-BigData data life cycle can be managed via Cluster Manager GUI (see Manage data lifecycle). If the data disk on your hosts begin to reach full capacity and are causing the Data Lake services to become unhealthy, you can follow the following steps:

1. From Cluster Manager > Services, manually Stop the Core service group.
2. Go to Cluster Manager > Data, expand the Root Storage Group and click Action > Manage Data Lifecycle.
3. In the Maximum Age field, reduce the number of days for storing data and click OK.
4. Go to Cluster Manager > Jobs, and locate and Run the Data Retention job in the job list.
5. Wait a for the Data Retention job to finish running.
6. From Cluster Manager > Services > Data Lake, manually Restart the Kudu service.
7. Check that the Kudu service has a healthy status.
8. If you still receive messages about the disk being full in the log, you might need to repeat steps 4-6.
9. Once you stop receiving messages, go to Cluster Manager > Services and manually Start the Core service group.