SonicBase provides multiple layers of reliability. When a single server dies, requests failover to another replica without any noticeable interruption of service. With this layer of reliability you can have multiple servers go down without interruption.

With redundant servers, SonicBase supports a rolling restart of the servers which allows all servers in the cluster to be restarted without experiencing any interruption of service.

Replication Factor

The administrator of the SonicBase cluster can choose the replication factor to use. For redundancy you must choose a replication factor of a least 2. A replication factor of 3 or more is recommended. A replication factor of 3 means that each piece of data is stored on 3 servers for redundancy.


All inserts, updates and deletes are saved in a disk-based snapshot. Additionally, inserts, updates and deletes are flushed to disk in the update logs. If a server dies and is restarted, the snapshots and logs are loaded into memory and reconstitute the in-memory state without loss of data.