Consistency Summary

In most cases, replicas can continue to service read requests as long as the underlying hardware remains functional. In its default configuration, there is nothing that stops a replica from doing this, even if it is the only node running after some catastrophic failure.

However, is is possible for a replica to stop servicing read requests following a network failure, if the consistency policy requires either version information, or disallows stale data relative to the master. Whether this happens depends on how your Replication Nodes are exactly partitioned as a result of the failure, and how long it takes to establish a new master. The replica's ability to service read requests is also determined by the consistency policy in use for each request. If the read requires tight consistency with the master, and the master is not available to ensure the consistency can be met, then the read will fail.