What happens when a database server goes down? With a single instance, you lose data since the last backup. With Oracle Data Guard, you switch to standby and keep going.
RPO and RTO¶
The client needed an RPO under 5 minutes and RTO under 30 minutes. Unachievable with RMAN backup alone. Data Guard maintains a physical standby — an identical copy continuously applying redo logs.
Synchronous vs. Asynchronous¶
Synchronous: RPO = 0, but higher write latency. Asynchronous: minimal performance impact, RPO of seconds to minutes. For the client: asynchronous with RPO under 60 seconds.
Switchover and Failover¶
Switchover: planned transition, zero data loss, 30–60 seconds. Failover: unplanned during a disaster. The old primary must be re-synchronized afterward.
Monitoring and Testing¶
Nagios: apply lag, transport lag, standby process status. Alert when lag exceeds 5 minutes. Quarterly switchover test — verifying that DR actually works.
Conclusion¶
Data Guard is the enterprise standard for DR. Implementation requires an experienced DBA, but the result — RPO in seconds, RTO in minutes — is essential for critical systems.
Need help with implementation?
Our experts can help with design, implementation, and operations. From architecture to production.
Contact us