Episode 2. DataGuard. Day 2 of your 10 day sprint.

Welcome back. Yesterday RAC kept nodes honest. Today DataGuard keeps data honest across sites. Say the key lines out loud as you hear them.

Concepts first. A physical standby receives redo and applies it. Transport lag means redo is not arriving; apply lag means redo arrives but is not applied. Standby redo logs must exist and be sized or SYNC shipping stalls. FSFO with the observer automates failover when thresholds trip.

Say this sentence: Switchover is planned and lossless, failover is unplanned and may lose data.

Now the scenario. The standby falls two hours behind before month-end close. Understand: transport versus apply lag from V dataguard stats, and when the gap started. Prepare: network throughput, MRP process state, disk rate on the standby. Identify: a stopped apply process versus slow storage; restart and measure before touching the network. Support: restart apply, clear the backlog, then add lag alerting so the class cannot recur.

Say this bridge once. Say: For a payments database I design for zero RPO with SYNC plus standby redo, and I monitor lag continuously because SYNC only protects while transport keeps up.

Your Day 2 lab: run the lag query from memory, then walk a switchover runbook aloud including the reinstate step after a failover.
Teaser: Tomorrow, Episode 3, GoldenGate. Extract pump replicat, and the loop risk in bi-directional replication.

Second scenario, fast rep. FSFO test fails to trigger. Understand: observer status and thresholds. Prepare: observer logs plus connectivity between sites. Identify: observer down versus threshold never met; check the observer first. Support: restart the observer, re-run the drill, record the drill as recurring evidence.

Pause and say the switchover sentence again: planned and lossless versus unplanned with possible loss. Say it twice.

Close with one STAR seed: a DataGuard save with an RPO number attached. You will need it on Power Day. Done.