Home / OCI

OCIRunbookAdvanced

OCI Full Stack Disaster Recovery: protection groups, DR plans and drills

Orchestrate cross-region failover of compute, volumes, databases and load balancers with OCI Full Stack Disaster Recovery, using DR protection groups, generated plans, custom steps and prechecks.

Deepak Bhardwaj2 min read

Applies toOCI Full Stack Disaster Recovery (FSDR) with Compute, Block Volume groups, Base DB / Autonomous Database Data Guard, Load Balancer

Concepts

Term Meaning
DR protection group (DRPG) A set of resources in one region that fail over together (instances, volume groups, databases, LBs, file systems, OKE...)
Peer The matching DRPG in the standby region
DR plan Ordered plan generated by FSDR: switchover, failover, start drill, stop drill
Plan group / step Built-in steps plus your custom steps (scripts on instances, Functions)

1. Prerequisites

  • Data replication in place: volume groups with cross-region replication, Data Guard for databases (FSDR orchestrates the role change but doesn't create the Data Guard config for Base DB).
  • DR region networking: VCN, subnets, NSGs, LB, DNS ready.
  • Object Storage bucket in each region for FSDR logs.
  • IAM policies for FSDR to manage the member resources.

2. Create protection groups

oci disaster-recovery dr-protection-group create --compartment-id <cmp> \
  --display-name drpg-payments-mum --log-location '{"bucket":"fsdr-logs-mum","namespace":"<ns>"}'

Add members (compute instances with their target VNIC mappings, volume groups, databases, load balancers), then create the standby DRPG in Hyderabad and associate them (primary ↔ standby).

3. Generate DR plans

In the standby DRPG: create plans of type Switchover, Failover, Start drill and Stop drill. FSDR generates built-in steps in dependency order (for example database role change → volume group activation → instance launch → LB backend update).

Add custom steps:

  • Before: stop the application cleanly (switchover only).
  • After: run a smoke test script, update a DNS record, send a notification.

4. Run prechecks

oci disaster-recovery dr-plan-execution create --plan-id <switchover-plan-ocid> \
  --execution-options '{"planExecutionType":"SWITCHOVER_PRECHECK"}' --display-name precheck-2026-10-22

Fix every warning before the real run. Prechecks catch missing policies, wrong mappings and replication problems.

5. Drill

Start drill creates the stack in the standby region from replicated data without stopping production (isolated network recommended). Validate, record timings, then Stop drill to clean up.

6. Real failover

  1. Declare disaster.
  2. Execute the Failover plan from the standby region (doesn't need the primary to respond).
  3. Watch step progress; FSDR pauses on failed steps so you can fix and resume.
  4. Switch traffic (DNS / Traffic Management steering policy).
  5. Afterwards, reverse replication and plan a switchover back.

Monitoring

Plan execution logs in the bucket; Events for plan execution state; regular prechecks scheduled (for example weekly) so configuration drift is caught before a disaster.

Tip

Keep custom step scripts in Git and idempotent. A step that works only once ("append line to config") breaks the second drill.

Related fixes