Files
hermes-recovery/references/4-DR-Testing-Schedule.md
T

144 lines
3.2 KiB
Markdown

IT PRO PARTNER — DOCUMENT 4 OF 4
DR Testing & Validation Schedule
The cadence that turns committed RTO/RPO targets into verified ones
Field
Value
Document status
ACTIVE
Date
July 9, 2026
Governing rule
A target in the Master Plan is “unverified” until it has passed a live test recorded here.
1. Test Cadence
Test
Frequency
Owner
Backup job success check
Daily (automated)
Monitoring/alerting
Backup age / RPO compliance check
Daily (automated)
Monitoring/alerting
Single database restore
Monthly
Network Services Team
Single app stack restore
Quarterly
Network Services Team
Core failover test (Core → app1-bu)
Quarterly
Network Services Team
Core failback test (app1-bu → Core)
Quarterly, immediately after failover test
Network Services Team
Full DR tabletop exercise
Twice yearly
Full team + stakeholders
Provider/S3 outage simulation
Twice yearly
Network Services Team
Do not skip the failback test: A failover test without an immediate failback test only proves half the design. Schedule them back-to-back so failback is never left untested for a full quarter.
2. Restore Test Log — Required Fields
Every restore test, whether scheduled or incident-driven, gets logged with:
Date and tester name
Server/application tested
Backup artifact path and timestamp used
Restore duration (actual, compared against the RTO target)
Data validation performed against a known record, file, or document
Issues encountered
Remediation ticket, if needed
Date
Tester
System
Backup Used
Duration
RTO Target Met?
Validation Passed?
Issues / Ticket
3. Failover / Failback Test Log
Date
Trigger Simulated
Detection Time
Time to Active
Failback Time
Split-Brain Check Passed?
Issues / Ticket
4. Immediate Priority Test Queue
Given uptime is the top priority, run these before anything else on the cadence above:
Standby capacity load test on app1-bu (validates DR Master Plan v2, Section 4).
First Core failover + failback test, end to end, to establish a real baseline against the 5-minute RTO target.
Verify MariaDB dump automation on wphost02 and fleettracker360, then restore one dump from each.
Restore one Docker volume on each of app1, app2, app3 into a test environment.
Confirm Object Lock and write-only IAM are actually enabled on the Wasabi bucket (Backup & Security Hardening Standard, Section 2), not just documented.
Stand up and validate the secondary backup destination replication job.
5. Sign-off
A target moves from “committed” to “verified” in the Master Plan only after it has a passing entry in this log, dated within the current quarter. Re-verify quarterly — infrastructure and application changes can silently break a previously-passing test.
Target
Status
Last Verified
Verified By
Core RTO: 5 minutes
Unverified — pending first test
Core RPO: 15 minutes
Unverified — pending first test
app1/2/3 RTO: 60 minutes
Unverified — pending first test
wphost02 / fleettracker360 RTO: 2 hours
Unverified — pending first test
Backup immutability (Object Lock)
Unverified — pending confirmation
Secondary backup destination
Not yet built
This document is the source of truth for whether the DR plan actually works, not just whether it is well written. Keep it current.