DevOps Professional · Module 3: Migration, restore and disaster recovery · Lab
Find a Double-Write Fault on a Small Dataset
45 min hands-on · Core
A local machine with a small fixture dataset and a dual-write script (simulated load standing in for scale; labeled as simulation).
Local guide: run the steps below on your own machine in order, then check the validation list.
Objectives
- Detect duplication with reconciliation queries
- Classify the cause: retry, overlap or second writer
- Fix the write path and dedup history by rule
Step 1
Detect the gap
Run the provided dual-write with the staged retry fault. Compare counts per window on both sides and quote the gap with the reconciliation query used.
Step 2
Classify and fix the path
Classify the cause from evidence (non-idempotent retry in this staging) and fix the write path with idempotency keys. State the dedup rule for history and apply it.
Step 3
Show the clean streak
Rerun the comparison across repeated runs and quote the clean streak. Write one paragraph on how the simulation differs from production scale.
How to confirm it worked
- Gap quoted with the reconciliation query
- Cause classified from evidence, path fixed
- History deduped by a stated rule
- Clean streak quoted with the simulation limit noted