Fence regional write authority before promoting a standby
An outage runbook says promote the standby but never explains how the unreachable former primary loses authority.
- Focused work estimate
- 6h + prerequisites
- Priority in the scenario
- High
- Engineering practice
- Fencing · Consistency
Estimated field mix
- Distributed systems60%
- System design40%
Field percentages are editorial estimates of the ticket's engineering focus. They total 100%; they are not measured time, proficiency scores, or ownership evidence.
Review it, then add it to your workspace.
The board opens an editable draft; nothing is saved until you confirm it. Sign-in and workspace permissions apply, and Demo boards remain ephemeral.
Project context
A fictional logistics dashboard serves distant customers from one primary region. Stakeholders want faster reads and outage recovery but have not agreed which data may be stale.
Setup prerequisites
- Create a local primary/replica simulator and synthetic shipment records.
- Use local models; no multi-region infrastructure is provisioned.
Preceding work
Complete these dependencies, or supply their agreed outputs before taking this ticket.
Acceptance criteria
- Define one durable authority epoch and promotion preconditions.
- Reject commands using an old epoch.
- Describe the availability tradeoff when fencing cannot be confirmed.
Implementation constraints
- Use a local two-writer model; do not claim split-brain prevention without enforced fencing.
Verification to include
- Promote a new epoch and accept only the new writer.
- Let the old writer recover and reject its stale-epoch command.
Deliverables
- Failover authority protocol and fencing probe
Rollout and recovery
Require the probe before any real failover procedure; refuse promotion when authority is ambiguous.
Value of the work
For the engineer: Practice consistency tradeoffs, failover authority and recovery assumptions.
For the team: Review regional availability proposals with explicit data-loss and staleness limits.
Evidence boundaries
Outcome Evidence: Tests, patches, and runbooks are requested deliverables. They become Outcome Evidence only through a qualified Mission and immutable Evidence IDs.
Ownership Evidence: Independent adaptation must be observed under a declared verification policy and cite immutable Evidence IDs. Completing a planning ticket establishes no Ownership Evidence.