阅读中文版本

Build a Distributed Lock Service

Distributed Lock Service — a complete walkthrough of mutual exclusion, covering lock acquirer, fence issuer, lease renewer, and testing.

distributed_lock_service/
├── src/index.ts
├── src/lock-acquirer.ts
├── src/fence-issuer.ts
├── tests/distributed_lock_service.test.ts
└── package.json

// Lock acquirer: mutual exclusion core
export function lockacquirer(input: LockacquirerInput): LockacquirerResult {
  // Build the lock acquirer that anchors mutual exclusion and defines the core data shape.
  return processLockacquirer(input);
}

// Fence issuer — handles mutual exclusion Add the fence issuer so
export class Fenceissuer {
  constructor(private opts: FenceissuerOptions) {}
  async run(ctx: Context): Promise<FenceissuerOutput> {
    return this.execute(ctx);
  }
}

// Lease renewer: mutual exclusion stage
export const leaserenewer = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

// Release handler for mutual exclusion
export async function releasehandler(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

// Watchdog timer: mutual exclusion finalization
export function watchdogtimer(state: State): Result {
  validate(state);
  return finalize(state);
}

Project overview

This project builds a mutual exclusion system whose reason for existing is granting locks with fencing tokens so a stale holder cannot corrupt shared state after a pause. The temptation in mutual exclusion is to bolt the interesting part onto a working prototype and discover the hard cases in production; this build inverts that order and treats a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject as the first thing designed, not the last.

You will implement Lock acquirer, Fence issuer, Lease renewer, Release handler, and Watchdog timer, each as a module with one responsibility and a test that proves it in isolation. The mutual exclusion pipeline is then the wiring that composes them, and the integration test is where the composition is shown to survive a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject.

The walkthrough is written for a developer who has used a mutual exclusion system but has not had to make one honest. It favors explicit data shapes over clever abstractions, and it names the failure each design choice prevents rather than presenting the choice as a style preference.

Each phase ends with something observable — a lock acquirer that returns the right shape, a fence issuer that rejects bad input, an integration test that goes green — so you can stop at any point and still have a mutual exclusion system that works for the cases covered so far.

What you will build

The finished mutual exclusion system exposes five capabilities, each with a focused test:

• Lock acquirer
• Fence issuer
• Lease renewer
• Release handler
• Watchdog timer

Lock acquirer is the foundation that every later stage assumes; Fence issuer is where mutual exclusion correctness lives; Lease renewer handles a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject; Release handler makes the system observable; and Watchdog timer is the graceful-degradation layer that keeps overload from becoming corruption.

The build keeps the mutual exclusion logic in ordinary modules you can read in one sitting, and pulls in dependencies only for the parts that are genuinely tedious to hand-roll. That keeps the diff between the tutorial and a real codebase small, and keeps the failure modes visible in code rather than hidden behind a framework.

Architecture and data model

The system is organized around Lock acquirer, Fence issuer, and Lease renewer, with Release handler tying them together. Data flows as a lock identified by a name, held with a lease and a monotonically increasing fencing token that guards against stale writers.

Each component is a module with a pure interface where practical: the same input produces the same output, and side effects are pushed to the edges. That discipline is what makes the phases independently testable, because a test for Lease renewer does not need to stand up Lock acquirer and Fence issuer to exercise the interesting branch.

The data model is deliberately explicit. A lock identified by a name, held with a lease and a monotonically increasing fencing token that guards against stale writers Keeping the shape written down — in a schema, a type, or a fixture — means a change to the shape is a visible diff rather than a silent runtime shift, and it means the integration test can assert against a concrete shape instead of guessing.

Configuration and code are kept apart: the configuration says what mutual exclusion should do, the modules say how. A reviewer can change a threshold or a policy without reading the implementation, which keeps the risky changes small and lets the diff stand as the review.

The failure path through Lease renewer is designed before the happy path is finished, because a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject is the case that distinguishes a mutual exclusion system from a demo. Every component declares what it does when its inputs are wrong, and the integration test asserts each of those declarations.

Prerequisites and dependencies

Have these ready before you install anything, so the mutual exclusion setup is not blocked by a missing tool:

• Node.js 20 LTS or later
• Familiarity with mutual exclusion fundamentals
• A local environment to run the test suite

The runtime dependencies are chosen to keep the mutual exclusion datapath legible:

• redis (^4.7.0) — Lease store
• fastify (^5.0.0) — Lock API

# setup for distributed-lock-service
pnpm init
pnpm add redis fastify
pnpm add -D typescript tsx vitest

Versions are pinned to a major range so the mutual exclusion install is reproducible; commit the lockfile. The lock acquirer reads its operational parameters from the environment and validates them at startup, so a missing value fails the boot rather than the first mutual exclusion request.

Directory structure

Keep each mutual exclusion component in its own file so it can be tested without starting the others:

distributed_lock_service/
├── src/index.ts
├── src/lock-acquirer.ts
├── src/fence-issuer.ts
├── tests/distributed_lock_service.test.ts
└── package.json

Phase 1: Lock acquirer

Build the lock acquirer that anchors mutual exclusion and defines the core data shape.

// Lock acquirer: mutual exclusion core
export function lockacquirer(input: LockacquirerInput): LockacquirerResult {
  // Build the lock acquirer that anchors mutual exclusion and defines the core data shape.
  return processLockacquirer(input);
}

After this phase the lock acquirer works in isolation and is covered by a unit test. The lock acquirer defines the data shape that every later stage assumes, so getting it right first means the later phases compile against a stable contract.

Treat the lock acquirer as the single source of truth for the mutual exclusion data shape. Validate inputs at the boundary and reject early, so a malformed value never reaches Fence issuer or Lease renewer where its effect would be harder to localize.

Phase 2: Fence issuer

Add the fence issuer so mutual exclusion can granting locks with fencing correctly.

// Fence issuer — handles mutual exclusion Add the fence issuer so
export class Fenceissuer {
  constructor(private opts: FenceissuerOptions) {}
  async run(ctx: Context): Promise<FenceissuerOutput> {
    return this.execute(ctx);
  }
}

After this phase the fence issuer works in isolation and is covered by a unit test. This is where mutual exclusion correctness lives: the fence issuer is what makes granting locks with fencing tokens so a stale holder cannot corrupt shared state after a pause true, and its test is the proof.

Keep the fence issuer pure where you can. A function that takes an input and returns an output is trivial to test and impossible to mock wrong; a class with hidden state is the opposite. Push side effects to the caller and keep the fence issuer about the transformation.

Phase 3: Lease renewer

Wire the lease renewer to handle a holder that pauses, without losing data.

// Lease renewer: mutual exclusion stage
export const leaserenewer = (next: Next) => async (req: Request) => {
  const out = await transform(req);
  return next(out);
};

After this phase the lease renewer works in isolation and is covered by a unit test. The lease renewer is the difference between a mutual exclusion system that works on the happy path and one that survives a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject.

Design the failure before the success. Decide what lease renewer does when a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject: reject, retry, shed, or degrade — and write that test first. A mutual exclusion system whose failure behavior is unspecified will pick one at random under pressure, and it is usually the wrong one.

Phase 4: Release handler

Integrate the release handler to make mutual exclusion observable and testable end to end.

// Release handler for mutual exclusion
export async function releasehandler(stream: AsyncIterable<Chunk>): Promise<Aggregate> {
  let acc = init();
  for await (const c of stream) acc = merge(acc, c);
  return acc;
}

After this phase the release handler works in isolation and is covered by a unit test. Observability is not a finishing touch; it is how you know the mutual exclusion system is doing what it claims in production.

Emit one structured event per meaningful action, with enough context to localize a failure without a second round-trip. The release handler is what turns “something is slow” into “the fence issuer is slow for inputs matching this shape,” which is the difference between an alert and an investigation.

Phase 5: Watchdog timer

Finalize the watchdog timer so the mutual exclusion system degrades gracefully under overload.

// Watchdog timer: mutual exclusion finalization
export function watchdogtimer(state: State): Result {
  validate(state);
  return finalize(state);
}

After this phase the watchdog timer works in isolation and is covered by a unit test. The watchdog timer is what keeps overload from becoming corruption: when the mutual exclusion system is asked for more than it can do, it degrades in a defined way instead of failing in an undefined one.

Pick a degradation strategy for watchdog timer — shed, queue, or backpressure — and make it explicit in code and in a test. Leaving the behavior to chance is how a mutual exclusion traffic spike becomes a data loss incident.

Testing

Each phase has a focused test: lock acquirer, fence issuer, and lease renewer are exercised in isolation, then an integration test drives the full mutual exclusion pipeline and asserts the a holder that pauses, case is handled.

Start with unit tests that pin each component’s contract: the lock acquirer returns the documented shape, the fence issuer transforms correctly, and the lease renewer fails in the documented way. These tests are fast and they localize regressions to the component that broke.

Then add an integration test that composes the components and drives a realistic mutual exclusion scenario end to end. Assert the happy path first, then deliberately trigger a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject and assert the system degrades rather than corrupts. Because the components are pure where practical, the integration test stays deterministic rather than flaky.

A mutual exclusion suite must also prove the negative: bad input is rejected at the boundary, overload is shed rather than queued indefinitely, and a partial failure does not leave the system in an inconsistent state. These are the tests that catch the regressions a happy-path suite never would.

The entire mutual exclusion suite runs with no external service; every dependency is either in-process or stubbed, so every case is reproducible in CI. That is what lets the lease renewer failure-mode assertions live as tests rather than as a manual checklist that drifts out of date.

# run the mutual exclusion test suite
pnpm vitest run

Deployment and operations

Run at least two mutual exclusion instances behind a load balancer and share state in Redis. Set resource limits so mutual exclusion overload surfaces as backpressure rather than failure.

Run at least two stateless instances behind a load balancer, and move any shared state into a shared store so the mutual exclusion behavior is global rather than per instance. Size limits so the system sheds load before it runs out of memory, because a mutual exclusion system that OOMs under load fails in a way that is hard to reason about.

Drain on deploy: stop accepting new work, let in-flight mutual exclusion operations finish, then exit. Because the system holds no state that cannot be reconstructed, a rollout is a matter of cycling instances one at a time while watching the error rate.

When a dependency is unavailable, decide explicitly whether the mutual exclusion system fails closed or degrades. Leaving that choice to chance is how a dependency outage becomes a silent data loss, so make the choice visible in code and in a test.

Security considerations

Validate all mutual exclusion inputs at the boundary, reject oversized payloads before buffering them, and keep secrets in the environment rather than the repository.

The mutual exclusion system is a choke point, so it is the right place to reject malformed input before it reaches fence issuer or lease renewer. Validate the shape and the size at the edge, and let the inner components assume what they receive is well-formed.

Never log secrets or full payloads; log enough to localize a failure without recording data that should not be on disk. Treat any external call as untrusted and validate its response before using it, because a mutual exclusion system that trusts its inputs blindly is one malformed response away from corruption.

Keep dependencies pinned and run an audit in CI. A transitive dependency with a known vulnerability is still a vulnerability in your mutual exclusion system, and the lockfile is what makes the audit meaningful.

Observability

Emit one structured event per mutual exclusion operation with the inputs’ shape, the outcome, and a correlation id. Track error rate and latency per component and alert on a step change rather than a single spike.

The correlation id is the thread that ties a user-visible symptom to the mutual exclusion decision and to the component log entry for the same operation. Without it, an investigation starts from a timestamp and a guess.

Watch the lease renewer failure rate separately from the overall error rate. A jump in lease renewer failures usually means a holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject, not a bug in fence issuer, and the two have different owners and different fixes.

Chart the share of mutual exclusion operations that shed, queue, or succeed. A change in the mix points at the cause faster than an aggregate latency can, because each class has a different root.

Troubleshooting

A holder that pauses, lets its lease expire, and then writes with a stale token that a fencing-aware resource must reject. Reproduce with a minimal fixture, then add a guard test so the mutual exclusion system rejects the condition loudly.

Performance regresses under mutual exclusion load. Profile with the sampling analyzer, locate the hot path in fence issuer, and add a targeted benchmark to lock the regression.

Keep the mutual exclusion dashboards separate from dependency dashboards. When an operation fails, the correlation id lets you decide in seconds whether the fault is lock acquirer, fence issuer, or lease renewer, instead of guessing from a timeout.

Limitations and next steps

• Single-region; multi-region replication is out of scope.
• The lock acquirer is in-process and not horizontally shared by default.
• No auth layer; assume deployment behind an edge that enforces identity.

The mutual exclusion system is single-region by design; extending to multi-region means making lock acquirer shareable across regions, which is a real change to the consistency model rather than a config flag. The lease renewer is in-process, which is correct for one instance but must move to a shared store before scaling out.

There is no authentication or authorization layer; the system assumes it is deployed behind an edge that enforces identity. Adding one is straightforward but out of scope for this build, and doing it here would distract from the mutual exclusion correctness that is the point.

The natural next step, once the mutual exclusion pipeline is trusted in production, is to add a short-TTL cache for idempotent reads and a richer release handler that exports histograms rather than counters. Both are additive and neither changes the contracts this build establishes.