Uh oh!
There was an error while loading. Please reload this page.
fix(world-postgres): serialize graphile-worker bootstrap with advisory lock - #1966
Closed
TooTallNate wants to merge 2 commits into
Closed
fix(world-postgres): serialize graphile-worker bootstrap with advisory lock#1966TooTallNate wants to merge 2 commits into
TooTallNate wants to merge 2 commits into
Conversation
…y lock Prevents "duplicate key value violates unique constraint pg_namespace_nspname_index" errors when multiple processes (e.g. dev server + test runner) call world.start() concurrently against a fresh database. PostgreSQL's CREATE SCHEMA IF NOT EXISTS is not race-safe across concurrent sessions, and graphile-worker's installSchema does not lock its bootstrap path.
Contributor
🦋 Changeset detectedLatest commit: f8f2980 The changes in this PR will be included in the next version bump. This PR includes changesets to release 2 packages
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
Contributor
🧪 E2E Test Results✅ All tests passed Summary
Details by Category✅ ▲ Vercel Production
✅ 💻 Local Development
✅ 📦 Local Production
✅ 🐘 Local Postgres
✅ 🪟 Windows
✅ 📋 Other
|
Contributor
📊 Benchmark Results
workflow with no steps💻 Local Development
▲ Production (Vercel)
workflow with 1 step💻 Local Development
▲ Production (Vercel)
workflow with 10 sequential steps💻 Local Development
▲ Production (Vercel)
workflow with 25 sequential steps💻 Local Development
▲ Production (Vercel)
workflow with 50 sequential steps💻 Local Development
▲ Production (Vercel)
Promise.all with 10 concurrent steps💻 Local Development
▲ Production (Vercel)
Promise.all with 25 concurrent steps💻 Local Development
▲ Production (Vercel)
Promise.all with 50 concurrent steps💻 Local Development
▲ Production (Vercel)
Promise.race with 10 concurrent steps💻 Local Development
▲ Production (Vercel)
Promise.race with 25 concurrent steps💻 Local Development
▲ Production (Vercel)
Promise.race with 50 concurrent steps💻 Local Development
▲ Production (Vercel)
workflow with 10 sequential data payload steps (10KB)💻 Local Development
▲ Production (Vercel)
workflow with 25 sequential data payload steps (10KB)💻 Local Development
▲ Production (Vercel)
workflow with 50 sequential data payload steps (10KB)💻 Local Development
▲ Production (Vercel)
workflow with 10 concurrent data payload steps (10KB)💻 Local Development
▲ Production (Vercel)
workflow with 25 concurrent data payload steps (10KB)💻 Local Development
▲ Production (Vercel)
workflow with 50 concurrent data payload steps (10KB)💻 Local Development
▲ Production (Vercel)
Stream Benchmarks(includes TTFB metrics)workflow with stream💻 Local Development
▲ Production (Vercel)
stream pipeline with 5 transform steps (1MB)💻 Local Development
▲ Production (Vercel)
10 parallel streams (1MB each)💻 Local Development
▲ Production (Vercel)
fan-out fan-in 10 streams (1MB each)💻 Local Development
▲ Production (Vercel)
SummaryFastest Framework by WorldWinner determined by most benchmark wins
Fastest World by FrameworkWinner determined by most benchmark wins
Column Definitions
Worlds:
❌ Some benchmark jobs failed:
Check the workflow run for details. |
Contributor
There was a problem hiding this comment.
Pull request overview
This PR addresses a flaky race during @workflow/world-postgres startup where concurrent processes can both attempt graphile-worker schema installation on a fresh database, triggering duplicate key value violates unique constraint "pg_namespace_nspname_index".
Changes:
- Serialize graphile-worker bootstrap (
makeWorkerUtils+migrate) using a Postgres advisory lock inpackages/world-postgres/src/queue.ts. - Update unit test pool mocks to include
pool.connect()to match the new startup path. - Add a changeset documenting the patch release.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.
| File | Description |
|---|---|
| packages/world-postgres/src/queue.ts | Adds transaction-scoped advisory lock wrapper around graphile-worker bootstrap to avoid concurrent schema install races. |
| packages/world-postgres/src/queue.test.ts | Updates pool mock to include connect() for the new locking code path. |
| packages/world-postgres/src/reenqueue.test.ts | Updates pool mock to include connect() for the new locking code path. |
| .changeset/fix-world-postgres-graphile-bootstrap-race.md | Declares a patch release and describes the race/lock fix. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
- Release the locked connection before makeWorkerUtils() so the bootstrap can't deadlock against the same pool when maxPoolSize is small (Copilot review #1). Achieved by replicating graphile-worker's installSchema DDL ourselves under the lock; once the schema exists graphile-worker's own installSchema is a no-op. - Add a unit test that asserts the lock-ordering: pool.connect -> BEGIN -> pg_advisory_xact_lock -> CREATE SCHEMA -> COMMIT -> release -> makeWorkerUtils (Copilot review #2).
VaguelySerious
left a comment
Member
There was a problem hiding this comment.
Not convinced we should duplicate migration code in the queue file and use a lock for this
| "@workflow/world-postgres": patch | ||
| --- | ||
| Pre-create the `graphile_worker` schema under a Postgres advisory lock so concurrent `world.start()` callers (e.g. dev server + test runner, or multiple serverless invocations) don't race on the not-race-safe `CREATE SCHEMA IF NOT EXISTS` and fail with `duplicate key value violates unique constraint "pg_namespace_nspname_index"`. The lock is held on a short-lived dedicated connection that is released before `makeWorkerUtils()` runs so the bootstrap can't deadlock against the same pool when its max size is small. |
There was a problem hiding this comment.
Suggested change
| Pre-create the `graphile_worker` schema under a Postgres advisory lock so concurrent `world.start()`callers (e.g. dev server + test runner, or multiple serverless invocations) don't race on the not-race-safe `CREATE SCHEMA IF NOT EXISTS`and fail with `duplicate key value violates unique constraint "pg_namespace_nspname_index"`. The lock is held on a short-lived dedicated connection that is released before `makeWorkerUtils()` runs so the bootstrap can't deadlock against the same pool when its max size is small. | |
| Ensure concurrent `world.start()`calls don't race on `CREATE SCHEMA IF NOT EXISTS`by using a lock. Fixes `duplicate key value violates unique constraint "pg_namespace_nspname_index"` errors. |
TooTallNate
commented
May 19, 2026
MemberAuthor
Closing in favor of #2019 — Peter's approach pre-bootstraps the See #2019 (review) for the full comparison. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes the
E2E Local Postgres Testsflake onmain(example failing job) where the first test (addTenWorkflow) fails with:Root cause
PR #1959 removed
instrumentation.tsfrom the workbenches, which previously pre-warmedworld.start()in the Next.js dev server before any HTTP traffic arrived. Without that pre-warm,world.start()is now called lazily on the first workflow request — and in the local-postgres CI matrix, the vitest test process itself also initializes a postgres world (becauseWORKFLOW_TARGET_WORLD=@workflow/world-postgresis set in the test env). So two independent Node processes (the test runner and the Next.js dev server) race to install thegraphile_workerschema on a fresh database.PostgreSQL's
CREATE SCHEMA IF NOT EXISTSis not race-safe across concurrent sessions — the existence check happens at one MVCC snapshot but thepg_namespaceinsert happens at commit, so two sessions can both pass the check and one then fails withduplicate key value violates unique constraint "pg_namespace_nspname_index". graphile-worker'sinstallSchemadoes no locking around this.Fix
Wrap the
makeWorkerUtils({pgPool}).migrate()call inpackages/world-postgres/src/queue.tswith a Postgres advisory lock (pg_advisory_xact_lock) so that concurrent callers across processes are serialized on graphile-worker schema bootstrap. The lock is transaction-scoped, so it is automatically released if the bootstrapping process dies mid-flight.Verification
pnpm --filter @workflow/world-postgres typecheck✓pnpm --filter @workflow/world-postgres build✓pnpm --filter @workflow/world-postgres test✓ (109/109, including the real-Postgres integration tests intest/spec.test.tsthat exercise the fullworld.start()path)