fix: implement Windows stale gateway process cleanup before restart by arifahmedjoy · Pull Request #60480 · openclaw/openclaw

arifahmedjoy · 2026-04-03T18:45:02Z

Summary

findGatewayPidsOnPortSync() in src/infra/restart-stale-pids.ts returned [] immediately on Windows (process.platform === 'win32'), causing cleanStaleGatewayProcessesSync() to silently skip killing old gateway processes during self-restart via the schtasks supervisor path (triggerOpenClawRestart).

This caused an infinite retry loop on Windows:

[gateway] already running under schtasks; waiting 5000ms before retrying startup

The new gateway instance could never bind port 18789 because the old process was never terminated.

Root Cause

The self-restart path on Windows (triggerOpenClawRestart -> relaunchGatewayScheduledTask) calls cleanStaleGatewayProcessesSync() before spawning the schtasks restart script. But since findGatewayPidsOnPortSync returns [] on Windows, no stale processes are ever found or killed. The new schtasks-launched gateway then races the old one for the port and enters an unbounded retry loop.

Note: openclaw daemon restart works correctly because it uses the restartScheduledTask() path in service.ts, which properly calls terminateScheduledTaskGatewayListeners() with the Windows-aware findVerifiedGatewayListenerPidsOnPortSync(). Only the in-process self-restart path was broken.

Changes

New: `src/infra/windows-port-pids.ts`

Extracted all Windows-specific port scanning and process-args helpers from gateway-processes.ts into a shared module with configurable timeoutMs parameter. This:

Breaks the circular import between restart-stale-pids.ts and gateway-processes.ts (both now import from windows-port-pids.ts instead of from each other)
Fixes poll budget overshoot: Windows poll calls use POLL_SPAWN_TIMEOUT_MS (400ms) instead of the 5000ms default, keeping each poll within the waitForPortFreeSync 2s budget

`src/infra/restart-stale-pids.ts`

findGatewayPidsOnPortSync: On win32, discovers listening PIDs via readWindowsListeningPidsOnPortSync + verifies each with readWindowsProcessArgsSync / isGatewayArgv
pollPortOnceWindows: Uses readWindowsListeningPidsOnPortSync(port, 400) — just checks if port has any listener, no verification needed (stale gateway already killed)
terminateStaleProcessesSync: Add terminateStaleProcessesWindows() using taskkill.exe (graceful /T first, then /F force-kill) instead of SIGTERM/SIGKILL

`src/infra/gateway-processes.ts`

Delegates Windows helpers to the new windows-port-pids.ts module
Removes ~100 lines of inlined Windows functions
Keeps findVerifiedGatewayListenerPidsOnPortSync public API unchanged

`src/infra/restart-stale-pids.test.ts`

Mocks windows-port-pids.js (port scanning + process args) for the win32 platform-mock tests
Updated win32 tests verify delegation to readWindowsListeningPidsOnPortSync and readWindowsProcessArgsSync
Tests use real isGatewayArgv for full integration coverage

Testing

Lightly tested: verified fix resolves the infinite restart loop on Windows 11
Confirmed openclaw daemon restart and openclaw gateway call health work after fix
Existing Unix tests unaffected — all 38 restart-stale-pids tests pass
All 6 gateway-processes tests pass after refactor
Updated test mocks verify win32 delegation with real isGatewayArgv validation

greptile-apps · 2026-04-03T18:47:51Z

Greptile Summary

This PR fixes a real Windows-only bug where findGatewayPidsOnPortSync returned [] on win32, causing cleanStaleGatewayProcessesSync to skip killing the old gateway before schtasks relaunched a new one, producing an unbounded port-conflict loop. The approach — delegating to findVerifiedGatewayListenerPidsOnPortSync and adding taskkill-based termination — is directionally correct, but the implementation has two structural issues that should be addressed before merging:

Circular import: restart-stale-pids.ts now imports from gateway-processes.ts, which already imports from restart-stale-pids.ts, creating a mutual cycle that can trip bundlers and violates the codebase's module-boundary conventions.
Poll budget overshoot: pollPortOnceWindows delegates to the same PowerShell helper, whose internal timeout is 5 000 ms — 2.5× the waitForPortFreeSync 2 000 ms budget — so a single slow PowerShell invocation blocks the entire polling loop well past the intended ceiling.

Confidence Score: 3/5

Not safe to merge as-is; two structural issues — a circular module dependency and a poll-budget mismatch — should be resolved first.

Both findings are P1: the circular import between restart-stale-pids.ts and gateway-processes.ts is a real module-graph defect that bundlers can mishandle, and the 5 000 ms PowerShell timeout exceeding the 2 000 ms waitForPortFreeSync budget means the port-free polling loop can stall for much longer than intended on Windows. The core fix for the infinite restart loop is valid, but these two issues need to be addressed before landing.

src/infra/restart-stale-pids.ts (circular import with gateway-processes.ts; poll timeout exceeds budget)

Prompt To Fix All With AI

This is a comment left during a code review.
Path: src/infra/restart-stale-pids.ts
Line: 5

Comment:
**Circular import between `restart-stale-pids.ts` and `gateway-processes.ts`**

This new import creates a mutual circular dependency: `restart-stale-pids.ts` → `gateway-processes.ts` → `restart-stale-pids.ts` (line 5 of `gateway-processes.ts` already imports `findGatewayPidsOnPortSync` from here). While ESM live bindings mean this won't throw a "cannot access before initialization" error at runtime (both exports are `function` declarations), it is a structural issue that bundlers (esbuild, rollup) can mishandle and that the codebase's dynamic-import guardrail policy is designed to avoid.

The cleaner fix is to lift the Windows-specific `readWindowsListeningPidsOnPortSync` (or the verified variant) out of `gateway-processes.ts` into a shared `windows-port-pids.ts` helper that neither file needs to cross-import.

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: src/infra/restart-stale-pids.ts
Line: 194-201

Comment:
**Windows poll can block for 5 s, far exceeding the 2 s budget in `waitForPortFreeSync`**

`findVerifiedGatewayListenerPidsOnPortSync` runs PowerShell with `WINDOWS_GATEWAY_DISCOVERY_TIMEOUT_MS = 5 000 ms`. The `waitForPortFreeSync` wall-clock budget is `PORT_FREE_TIMEOUT_MS = 2 000 ms`. A single hung PowerShell spawn will block the entire polling loop for 5 s — 2.5× the intended ceiling — before `waitForPortFreeSync` even gets to re-check the deadline. On Unix, `POLL_SPAWN_TIMEOUT_MS = 400 ms` keeps each poll well under the budget.

The Windows helper should respect the same short per-call timeout, or `pollPortOnceWindows` should accept a timeout and pass it through to the underlying PowerShell / netstat calls.

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (1): Last reviewed commit: "fix: implement Windows stale gateway pro..." | Re-trigger Greptile}

Copilot

Pull request overview

Fixes Windows self-restart getting stuck in an infinite “already running under schtasks” retry loop by ensuring stale gateway listeners are discovered and terminated before attempting to relaunch.

Changes:

On Windows, findGatewayPidsOnPortSync() now uses the existing verified Windows listener discovery instead of returning [].
Adds a Windows-specific port polling path for waitForPortFreeSync() using the same listener discovery.
Adds a Windows-specific stale-process terminator using taskkill (tree kill first, then /F escalation).

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d492f2c6d2

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a8e2bead83

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

arifahmedjoy · 2026-04-03T19:45:05Z

CI Failure Analysis

All current CI failures are pre-existing on main and unrelated to this PR (which only modifies src/infra/restart-stale-pids.ts and src/infra/restart-stale-pids.test.ts):

Job	Error	Related?
`check`	`extensions/slack/src/outbound-adapter.ts:66` — `Type 'OpenClawConfig \| undefined' is not assignable to type 'OpenClawConfig'`	❌
`checks-windows-node-test`	`src/cli/update-cli.test.ts:817` — `expected undefined to be defined`	❌
`checks-node-test`	`src/process/exec.test.ts:92,107` — test timeouts (15s)	❌
`checks-fast-contracts-protocol`	`src/plugins/contracts/plugin-sdk-subpaths.test.ts:130` — missing `command-auth` exports	❌
`checks-fast-extensions-shard-3`	`extensions/feishu/src/tool-account-routing.test.ts:97` — assertion mismatch	❌
`checks-fast-extensions-shard-6` / `check-additional`	`lint:plugins:no-extension-test-core-imports` guard failure	❌

Happy to rebase onto main once these are resolved upstream.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d4666f364c

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

arifahmedjoy · 2026-04-03T20:05:51Z

Update: Greptile P1 findings resolved

Both structural issues flagged by Greptile have been addressed in the latest push (f97dd5c):

1. Circular import — fixed

Extracted all Windows-specific port scanning and process-args helpers from gateway-processes.ts into a new shared src/infra/windows-port-pids.ts module. Both restart-stale-pids.ts and gateway-processes.ts now import from windows-port-pids.ts — no circular dependency.

2. Poll budget overshoot — fixed

pollPortOnceWindows now passes POLL_SPAWN_TIMEOUT_MS (400ms) to readWindowsListeningPidsOnPortSync instead of using the 5000ms default. Each poll stays within the waitForPortFreeSync 2s budget, matching the Unix lsof behavior. The poll also skips per-PID command-line verification (unnecessary since stale gateway is already killed in the prior step).

CI status

All PR-related checks are now passing (including check, check-additional, and all extension shards). The 3 remaining failures (checks-fast-contracts-protocol, checks-node-test, checks-windows-node-test) are pre-existing on main and unrelated to this change.

@steipete @vincentkoc — this is ready for review when you have a chance. Happy to address any further feedback.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 30467bbe0c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5cf46b7588

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

findGatewayPidsOnPortSync() returned [] immediately on Windows, causing cleanStaleGatewayProcessesSync() to skip killing old gateway processes during self-restart (triggerOpenClawRestart -> schtasks path). This led to an infinite retry loop: 'gateway already running under schtasks; waiting 5000ms before retrying startup'. Changes: - Extract Windows port/process helpers into shared windows-port-pids.ts to break the circular import between restart-stale-pids.ts and gateway-processes.ts, with configurable timeoutMs for poll compliance - findGatewayPidsOnPortSync: discover + verify Windows gateway PIDs via readWindowsListeningPidsOnPortSync + readWindowsProcessArgsSync - pollPortOnceWindows: use short POLL_SPAWN_TIMEOUT_MS (400ms) so a single slow PowerShell call cannot exceed the 2s polling budget - terminateStaleProcessesSync: add terminateStaleProcessesWindows using taskkill.exe (graceful /T first, then /F force-kill) Fixes the Windows gateway restart infinite loop caused by the schtasks supervisor detecting a port conflict it cannot resolve.

@arifahmedjoy

…nks @arifahmedjoy)

obviyus

Reviewed latest changes; landing now.

obviyus · 2026-04-05T15:31:37Z

Landed on main.

Landed source commit: 9751f62
Merge commit: 63fcc52

Thanks @arifahmedjoy.

arifahmedjoy · 2026-04-05T15:51:21Z

You're welcome, @obviyus.

@arifahmedjoy

…nks @arifahmedjoy) * fix: implement Windows stale gateway process cleanup before restart findGatewayPidsOnPortSync() returned [] immediately on Windows, causing cleanStaleGatewayProcessesSync() to skip killing old gateway processes during self-restart (triggerOpenClawRestart -> schtasks path). This led to an infinite retry loop: 'gateway already running under schtasks; waiting 5000ms before retrying startup'. Changes: - Extract Windows port/process helpers into shared windows-port-pids.ts to break the circular import between restart-stale-pids.ts and gateway-processes.ts, with configurable timeoutMs for poll compliance - findGatewayPidsOnPortSync: discover + verify Windows gateway PIDs via readWindowsListeningPidsOnPortSync + readWindowsProcessArgsSync - pollPortOnceWindows: use short POLL_SPAWN_TIMEOUT_MS (400ms) so a single slow PowerShell call cannot exceed the 2s polling budget - terminateStaleProcessesSync: add terminateStaleProcessesWindows using taskkill.exe (graceful /T first, then /F force-kill) Fixes the Windows gateway restart infinite loop caused by the schtasks supervisor detecting a port conflict it cannot resolve. * fix: tighten windows stale gateway cleanup * fix: preserve windows restart probe failures * refactor: unify windows gateway pid verification * fix: preserve windows argv probe failures * fix: windows self-restart stale gateway cleanup (openclaw#60480) (thanks @arifahmedjoy) --------- Co-authored-by: Ayaan Zaidi <hi@obviy.us>

@arifahmedjoy

…nks @arifahmedjoy) * fix: implement Windows stale gateway process cleanup before restart findGatewayPidsOnPortSync() returned [] immediately on Windows, causing cleanStaleGatewayProcessesSync() to skip killing old gateway processes during self-restart (triggerOpenClawRestart -> schtasks path). This led to an infinite retry loop: 'gateway already running under schtasks; waiting 5000ms before retrying startup'. Changes: - Extract Windows port/process helpers into shared windows-port-pids.ts to break the circular import between restart-stale-pids.ts and gateway-processes.ts, with configurable timeoutMs for poll compliance - findGatewayPidsOnPortSync: discover + verify Windows gateway PIDs via readWindowsListeningPidsOnPortSync + readWindowsProcessArgsSync - pollPortOnceWindows: use short POLL_SPAWN_TIMEOUT_MS (400ms) so a single slow PowerShell call cannot exceed the 2s polling budget - terminateStaleProcessesSync: add terminateStaleProcessesWindows using taskkill.exe (graceful /T first, then /F force-kill) Fixes the Windows gateway restart infinite loop caused by the schtasks supervisor detecting a port conflict it cannot resolve. * fix: tighten windows stale gateway cleanup * fix: preserve windows restart probe failures * refactor: unify windows gateway pid verification * fix: preserve windows argv probe failures * fix: windows self-restart stale gateway cleanup (openclaw#60480) (thanks @arifahmedjoy) --------- Co-authored-by: Ayaan Zaidi <hi@obviy.us>

@arifahmedjoy

…nks @arifahmedjoy) * fix: implement Windows stale gateway process cleanup before restart findGatewayPidsOnPortSync() returned [] immediately on Windows, causing cleanStaleGatewayProcessesSync() to skip killing old gateway processes during self-restart (triggerOpenClawRestart -> schtasks path). This led to an infinite retry loop: 'gateway already running under schtasks; waiting 5000ms before retrying startup'. Changes: - Extract Windows port/process helpers into shared windows-port-pids.ts to break the circular import between restart-stale-pids.ts and gateway-processes.ts, with configurable timeoutMs for poll compliance - findGatewayPidsOnPortSync: discover + verify Windows gateway PIDs via readWindowsListeningPidsOnPortSync + readWindowsProcessArgsSync - pollPortOnceWindows: use short POLL_SPAWN_TIMEOUT_MS (400ms) so a single slow PowerShell call cannot exceed the 2s polling budget - terminateStaleProcessesSync: add terminateStaleProcessesWindows using taskkill.exe (graceful /T first, then /F force-kill) Fixes the Windows gateway restart infinite loop caused by the schtasks supervisor detecting a port conflict it cannot resolve. * fix: tighten windows stale gateway cleanup * fix: preserve windows restart probe failures * refactor: unify windows gateway pid verification * fix: preserve windows argv probe failures * fix: windows self-restart stale gateway cleanup (openclaw#60480) (thanks @arifahmedjoy) --------- Co-authored-by: Ayaan Zaidi <hi@obviy.us>

Copilot AI review requested due to automatic review settings April 3, 2026 18:45

openclaw-barnacle Bot added the size: S label Apr 3, 2026

Copilot started reviewing on behalf of arifahmedjoy April 3, 2026 18:45 View session

greptile-apps Bot reviewed Apr 3, 2026

View reviewed changes

Comment thread src/infra/restart-stale-pids.ts Outdated

Comment thread src/infra/restart-stale-pids.ts

Copilot AI reviewed Apr 3, 2026

View reviewed changes

Comment thread src/infra/restart-stale-pids.ts

Comment thread src/infra/restart-stale-pids.ts Outdated

Comment thread src/infra/restart-stale-pids.ts

Comment thread src/infra/restart-stale-pids.ts

arifahmedjoy force-pushed the fix/windows-stale-gateway-cleanup branch 2 times, most recently from 9cca773 to d492f2c Compare April 3, 2026 19:25

chatgpt-codex-connector Bot reviewed Apr 3, 2026

View reviewed changes

Comment thread src/infra/restart-stale-pids.ts Outdated

Comment thread src/infra/restart-stale-pids.ts

arifahmedjoy force-pushed the fix/windows-stale-gateway-cleanup branch from d492f2c to a8e2bea Compare April 3, 2026 19:32

chatgpt-codex-connector Bot reviewed Apr 3, 2026

View reviewed changes

Comment thread src/infra/restart-stale-pids.ts Outdated

Comment thread src/infra/restart-stale-pids.ts

chatgpt-codex-connector Bot reviewed Apr 3, 2026

View reviewed changes

Comment thread src/infra/restart-stale-pids.ts Outdated

arifahmedjoy force-pushed the fix/windows-stale-gateway-cleanup branch from d4666f3 to f97dd5c Compare April 3, 2026 19:53

openclaw-barnacle Bot added size: M and removed size: S labels Apr 3, 2026

arifahmedjoy mentioned this pull request Apr 4, 2026

[Bug]: Windows gateway self-restart enters infinite retry loop — stale process never killed #60878

Closed

obviyus force-pushed the fix/windows-stale-gateway-cleanup branch from f97dd5c to 30467bb Compare April 5, 2026 14:57

openclaw-barnacle Bot added size: L and removed size: M labels Apr 5, 2026

chatgpt-codex-connector Bot reviewed Apr 5, 2026

View reviewed changes

Comment thread src/infra/windows-port-pids.ts

Comment thread src/infra/restart-stale-pids.ts Outdated

chatgpt-codex-connector Bot reviewed Apr 5, 2026

View reviewed changes

Comment thread src/infra/restart-stale-pids.ts

obviyus self-assigned this Apr 5, 2026

arifahmedjoy and others added 6 commits April 5, 2026 20:59

fix: tighten windows stale gateway cleanup

673bd09

fix: preserve windows restart probe failures

f86eccc

refactor: unify windows gateway pid verification

e42d7ba

fix: preserve windows argv probe failures

d9cbad4

fix: windows self-restart stale gateway cleanup (openclaw#60480) (tha…

9751f62

…nks @arifahmedjoy)

obviyus force-pushed the fix/windows-stale-gateway-cleanup branch from 535c323 to 9751f62 Compare April 5, 2026 15:30

obviyus approved these changes Apr 5, 2026

View reviewed changes

obviyus merged commit 63fcc52 into openclaw:main Apr 5, 2026
17 of 31 checks passed

github-actions Bot mentioned this pull request Apr 5, 2026

📡 Upstream Digest — 2026-04-05 16:27 UTC curtismercier/openclaw-mods#478

Open

andyk-ms mentioned this pull request Apr 5, 2026

fix(windows): detect stale gateway PIDs via netstat #52485

Closed

BradGroux mentioned this pull request Apr 9, 2026

WORKING: All Microsoft Issues and PRs #49126

Closed

DanielGao511 mentioned this pull request Apr 10, 2026

Bug: gateway restart spawns duplicate processes on Windows (3 windows) #52044

Closed

clawsweeper Bot mentioned this pull request Apr 30, 2026

Windows: gateway restart does not wait for active tasks and loses session state #56284

Open

4 tasks

clawsweeper Bot mentioned this pull request May 2, 2026

[Bug]: Editing openclaw.json while gateway is running triggers hot-reload crash loop on Windows (stale lock file + EADDRINUSE) #68493

Open

Uh oh!

Conversation

arifahmedjoy commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Root Cause

Changes

New: src/infra/windows-port-pids.ts

src/infra/restart-stale-pids.ts

src/infra/gateway-processes.ts

src/infra/restart-stale-pids.test.ts

Testing

Uh oh!

greptile-apps Bot commented Apr 3, 2026

Greptile Summary

Confidence Score: 3/5

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

arifahmedjoy commented Apr 3, 2026

CI Failure Analysis

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

arifahmedjoy commented Apr 3, 2026

Update: Greptile P1 findings resolved

1. Circular import — fixed

2. Poll budget overshoot — fixed

CI status

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

obviyus left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

obviyus commented Apr 5, 2026

Uh oh!

arifahmedjoy commented Apr 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

arifahmedjoy commented Apr 3, 2026 •

edited

Loading

New: `src/infra/windows-port-pids.ts`

`src/infra/restart-stale-pids.ts`

`src/infra/gateway-processes.ts`

`src/infra/restart-stale-pids.test.ts`