Need help — gateway stuck unreachable | Friends of the Crustacean 🦞🤝 | Page 1

halcyon oasis Apr 28, 2026, 2:38 PM

#

Gateway stuck in unreachable state — survives reboot
macOS 26.3.1 (arm64), Node 25.6.1, OpenClaw 2026.4.24 (cbcfdf6)
Symptoms:

Gateway service shows "loaded · running" but gateway probe always fails
Connection error rotates between ECONNREFUSED, timeout, and ECONNRESET across attempts
Survives a full Mac reboot — not transient
tasks maintenance --apply reports 8 running tasks but cleans up 0 (stuck state)
1,194 audit errors, 300 warnings on tasks
Bonjour discovers the gateway, but loopback connections fail

Tried: gateway stop, gateway start, doctor, doctor --fix, tasks maintenance --apply, full Mac reboot — all unsuccessful.Gateway stuck in unreachable state — survives reboot
macOS 26.3.1 (arm64), Node 25.6.1, OpenClaw 2026.4.24 (cbcfdf6)
Symptoms:

Gateway service shows "loaded · running" but gateway probe always fails
Connection error rotates between ECONNREFUSED, timeout, and ECONNRESET across attempts
Survives a full Mac reboot — not transient
tasks maintenance --apply reports 8 running tasks but cleans up 0 (stuck state)
1,194 audit errors, 300 warnings on tasks
Bonjour discovers the gateway, but loopback connections fail

Tried: gateway stop, gateway start, doctor, doctor --fix, tasks maintenance --apply, full Mac reboot — all unsuccessful.
Diagnostic file attached: openclaw-diagnostic.txt
Diagnostic file attached: openclaw-diagnostic.txt

thin moatBOT Apr 28, 2026, 2:38 PM

#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

long impBOT Apr 28, 2026, 2:41 PM

#

I don’t see the diagnostic attachment on this side — Discord bridge shows zero attachments — so please reattach it or paste openclaw status --all in a code block.

My read: don’t chase the 8 tasks first. On 2026.4.24, this smells more like the Gateway is dying/stalling during startup/runtime-deps/Bonjour/service handoff, and the task audit mess is likely fallout. 2026.4.26 is the current GitHub latest release and has fixes in exactly this startup/plugin-runtime-deps area, including staging bundled plugin runtime deps before Gateway startup. (github.com)

Try this recovery path:

openclaw --version
which openclaw
node -v
openclaw gateway status --deep
openclaw status --all
tail -n 200 "$(ls -t /tmp/openclaw/openclaw-*.log | head -1)"
lsof -nP -iTCP:18789 -sTCP:LISTEN

Then upgrade/reinstall the service:

#

npm install -g openclaw@2026.4.26
hash -r
openclaw --version
openclaw doctor --repair
openclaw gateway install --force
openclaw gateway restart
openclaw gateway status --deep

Also: Node 25.6.1 is not where I’d debug this. Docs require Node >=22, but for recovery I’d use Node 22 LTS/24 rather than bleeding-edge 25. pnpm is recommended; Bun is not. (docs.openclaw.ai)

If it’s still unreachable after that, paste the gateway status --deep output plus the last 200 log lines. Docs for the debug ladder are here: https://docs.openclaw.ai/help/faq / https://docs.openclaw.ai/gateway/troubleshooting

halcyon oasis Apr 28, 2026, 2:47 PM

#

/Thank you. I'm new here, so it is not allowing me to attach anything. I will try your recommendation

long impBOT Apr 28, 2026, 2:47 PM

#

No worries — Discord does that sometimes for new accounts.

If attach isn’t allowed, paste just these in chunks/code blocks:

openclaw --version
which openclaw
node -v
openclaw gateway status --deep
openclaw status --all

And if still broken after the upgrade:

tail -n 200 "$(ls -t /tmp/openclaw/openclaw-*.log | head -1)"

#

One important bit: after changing Node/OpenClaw versions, reinstall the gateway service so launchd isn’t still pointing at the old binary:

openclaw gateway install --force
openclaw gateway restart

Good luck — this looks recoverable, just probably a bad 2026.4.24/startup state rather than your config. 🦐

halcyon oasis Apr 28, 2026, 2:50 PM

#

Found the root cause in the logs — Bonjour plugin is crashing the gateway every ~10 seconds with Unhandled promise rejection: CIAO ANNOUNCEMENT CANCELLED. Watchdog keeps restarting it, but each restart triggers another unhandled rejection that kills the gateway. The Bonjour service is stuck in "announcing" state and never reaches "active" before being killed.
Repeating log pattern:
[plugins] bonjour: restarting advertiser (service stuck in announcing for ~9500ms)
[openclaw] Unhandled promise rejection: CIAO ANNOUNCEMENT CANCELLED
[openclaw] wrote stability bundle: ...unhandled_rejection.json
Should I run gateway install --force first, or disable the Bonjour plugin? I only need local loopback access, no network discovery needed.
Version: OpenClaw 2026.4.24, Node 25.6.1, macOS 26.3.1, installed via Homebrew at /opt/homebrew/bin/openclaw.

long impBOT Apr 28, 2026, 2:51 PM

#

Disable Bonjour first. That’s the direct root-cause isolation, and you don’t need discovery for loopback.

Try:

openclaw plugins disable bonjour
openclaw gateway restart
openclaw gateway status --deep

If that command doesn’t work while the gateway is crash-looping, use the config/env path instead:

openclaw config set discovery.mdns.mode off
openclaw gateway restart

#

Then verify loopback:

curl -fsS http://127.0.0.1:18789/healthz

I’d not run gateway install --force until after the gateway stays up. Once stable, update off 2026.4.24 and then reinstall the service so launchd points at the updated Homebrew binary:

brew update
brew upgrade openclaw
openclaw gateway install --force
openclaw gateway restart

Docs: https://docs.openclaw.ai/gateway/bonjour

halcyon oasis Apr 28, 2026, 2:55 PM

#

Remedied. I'm indebted. Thank you

#Need help — gateway stuck unreachable