Summary

Environment

Problem

openclaw cron run <id> always returns gateway timeout after 30000ms,
even though the daemon is confirmed running and all other commands work normally.

Steps to Reproduce

openclaw cron run <any-job-id>
→ Error: gateway timeout after 30000ms

What Works

• openclaw cron list ✅
• openclaw cron runs --id <id> ✅
• openclaw status (RPC probe: ok) ✅
• openclaw daemon status (running, active) ✅
• Scheduled cron jobs run normally at their set time ✅

What Fails

• openclaw cron run <id> ❌ (always timeout)
• openclaw cron run <id> --expect-final ❌ (always timeout)

Daemon Log at Time of Failure

Gateway is running (pid confirmed), RPC probe ok, Discord connected.
No error logged on the gateway side during cron run attempt.

Expected Behavior

cron run should trigger immediate job execution for debugging purposes.

Steps to reproduce

1. openclaw cron add --name test --cron "0 6 * * *" --tz Asia/Tokyo --session main --system-event "test"
2. openclaw cron run
   → Error: gateway timeout after 30000ms

Expected behavior

The job runs immediately and returns the result.

Actual behavior

Error: gateway timeout after 30000ms
Gateway target: ws://127.0.0.1:18789
Source: local loopback

This happens 100% of the time regardless of job ID or session type.
Other commands (cron list, cron runs, status) work fine on the same gateway.

OpenClaw version

2026.2.26 (bc50708)

Operating system

Linux 6.17.0-1008-nvidia (arm64) / Ubuntu 24

Install method

npm install -g openclaw Managed by systemd (openclaw-gateway.service)

Logs, screenshots, and evidence

Gateway status at time of failure:
- Runtime: running (pid 2377345, state active)
- RPC probe: ok
- Listening: 127.0.0.1:18789
- Discord: connected

No error logged on gateway side during cron run attempt.

Impact and severity

Affected: All users using cron jobs with local loopback gateway on Linux/arm64
Severity: Blocks workflow — cannot manually test or debug cron jobs
Frequency: Always (100% reproducible)
Consequence: Cannot verify job behavior before scheduled run; forced to wait for next scheduled execution to test changes

Additional information

• Scheduled runs (at cron time) work correctly — only manual trigger fails
• Occurs on both --session main and --session isolated jobs
• Persists after openclaw update (2026.2.25 → 2026.2.26)
• Persists after daemon restart
• cron run without --expect-final also times out