openclaw - 💡(How to fix) Fix [Bug]: Gateway ERR_MODULE_NOT_FOUND after update + repeated unhandled rejections (empty snapshots) [1 participants]

Bug type

Crash (process/app exits or hangs)

Beta release blocker

No

Summary

Gateway startup fails with ERR_MODULE_NOT_FOUND after update, accompanied by repeated unhandled rejections with empty event snapshots

Steps to reproduce

1. Trigger an OpenClaw update on 2026.4.25-beta.1 (63803d7)
2. The update causes a gateway restart
3. Observe telemetry / error logs

Note: exact reproduction steps are NOT_ENOUGH_INFO. The error set occurred during a natural update cycle; no manual reproduction has been attempted.

Expected behavior

After an update-triggered gateway restart, the gateway should start cleanly with all required modules resolving correctly and no unhandled promise rejections.

Actual behavior

1. gateway.startup_failed — one occurrence at 2026-04-26T12:39:44.499Z. The gateway process spent 8822 ms in startup before failing with ERR_MODULE_NOT_FOUND. No further detail about which module was missing is available.

2. unhandled_rejection — nineteen occurrences clustered in three windows on 2026-04-26:

   • 06:36–06:44 UTC (roughly every 45 seconds)
   • 08:24–08:25 UTC
   • 12:35–12:39 UTC

   Every unhandled rejection snapshot reports an empty events array (count=0), meaning no rejection payload was captured. The rejections themselves are opaque — no stack trace, message, or rejection reason is available.

3. The gateway currently starts and operates normally; the failures were transient.

OpenClaw version

OpenClaw 2026.4.25-beta.1 (63803d7)

Operating system

Linux WSL2, 6.6.87.2-microsoft-standard-WSL2, x86_64

Install method

nvm global (npm)

Model

minimax/MiniMax-M2.7

Provider / routing chain

openclaw -> minimax

Additional provider/model setup details

NOT_ENOUGH_INFO

Logs, screenshots, and evidence

telemetry error: gateway.startup_failed
  timestamp: 2026-04-26T12:39:44.499Z
  error: Error
  code: ERR_MODULE_NOT_FOUND
  process uptime: 8822 ms
  context: occurred immediately after a gateway restart triggered by an update

telemetry error: unhandled_rejection x 19
  timestamps:
    cluster A - 2026-04-26 06:36-06:44 UTC (~45 s period)
    cluster B - 2026-04-26 08:24-08:25 UTC
    cluster C - 2026-04-26 12:35-12:39 UTC
  snapshot events: [] (count=0 for all 19 occurrences)
  rejection detail: none captured

commands.log: no matching error lines found for the above timestamps

Impact and severity

• Gateway availability: One transient startup failure; the gateway was able to start on subsequent attempts.
• Observability gap: Nineteen unhandled rejections fired with zero captured detail, making root-cause analysis impossible without additional instrumentation.
• Update disruption: All errors are correlated with the 2026-04-26 update cycle. Users updating to this beta may experience a brief gateway outage plus silent rejections.
• Severity: Medium. The gateway recovers, but the silent rejections mask potentially real bugs, and the ERR_MODULE_NOT_FOUND on startup is a regression risk.

Additional information

• All errors occurred within the same 2026-04-26 update window.
• commands.log contains no entries matching these error timestamps, suggesting the failures happen outside the command-logging scope.
• The empty events array on every unhandled rejection snapshot may itself be a secondary bug — the rejection payload is being dropped before it can be recorded.