openclaw-lighthouse

install.sh + Feishu setup: duplicate plugin id warning spam and gateway health check failure

Summary

Running the installer (curl -fsSL https://openclaw.ai/install.sh | bash) and then enabling Feishu produced two problem signals:

1. repeated config warnings about duplicate plugin id for feishu
2. health check failure because gateway was not reachable (1006 abnormal closure)

The logs also show:

• systemd user services unavailable in this runtime
• service install skipped during onboarding step

Environment

• Install path style in report: global npm under /usr/lib/node_modules/openclaw
• Runtime user in report: root
• Setup path: install script + Feishu plugin install
• Channel selected: Feishu/Lark (飞书)

Reproduction

1. Run installer:
   • curl -fsSL https://openclaw.ai/install.sh | bash
2. Install/enable Feishu plugin in onboarding flow.
3. Observe repeated warnings:
   • plugins.entries.feishu: plugin feishu: duplicate plugin id detected ...
4. Observe systemd message:
   • Systemd user services are unavailable. Skipping lingering checks and service install.
5. Health check later fails with:
   • gateway closed (1006 abnormal closure (no close frame))

Findings

1. The duplicate id warning indicates two Feishu plugin sources are discoverable with the same id (feishu).
2. One warning path points to bundled plugin location:
   • /usr/lib/node_modules/openclaw/extensions/feishu/index.ts
3. Repeated warning lines suggest the config/validation path emits the same warning multiple times in this flow.
4. Gateway health check failure is likely operational (gateway not running/reachable) rather than plugin crash.
5. The runtime explicitly reports systemd user services unavailable, which explains skipped service install and later gateway reachability issues.

Mitigation / Workaround

1. Keep only one Feishu plugin source active (avoid duplicate id collision):
   • Use bundled plugin or installed plugin under ~/.openclaw/extensions/feishu, not both.
2. Set explicit plugin allowlist:
   • plugins.allow: ["feishu"]
3. Start gateway with a mode that works on this host:
   • supported user systemd service, or
   • foreground/tmux mode when systemctl --user is unavailable.
4. Verify after changes:
   • openclaw gateway status
   • openclaw channels status --probe
   • openclaw logs --follow

Risk / Impact

• UX noise: repeated warnings make onboarding look broken.
• Config ambiguity: uncertain which plugin copy is active.
• Availability risk: channel setup appears complete while gateway health is down.

Related Issues/PRs

• Related local Lighthouse note:
  • Feishu QuickStart install warns: plugins.allow is empty and non-bundled plugins may auto-load
• Solved workaround doc:
  • Gateway health check fails with 1006 on root VPS (systemctl --user unavailable)

Next actions

• Confirm plugin precedence with a minimal clean host (bundled-only vs installed-only).
• Reduce duplicate warning spam in onboarding/health flow (single consolidated warning preferred).
• Add explicit operator hint when systemd user service is unavailable (offer tmux foreground fallback command).

References

• Plugins (Extensions)
• Gateway runbook
• Gateway CLI
• Troubleshooting