zeroclaw-labs/zeroclaw · critical · anyhow::Error

systemctl restart failed: {}

Error message

systemctl restart failed: {}

What it means

Thrown by maybe_restart_managed_daemon_service on Linux when a user-level systemd unit exists (~/.config/systemd/user/zeroclaw.service), `systemctl --user is-active` (SYSTEMD_STATUS_ARGS) reported `active`, and the subsequent `systemctl --user restart zeroclaw.service` (SYSTEMD_RESTART_ARGS) exits non-zero. Because the unit was active before, a failed restart usually leaves it inactive — check and restore it promptly.

Source

Thrown at crates/zeroclaw-channels/src/orchestrator/mod.rs:8825

            return Ok(false);
        }

        let active_output = Command::new("systemctl")
            .args(SYSTEMD_STATUS_ARGS)
            .output()
            .context("Failed to query systemd service state")?;
        let state = String::from_utf8_lossy(&active_output.stdout);
        if !state.trim().eq_ignore_ascii_case("active") {
            return Ok(false);
        }

        let restart_output = Command::new("systemctl")
            .args(SYSTEMD_RESTART_ARGS)
            .output()
            .context("Failed to restart systemd daemon service")?;
        if !restart_output.status.success() {
            let stderr = String::from_utf8_lossy(&restart_output.stderr);
            anyhow::bail!("systemctl restart failed: {}", stderr.trim());
        }

        return Ok(true);
    }

    Ok(false)
}

#[cfg(any(
    test,
    feature = "channel-discord",
    feature = "channel-lark",
    feature = "channel-matrix",
    feature = "channel-slack",
    feature = "channel-telegram",
    feature = "channel-wechat",
    feature = "whatsapp-web",
))]

View on GitHub (pinned to 88bb9c8533)

Solutions

1. Restore immediately: `systemctl --user restart zeroclaw.service` and run `journalctl --user -u zeroclaw.service -n 50` for the exact startup error
2. If ExecStart is stale, re-run the service install (e.g. `zeroclaw service install`) or edit ~/.config/systemd/user/zeroclaw.service to the current binary path, then `systemctl --user daemon-reload`
3. Fix the underlying config error if the daemon crashes on the new config; validate with a foreground run first
4. For headless/cron contexts, ensure `loginctl enable-linger <user>` and XDG_RUNTIME_DIR are set so the user manager is reachable

Example fix

# before
# systemctl restart failed: Failed to start zeroclaw.service: Unit ... 

# after
systemctl --user daemon-reload
systemctl --user restart zeroclaw.service
journalctl --user -u zeroclaw.service -n 50   # confirm active

let unit = dirs::home_dir().unwrap().join(".config/systemd/user/zeroclaw.service");
if !unit.exists() { return Ok(()); }
let active = Command::new("systemctl").args(["--user", "is-active", "zeroclaw.service"]).output()?;
if !String::from_utf8_lossy(&active.stdout).trim().eq_ignore_ascii_case("active") { return Ok(()); }

if let Err(e) = maybe_restart_managed_daemon_service() {
    eprintln!("restart failed ({e:#}); run `systemctl --user restart zeroclaw.service`");
    // check `journalctl --user -u zeroclaw.service` before retrying
}

When it happens

Trigger: Any config-save path that auto-reloads the daemon (e.g. allowlist bind save) on a systemd desktop/server host where the user unit is active. Common failure sources: the unit's ExecStart path is stale after reinstall, the new config crashes the binary on start, or the user systemd session (DBUS/XDG_RUNTIME_DIR) is unavailable from the invoking context (cron, ssh without lingering).

Common situations: Binary reinstalled to a different path while ~/.config/systemd/user/zeroclaw.service kept the old ExecStart; running over SSH/cron where the user manager is not reachable; daemon exits at boot due to invalid config edited just before the restart; `loginctl enable-linger` missing so the user bus disappears.

Related errors

• rc-service restart failed: {}
• launchctl start failed: {}
• systemd user lingering disabled; user service may stop after
• Could not detect init system. Supported: systemd, OpenRC. Us
• Linux service {action} only manages the default zeroclaw ser

AI-assisted analysis of zeroclaw-labs/zeroclaw@88bb9c8533 (2026-08-23). Data as JSON: /api/errors/93233132c584e3da. Report an issue: GitHub.