Admin CLI Overview

simai-admin.sh provides maintenance commands with two modes:

• Direct CLI: sudo /root/simai-env/simai-admin.sh <section> <command> [options]
• Interactive menu: sudo /root/simai-env/simai-admin.sh menu (numeric choices, stays in section; self-update reloads menu)
  • Text backend is the default for stable console UX.
  • After each command, menu shows status/output and waits for Enter before returning.
• Managed file access now lives under Access; this batch covers global access to all projects under WWW_ROOT and project-scoped access with chroot isolation.

Supported OS: Ubuntu 22.04/24.04. Run as root.

Profiles

• Registry-driven: defined in profiles/*.profile.sh (declarative PROFILE_ variables only).
• Current profiles:
• generic (default): nginx root <project-root>/public, creates placeholder page and php-mode healthcheck; optional DB; PHP required.
• laravel: nginx root <project-root>/public, requires artisan/bootstrap/app.php; DB required; cron/queue supported.
• static: serves <project-root>/public/index.html; no PHP/DB; no cron/queue; nginx-mode healthcheck at /healthcheck (local-only).
• alias: points a new domain to an existing site (reuses target PHP-FPM pool and root); no DB/pool creation.
• wordpress: nginx root <project-root>/public, requires DB, enables WordPress-oriented nginx routing and recommended cron.
• bitrix: nginx root <project-root>/public, requires DB, enables Bitrix-oriented routing with basic hardening and recommended cron.
• Public web root is always <project-root>/public; default filesystem path is /home/simai/www/<domain>, while slug is used for IDs (pool/cron/socket/logs).

See docs/architecture/profiles.md.

Common behaviors

• Domain -> project slug auto-derived if not provided.
• PHP version selection from installed /etc/php/* when not passed.
• Healthcheck copied to public/healthcheck.php for php-mode profiles; static uses nginx-served /healthcheck (local-only); alias inherits target.
• Site removal cleans nginx and PHP-FPM pools; optional files/DB/user removal via prompts (alias removal only drops nginx).
• Site doctor: read-only diagnostics against profile (filesystem, nginx, PHP, cron, SSL, DB) with PASS/WARN/FAIL summary; for Bitrix it also warns about root-owned docroot files, fails when nginx SEF fallback does not point to /bitrix/urlrewrite.php, fails when the nginx upload body limit is missing, and warns when static asset compression/cache headers are missing. It does not apply changes.
• Laravel daily ops (laravel status, laravel app-ready, laravel finalize, laravel perf-status, laravel perf-apply) provide low-risk application bootstrap, post-bootstrap baseline, and optimization checks for laravel profile sites.
• WordPress daily ops (wp status, wp installer-ready, wp finalize, wp cron-status, wp cron-sync, wp cache-clear) provide low-risk operational checks and install/scheduler/cache maintenance for wordpress profile sites.
• Bitrix daily ops (bitrix status, bitrix finalize, bitrix ownership, bitrix cron-status, bitrix cron-sync, bitrix agents-status, bitrix agents-sync, bitrix cache-clear) provide low-risk operational checks and post-install/scheduler/cache/ownership maintenance for bitrix profile sites.
• In the interactive menu, these daily application/CMS actions live under Applications, which opens separate submenus for Laravel, WordPress, and Bitrix.
• PHP commands: list/reload installed versions, and php install to add a new PHP version (uses ondrej/php, installs FPM/CLI/common extensions, with post-install tests).
• Site list shows domain, profile, PHP version, root/alias target, and brief SSL status.
• Logging:
  • Platform log: /var/log/simai-admin.log.
  • Command audit log: /var/log/simai-audit.log records command start/finish with user, section/command, redacted args, exit code, and correlation ID.
  • Setup log: /var/log/simai-env.log.

System and automation

• User guide for menu-driven work: docs/guide/menu-user-guide.md.
• Command reference for this section: docs/commands/self.md.
• Production rollout gate: docs/operations/production-readiness.md.

Self commands

• self update: update scripts in place (reloads menu when invoked from menu).
  • Honors update source from /etc/simai-env.conf: SIMAI_UPDATE_REF (refs/heads/... or refs/tags/...) or SIMAI_UPDATE_BRANCH.
  • Creates best-effort pre-update backup at /root/simai-backups/simai-env-preupdate-<timestamp>.tar.gz for manual rollback.
  • Runs a fast post-update smoke check (bash -n + executable presence). Set SIMAI_UPDATE_SMOKE_STRICT=yes to fail update on smoke errors.
• self version: show local/remote versions to know if an update is available (including configured update ref).
• self auto-update-status: show cached automatic update state (off|check|apply-safe), last check time, and cached local/remote versions.
• self auto-update-enable-check / self auto-update-enable-apply / self auto-update-disable: control automatic update checks in /etc/simai-env.conf.
• self auto-update-run-check: refresh update availability now without applying the update.
• In apply-safe mode, the interactive menu applies updates only at safe points (section boundaries / post-command return) and restores the same menu section after restart.
• self scheduler: run one internal scheduler tick (the command used by the managed system cron entry).
• self scheduler-status: show whether the shared scheduler cron is installed, plus job state/last run/next due.
• self scheduler-enable --job all|auto-optimize|health-review|site-review / self scheduler-disable --job all|auto-optimize|health-review|site-review: enable or disable the shared scheduler globally or per job without touching the cron entry itself.
• self scheduler-run --job auto-optimize|health-review|site-review: run one scheduler job immediately for testing/debugging.
• self auto-optimize-status: show the simple user-facing automatic optimization state without exposing low-level scheduler details.
• self auto-optimize-enable / self auto-optimize-disable: turn automatic optimization on or off globally while leaving the shared scheduler infrastructure in place.
• self health-review-status: show the latest read-only platform review produced by the shared scheduler.
• self site-review-status: show the latest read-only review of unfinished sites, stale setup candidates, and sites that are good pause candidates.
• self perf-status: show current server optimization status, detected server size, recommended preset, live nginx/mysql/redis/FPM pressure signals, and estimated FPM oversubscription.
• self perf-plan --limit <n>: show the heaviest PHP-FPM pools on the server, plus usage-aware suggested target modes (safe, balanced, parked) for the full server footprint.
• self perf-rebalance --limit <n> --mode auto|safe|parked --confirm yes: apply site perf-tune to the heaviest eligible pools, reducing global FPM oversubscription in controlled batches. auto respects the site usage class.
• self perf-apply --preset small|medium|large --confirm yes: apply a managed server baseline for future PHP-FPM pools, PHP OPcache, nginx, MySQL, and Redis (when installed).
• The regular System menu uses user-facing labels such as Platform status, Optimization status, Automatic updates, and Check for updates now; scheduler internals, Health review, and Site review remain in Advanced mode.
• site perf-status --domain <domain>: inspect current per-site PHP-FPM governance, socket/service state, pool share, estimated global FPM oversubscription, memory risk, and cron/queue footprint.
• site info, site usage-status, and site perf-status now include a simple optimization posture plus a plain-language recommendation so ordinary users can understand the current state without reading raw FPM values first.
• site perf-tune --domain <domain> --mode parked|safe|balanced|aggressive --confirm yes: apply site-level FPM governance without touching nginx/MySQL/Redis.
• site usage-status --domain <domain> / site usage-set --domain <domain> --class standard|high-traffic|rarely-used --confirm yes: simple user-facing site activity class mapped to the internal performance engine.
• site auto-optimize-status --domain <domain> / site auto-optimize-enable|disable|reset --domain <domain> --confirm yes: per-site override for automatic optimization; useful when one site should be excluded from scheduler-driven automatic changes.
• site runtime-status --domain <domain>: show whether the site runtime is active or suspended and whether its pool/cron/queue are currently enabled.
• site add --domain <domain> --host-mode wildcard: create one site that serves both the main domain and all first-level subdomains (*.domain) on the same nginx vhost.
• ssl letsencrypt --domain <domain> --wildcard yes --dns-provider cloudflare --dns-credentials <file>: request one Let’s Encrypt cert for both the main domain and all first-level subdomains of a wildcard-host site.
• In the menu, these are intentionally phrased in simpler language such as Activity & optimization, Automatic optimization for this site, Site availability, Pause site, Resume site, Optimization plan, Database server status, and Configuration check.
• site runtime-suspend --domain <domain> --confirm yes: suspend a site runtime by disabling its PHP-FPM pool, parking nginx behind a managed 503, and disabling cron/queue where applicable.
• site runtime-resume --domain <domain> --confirm yes: restore a previously suspended site runtime.
• laravel status --domain <domain>: show Laravel lifecycle state (placeholder, app-ready, post-install) and the next practical action.
• laravel app-ready --domain <domain>: bootstrap a real Laravel application via Composer and prepare .env from SIMAI DB credentials.
• laravel finalize --domain <domain> --confirm yes [--migrate yes] [--ssl yes --email <email>]: complete Laravel post-bootstrap baseline (APP_KEY, storage link, scheduler, baseline optimization, optional migrations and Let’s Encrypt).
• laravel perf-status --domain <domain> / laravel perf-apply --domain <domain> --mode safe|balanced|aggressive --confirm yes: Laravel optimization status and apply flow.
• wp perf-status --domain <domain> / wp perf-apply --domain <domain> --mode standard|woocommerce-safe --confirm yes: WordPress optimization status and apply flow.
• wp installer-ready --domain <domain>: prepare WordPress core files, generated wp-config.php, and wp-cli so the browser installer can run cleanly.
• wp finalize --domain <domain> --confirm yes [--ssl yes --email <email>]: complete a finished WordPress web install in one step (optimization baseline + scheduler baseline + optional Let’s Encrypt).
• bitrix perf-status --domain <domain> / bitrix perf-apply --domain <domain> --mode standard|high-load --confirm yes: Bitrix optimization status and apply flow (site tune + PHP baseline + installer-aware agents/cache steps).
• bitrix finalize --domain <domain> --confirm yes [--ssl yes --email <email>]: finalize a completed Bitrix web install in one step (PHP baseline + agents sync + optional Let’s Encrypt).
• bitrix ownership --domain <domain> [--apply yes --confirm yes]: detect and repair root-owned Bitrix web files and managed symlinked module checkouts after root-run installs, restores, or deployments.

Internal scheduler

• Bootstrap now installs one shared cron entry: /etc/cron.d/simai-scheduler.
• The cron entry stays stable and always calls simai-admin.sh self scheduler.
• Individual background capabilities are enabled/disabled inside simai-env, not by editing cron lines.
• Ordinary users can treat this as one switch via self auto-optimize-enable / self auto-optimize-disable; Advanced mode still exposes scheduler details and manual job control.
• Current built-in job:
• auto_optimize: policy-driven performance maintenance for oversubscribed PHP-FPM servers.
• health_review: recurring read-only platform health summary.
• site_review: recurring read-only review of unfinished, stale, or pause-candidate sites.
• auto_optimize now uses usage-aware targets by default:
  • standard -> safe
  • high-traffic -> balanced
  • rarely-used -> parked
• Managed config keys in /etc/simai-env.conf:
  • SIMAI_SCHEDULER_ENABLED=yes|no
  • SIMAI_AUTO_OPTIMIZE_ENABLED=yes|no
  • SIMAI_AUTO_OPTIMIZE_MODE=observe|assist|manual
  • SIMAI_AUTO_OPTIMIZE_INTERVAL_MINUTES=<n>
  • SIMAI_AUTO_OPTIMIZE_COOLDOWN_MINUTES=<n>
  • SIMAI_AUTO_OPTIMIZE_LIMIT=<n>
  • SIMAI_AUTO_OPTIMIZE_REBALANCE_MODE=auto|safe|parked

Non-Interactive Contract

Use this section when running commands from automation (CI, cron, deploy scripts).

• Prefer direct CLI mode (simai-admin.sh <section> <command> ...), not menu.
• Treat exit code 0 as success and non-zero as failure.
• In menu mode, cancel actions are intentionally non-fatal to the menu session, but they are now reported distinctly as CANCELLED instead of SUCCESS.

Output channels

• Command output is primarily printed to stdout.
• Operational logs are also written to /var/log/simai-admin.log.
• Audit trail is written to /var/log/simai-audit.log with redacted arguments.

Color control

• Disable ANSI colors in automation with NO_COLOR=1 or SIMAI_UI_COLOR=never.
• Force colors for interactive troubleshooting with SIMAI_UI_COLOR=always.

Recommended automation pattern

NO_COLOR=1 /root/simai-env/simai-admin.sh self status >/tmp/simai-status.txt
rc=$?
if [[ $rc -ne 0 ]]; then
  echo "simai status failed" >&2
  exit $rc
fi

Menu backend switches

• SIMAI_MENU_BACKEND=text forces text menu backend.
• SIMAI_MENU_BACKEND=whiptail enables optional whiptail backend (falls back safely when unavailable).
• Backend can also be switched during menu session: System -> Menu backend.

Commands detail: see docs/commands/*. Daily profile mapping: docs/operations/profile-ops-matrix.md.