The `sixpanel` command line

What this page is for: run your server from an SSH prompt — especially when the panel itself will not open.

You need

• SSH access to the server.
• Nothing else. The installer put the tool on the box. It is called sixpanel, and sp is the same tool with a shorter name.

The tool must run as root. On most cloud images you log in as ubuntu or opc — type the command anyway and it re-runs itself through sudo, asking for your password once. In a script it prints the exact sudo sixpanel … line to use instead, so nothing hangs waiting on a prompt you cannot see.

Start with these three

sixpanel                  the menu — everything below, by number
sixpanel doctor           what is wrong, and the fix under each failure
sixpanel logs nginx -f    watch a service's log live

sixpanel on its own opens a numbered menu. It starts with your panel link and how many services are running, then lists actions. Pick a number. If an action fails you get the reason and the menu again — never a bare shell prompt.

The five you will actually use

1. Find out what is wrong

sudo sixpanel doctor
sudo sixpanel doctor --self

Every check prints a tick, a warning or a cross, and every failure carries the same one-line fix the panel's Health tab shows.

2. Read a log

sudo sixpanel logs nginx -n 50
sudo sixpanel logs app -f
sudo sixpanel logs php --since 10m --grep 'PHP Fatal'
sudo sixpanel logs laravel -n 100

Targets: any service name, app (php + queue + scheduler + nginx together), all, or laravel for the application's own log file. Everything except laravel comes out of the system journal, which is where systemd keeps each service's output. -f follows the log until you press Ctrl-C.

3. Clear the caches after a settings change

sudo sixpanel cache clear

This is the answer to most "I changed a setting and nothing happened" reports. It clears the application's cached configuration, rebuilds the fast caches and restarts the workers.

4. Restart the website

sudo sixpanel service restart app

The classic "turn it off and on again" for the site. The panel keeps running.

5. Send support one file

sudo sixpanel support-bundle

It writes one .tgz — by default into /root/ — with versions, the health report, service states, disk usage and the last 500 lines of every log. Send us that file and most questions are answered in one round trip. What is removed before it is written: the stack settings file is reduced to key names only, the panel's own state file is not included at all, and the secret code in your panel link plus common password-looking values are masked in the collected logs.

The seven that get you back in

Each of these undoes one lock, over SSH, in one line. This is the reason root SSH matters as much as your panel password.

sudo sixpanel username           forgot the name you type on the login page
sudo sixpanel password reset     forgot the panel password
sudo sixpanel 2fa off            lost the phone with the 6-digit codes
sudo sixpanel domain none        the panel's own domain broke
sudo sixpanel entry-code reset   the secret link leaked
sudo sixpanel allowlist clear    the login IP allowlist shut you out
sudo sixpanel unlock             too many failed logins locked your address out

sixpanel unlock on its own lists what is locked and why, and changes nothing. Add an address, or all, to let it log in again. It is the answer to "my password is right and the panel still will not take it" — after enough failed tries the panel refuses that address for 30 minutes, and one of the two counters doing it is not written to any file, so there is nothing on the server to read that would explain it.

sixpanel 2fa off forgets the phone you set up, so your next login shows a fresh QR code to scan with the new one. Two-factor stays on and staying on is the recommendation. If you have no phone at that moment the enrolment screen also offers Not now — continue without two-factor login, which is a real way through — it is written to your activity log, and from then on your password alone opens the panel until you enrol again from the Security page.

Two more, for the logins you gave to other people:

sudo sixpanel temp-login              list the temporary logins
sudo sixpanel temp-login revoke-all   delete every one of them
sudo sixpanel demo off                switch the public demo login off

Every one of them says what it will change and asks before it does it. See Security for what each lock does, Let someone else in for the two shared logins, and When something is broken for the symptoms that lead here.

Which of these work when the panel is down

The sixpanel tool is an ordinary program on the machine, not a client for a server that has to be up. The services it talks to are systemd units and the files it reads are on disk, so almost everything keeps working with the panel stopped: info, url, version, logs, every service command, ssl status, disk, prune, top, cache clear, the db commands, sql, exec, artisan, support-bundle, server firewall, server os-update, server reboot, password reset, username, entry-code, domain, port, 2fa off, temp-login, demo off, allowlist, unlock, backup targets, backup password and deploy history.

Some of those change something the panel keeps in memory. With the panel up they are applied straight away. With it down, the tool stops the panel, writes, and starts it again — that takes ten to twenty seconds and it announces each stage, so a pause is not a hang.

Needs the panel, or its own switch: doctor. Use sixpanel doctor --self for the host-only checks that need nothing from the panel. That is the one to run when the panel will not start.

Runs the panel's own engine: backup run, backup list, backup restore, deploy and deploy rollback. These work when the panel is running normally, and also when it is fully stopped. They refuse in the middle state — a running panel this tool cannot talk to — and tell you which of two things to do: install the small helper (sudo apt-get install -y python3) or do that one action in the panel.

Getting help

sixpanel help              every command, grouped
sixpanel help logs         one command, with examples
sixpanel logs --help       the same thing
sixpanel cheatsheet        the ones people actually use, on one screen
man sixpanel               the manual page

Press Tab to complete commands, service names and project names.

Typed something that is not a command? It suggests the closest ones, including two-word commands: sixpanel restore points at sixpanel backup restore, sixpanel firewall at sixpanel server firewall. It will never offer a command that reboots or deletes something as a guess for an unrelated word.

Safety rules that are true everywhere

• Anything that cannot be undone tells you what changes, what restarts and how to undo it, then asks.
• Anything that deletes data makes you type the target's name — the database name, or the stack path. --yes does not skip that one.
• A password is never accepted as an argument. Use sudo sixpanel password reset (it generates one and prints it), or let your shell read it without showing it:

  read -rs NEW && printf '%s' "$NEW" | sudo sixpanel password --stdin

• sixpanel server update updates SixPanel. sixpanel deploy updates your own 6ammart code. See Update SixPanel itself.
• sixpanel port random moves the panel to a new random port. It does not restore an old one; to go back, name it: sixpanel port 41397.

Full reference — you will not need most of this

Everything above is the part people use. What follows is the complete surface, for the day you want to look something up. Skip it until then.

This list mirrors the CLI's own command table. If the two ever disagree, sudo sixpanel help is the truth.

Panel — its address, password and lifecycle

Creating a temporary login, and switching the demo login on, are done in the panel and not here. Both hand out a password that is shown exactly once, and both deserve the warning that is on the screen at the time. See Let someone else in.

Each of these also works without the word panel: sixpanel url, sixpanel password reset, sixpanel allowlist clear, and so on.

Services — the systemd units that run your site

Service names: db, redis, nginx, panel, php, queue, scheduler, reverb, react. With several projects, each project has its own php-SLUG, queue-SLUG, scheduler-SLUG, reverb-SLUG and react-SLUG.

Behind those short names are ordinary Ubuntu services: db is mariadb.service, redis is redis-server@six-<project>.service (one per project — the bare name is the default project's, and redis-<project> names another's), php is php8.3-fpm.service (one service, with one pool per project), and the per-project workers are sixpanel-queue@<slug>.service, sixpanel-schedule@<slug>.timer, sixpanel-reverb@<slug>.service and sixpanel-react@<slug>.service. You never have to use those names — but if you already know systemd, systemctl status on any of them works exactly as you expect.

Two shortcuts: all is every service this server defines, including every project's. app is php + queue + scheduler + nginx for the current project — the "restart my website" set.

Projects — one 6ammart install each

Database

db import replaces everything in that database. Every table in this project's database is replaced by the contents of the file you name. Orders, users and settings that the file does not contain are gone, and nothing takes a copy for you first. Take one yourself before you start: sudo sixpanel db dump -o /root/before.sql.gz. The command then makes you type the database name to confirm, and --yes does not skip that. When it finishes, run sudo sixpanel db migrate and then sudo sixpanel cache clear — in that order.

Backups

A restore writes over your live database and files, and backup password prints a secret to your terminal in plain text. Both are explained on Backups.

Deploying your 6ammart code

This server

server uninstall --purge-data cannot be undone by anything on this server. Without --purge-data the command only stops and switches off the services: your databases, uploaded files, certificates and local backups stay on disk, and re-installing picks them up again. With --purge-data the panel's own folder is deleted — its settings and every backup kept on this server. Your databases, your site files (/var/www) and your certificates (/etc/letsencrypt) are not deleted by either form; take them off the server yourself if you want them gone. sixpanel disk names the directory your database files are in (Ubuntu keeps them in /var/lib/mysql or /var/lib/mariadb, depending on the release). Backups you sent to S3, SFTP or Google Drive are not touched, but they are encrypted, and the password that opens them is stored in what you are about to delete. Run sudo sixpanel backup password and save that password somewhere else first. The command makes you type the stack path to confirm, and --yes does not skip that.

Two things uninstall does not do, whichever form you use: nginx keeps serving your shop's PHP (the site files, the vhosts and the databases are all still there), and certificate renewal stops, because the panel was the only thing running it.

server update and server os-update are different things. See Server updates and restarts.

Diagnose and fix

prune takes no arguments. It trims the system journal to the last seven days, clears the package manager's cache and empties this project's application log. It asks first, and no database, upload or backup can be removed by it.

Getting inside

You will probably never need this group. It is for support, and for people who already know Linux, Laravel and SQL. A wrong command here can break the site in ways the panel cannot repair.

exec matters more than it looks. There is nothing to "enter" on this server — the programs are simply here — so exec runs your command as the project's own Linux user, in the project's own directory. That is what keeps file ownership right. A root shell in your code folder leaves root-owned files that break the app days later, and this is the command that stops you doing it by accident.

Reference

[panel] marks commands whose work happens in the panel's web interface, not here. The tool says so before you type it — in help, in the cheatsheet and in man sixpanel — instead of failing at you afterwards.

Flags every command understands

What the exit codes mean

0 fine · 1 you typed something wrong · 2 SixPanel is not installed · 3 the panel is unreachable · 4 the check or the action failed.

So a monitoring hook is one line:

sudo sixpanel doctor --json || alert-someone

If it went wrong

• "sixpanel must run as root" — put sudo in front, exactly as the message prints it.
• "no /etc/sixpanel.conf found" — SixPanel is not installed on this server, or the stack folder moved.
• "sixpanel libraries not found" — the tool was copied by hand. Re-run the installer for this server.
• "… is read out of the panel's JSON, and this host has no JSON reader" — a few commands read the panel's own file to answer: backup list, backup targets, deploy history, deploy rollback last and allowlist. They stop with that message rather than print an empty list you would believe. Install a reader and run it again: sudo apt-get install -y python3 (or jq). sudo sixpanel allowlist clear does not need one, so a lockout is still one command away.

More symptoms, in the order people hit them: see When something is broken.