Just Dashboard Uninstall
Take Just Dashboard off the server, keeping its deployments and data for a reinstall or wiping them.
Just Dashboard 0.8.4 uninstalls itself from its own Configuration page. There is no separate uninstall script: the dashboard knows which containers, volumes and folders it owns, so it does the removal. You decide whether what it made stays for a reinstall or goes with it.
Before you start
You need an administrator account. Keep an SSH session to the server open: once the stack is down, the page that started the uninstall stops answering, and the rest of the transcript is only on the server.
If you might come back, back up the data directory (JD_DATA_DIR, normally /var/lib/just-dashboard) and keep JD_MASTER_KEY somewhere safe, even if you choose to keep everything.
Uninstall from Configuration
- Open Settings → Configuration and scroll to Danger zone.
- Choose Uninstall….
- In Uninstall Just Dashboard, choose Keep for a reinstall or Wipe everything. The dialog lists what will be removed and what will be kept, with the real counts and sizes read from this server.
- For a wipe, decide whether to turn on Delete the backups too, then type the server's hostname to confirm.
- Choose Uninstall or Wipe and uninstall.
The Danger zone doesn't appear while the dashboard is restarting or already uninstalling, or when it can't find the Compose checkout it was started from. An uninstall also won't start while an update is running.
What each choice does
| Keep for a reinstall | Wipe everything | |
|---|---|---|
| The dashboard's three containers, its two images and open web-terminal sessions | Removed | Removed |
| Deployments | Kept, still running | Removed with their containers, volumes, folders, images, networks, routes and public ingress |
| The data directory: accounts, settings, secrets, deployment history, audit trail | Kept | Deleted |
.env with JD_MASTER_KEY | Kept, with a copy beside the data as install.env | Deleted, with its backup copies |
Backups in JD_BACKUP_DIR | Kept | Kept unless you turned on Delete the backups too |
| Databases it created, proxy sites, firewall rules, WireGuard and other network settings | Kept | Kept |
| The cloned checkout | Kept | Kept; delete it yourself afterwards |
A wipe keeps the data directory's nginx cache if a site on the server still caches into it. A filesystem mounted inside a folder being deleted, such as a disk under the backup directory, is left alone and named in the transcript.
What happens during the uninstall
The dashboard first pauses everything that could put a deployment back: the deployment engine, Git watching, schedules and backups. It gives running deployments up to 15 minutes to reach a safe stopping point. For a wipe, it then removes the deployments while the page is still there to show it. If pausing fails, the uninstall stops and the dashboard restarts to resume what was paused.
A separate container, just-dashboard-uninstall, then takes the stack down and removes the rest. If taking the stack down fails, it starts the dashboard again. A file that can't be deleted is a line in the transcript, not a reason to stop.
The full transcript is written to /var/log/just-dashboard-uninstall.log on the server, which survives a wipe:
sudo cat /var/log/just-dashboard-uninstall.logReinstall after keeping
Run the installer again, either in the original checkout or in a fresh clone:
sudo ./install.shIn the original checkout, accept the default to keep the existing .env. In a fresh clone, the installer finds the settings the uninstall kept and offers to restore them, so the new install opens the old data with the old key. The dashboard removes the kept install.env once it is running on that key. See install for what happens if the key is missing.