9tSELF-HOSTED WORKSPACEDocs
Operations

DocsOperations

Keeping 9t running

Backups and retention

Systemd timers clean expired and trashed objects hourly and create a daily verified archive; fourteen backups are retained by default. Verify an archive without changing data:

node scripts/restore.mjs backups/9t-….tar.gz --verify

Stop 9t before a real restore, then omit --verify. The restore tool keeps the previous data directory beside the restored copy.

CLI reference

Everything below runs from the checkout (./9t …) or, once linked, as 9t … from anywhere. Object commands need a login first; server commands don't.

9t setup                         Install and configure this checkout
9t setup --help                  Installation options
9t setup --service               Install the boot service on an existing checkout
9t setup --service --dry-run     Preview the unit, writes nothing
9t setup --check                 Pre-flight only: disk, RAM, data dir, port
9t start                         Run the configured server
9t status                        Service, listener and app health
9t update                        Backup, pull, rebuild and restart

Workspace commands (run 9t login first):

9t login --url https://9t.example.com --username you
9t push <text|url|file> [--name name] [--lifetime forever|1h|1d|7d]
9t list [--trash]
9t get <id|name> [--output file]
9t share <id|name> [--lifetime 1d] [--password secret]
9t trash <id|name>
9t restore <id|name>
9t logout

push figures out the type itself: an existing file path becomes a file, an http(s) URL becomes a link, anything else becomes a snippet. get prints snippets and links, downloads files. IDs are 4 letters — any unambiguous prefix works.

Configuration:

9t config export [--output file]   # dump config as JSON (no secrets)
9t config import <file|->          # restore config, e.g. onto a new install
9t configure                       # interactive: modules, exposure, domain,
                                   # limits, theme — needs login except server
9t configure --section server      # port, address, data dir, HTTPS flag (offline)
9t doctor [--domain host]          # exposure, listener, DNS and HTTPS diagnostics

The token is stored with owner-only permissions in ~/.config/9t/config.json. 9t logout revokes it on the server.

Systemd service

Installs are named 9t-<checkout-hash>.service so they never overwrite each other. On an existing checkout:

./9t setup --service --dry-run   # preview the unit, writes nothing
sudo ./9t setup --service        # install + enable at boot

The unit runs node scripts/run-server.mjs, which reads your existing .env.production — data is never touched. Diagnostics: sudo systemctl status NAME, sudo journalctl -u NAME -f.

Updates and health

./9t status     # service state, build presence, listener + initialized flag
./9t update     # backup, git pull, rebuild into staging, migrate, restart

update refuses when the working tree has real local changes (regenerable build churn is auto-stashed) and backs up data before touching anything. Managed installs restart on the new build automatically; foreground installs print the build dir to restart with. start warns when sources are newer than the running build, which usually means an update was pulled but not rebuilt.

Edit this page on GitHub