Troubleshooting — Noodara

Every failure the installer can produce has its own exit code and its own actionable message. This is the complete table, taken directly from the script:

Exit code	Reason	What to do
10	Not running as root	Re-run the installer with `sudo`.
11	A required command is missing	Install the named command (`curl`, `openssl`, `ss`, `ip`, `awk` or `grep`) and re-run.
12	Unsupported operating system	Noodara requires Ubuntu 22.04 or 24.04.
13	Unsupported architecture	Noodara requires `amd64` or `arm64`.
14	Insufficient RAM	Free up memory, add RAM, or set `NOODARA_SKIP_RESOURCE_CHECK=1` to override deliberately.
15	Insufficient free disk	Free up disk space, or set `NOODARA_SKIP_RESOURCE_CHECK=1` to override deliberately.
16	The panel port is already in use	Set `NOODARA_PORT=<other>` and re-run.
17	Docker was installed via snap	Remove it (`sudo snap remove docker`) and re-run — Noodara does not support a snap-installed Docker.
20	Docker Engine installation failed	Read the `apt` output printed above the error and fix the underlying issue, or install Docker manually.
21	The Compose plugin is still missing after installation	Install `docker-compose-plugin` manually and re-run.
22	Docker is installed but its daemon is not responding	Start it (`sudo systemctl start docker`) and re-run. Nothing is touched — the installer never removes or reinstalls an existing Docker.
30	Writing `.env` (or a value in it) failed	The message names the exact variable or file operation that failed; fix it and re-run.
40	Could not resolve which version to install	Set `NOODARA_VERSION=<tag>` explicitly and re-run.
41	Could not resolve a public URL	Set `NOODARA_PUBLIC_URL=<url>` explicitly and re-run.
50	Pulling images failed	Check network connectivity and the resolved image tags, then re-run.
51	`docker compose up` failed	Read Compose's own error and the service state printed above. Data and secrets are untouched.
52	Database migrations failed	Read the `migrate` service's log tail printed above. Data and secrets are untouched.
53	A service did not become healthy in time	Read the named service's log tail printed above. Data and secrets are untouched; see Rollback if this happened during an upgrade.

Reading logs:

docker compose -f /opt/noodara/docker-compose.yml logs <service>

Checking health:

docker compose -f /opt/noodara/docker-compose.yml ps
docker compose -f /opt/noodara/docker-compose.yml exec -T api node -e "fetch('http://127.0.0.1:3000/health').then(r=>process.exit(r.ok?0:1)).catch(()=>process.exit(1))" && echo healthy
curl -I http://127.0.0.1:<port>/login

The ps HEALTH column reflects each service's own Compose healthcheck:. The middle command re-runs the api container's own healthcheck by hand, from inside the container against its /health route directly (the same call docker-compose.yml's api healthcheck itself makes) — only web publishes a port on the host, and web only proxies /api/* to api (see What gets installed), so /health is never reachable from outside the container. The last command confirms the panel itself is answering on the port you published.

Running the operator CLI inside the running api container:

docker compose -f /opt/noodara/docker-compose.yml exec api node dist/cli/index.js admin reset
docker compose -f /opt/noodara/docker-compose.yml exec api node dist/cli/index.js secrets rotate

admin reset issues a one-time recovery token for the existing admin account. secrets rotate re-encrypts every stored credential under a newly rotated master key.