Appliance Deployment Guide
Welcome to the official NoCloudGPT support documentation. This guide maps directly to the automated installation sequence of your appliance. To protect the security and integrity of your private cloud environment, our deployment logic is closed-source. However, the exact configuration requirements and verification steps for each stage are detailed in this guide.
Understanding Your Installation Summary
At the end of a successful boot, your instance will generate an install-summary.txt and instance.env file. These files contain the routing information needed to access your appliance. Below is an example of what to look for, with sensitive IPs and domains redacted.
The Boot Sequence (Stages 0-7)
If you encounter an error in your ycgpt-log.txt, locate the stage number below and click through to find the corresponding troubleshooting steps, expected variables, and path structures.
- 0.0 Environment: Validates instance configurations, swap space, and base Docker dependencies.
- 1.0 Networking: Establishes NGINX reverse proxies, self-signed TLS certificates, and domain routing.
- 2.0 Ollama AI Engine: Configures local bind addresses and verifies the internal AI API layer.
- 3.0 OpenWebUI: Pulls canary models, removes legacy containers, and handles Docker Compose startup.
- 4.0 Landing Screen: Generates the offline appliance directory structure (eNotes) and HTML front-end.
- 5.0 Team Wiki: Installs PHP 8.3, assigns DokuWiki permissions, and binds the NGINX fastcgi socket.
- 6.0 Extractor Control Plane: Configures Python export APIs and Systemd services for background administration.
- 7.0 Administration: Cleans up installation artifacts and records the final service validation summary.