xbot-server

/xbot-server — Set up remote service server access

Initialize the service catalog to discover ML services on a remote GPU server.

Usage

/xbot-server

What to do

1. Ask the user for server configuration. Use AskUserQuestion to collect:

   • Server IP — the remote GPU server address (e.g., 192.168.1.100)
   • SSH username — the user account on the server (e.g., gpu_user)
   • Service directory — the directory on the server containing deployed services (e.g., /home/gpu_user/services)

   Ask all three in one AskUserQuestion call. Do NOT hardcode or assume any values.

2. Set up SSH access. Test connectivity:

   ssh -o ConnectTimeout=5 -o BatchMode=yes USERNAME@SERVER_IP "echo connected"
   

   If it fails, ask the user to run ssh-copy-id USERNAME@SERVER_IP interactively:

   ! ssh-copy-id USERNAME@SERVER_IP
   

   Then verify again.

3. Add SSH config alias (if not already present). Add an entry to ~/.ssh/config:

   Host service-server
       HostName SERVER_IP
       User USERNAME
       IdentityFile ~/.ssh/id_ed25519
   

4. Run the setup script. Execute:

   cd ~/tidybot_uni/Tidybot-Universe/service-server-setup
   bash setup.sh --server-ip SERVER_IP --username USERNAME --service-dir SERVICE_DIR
   

   This will:

   • Verify SSH connectivity
   • Verify the service directory exists
   • Write config.json
   • Run an initial scan to discover services

5. Start the catalog server. Launch in the background:

   cd ~/tidybot_uni/Tidybot-Universe/service-server-setup
   python3 service_scanner.py &
   

   Wait for it to be ready:

   sleep 10 && curl -s http://localhost:8090/health
   

6. Show discovered services. Fetch and display the catalog:

   curl -s http://localhost:8090/services
   

   Format the output as a clear table for the user showing:

   • Service name
   • Port
   • Status (running/stopped)
   • URL
   • Number of endpoints