Error connecting to local llm

[ashish-kumar-hpe]

Hi,
I have configured my local vscode continue extension to point to local nvidia nim codellama model and my configuration looks something like this.

name: Config
version: 1.0.0
schema: v1

assistants:

• name: default
  model: CodeLlama

models:

• name: CodeLlama
  provider: ollama
  model: codellama/codellama-13b-instruct
  apiBase: https://ashish-code-llama-1.project-user-ashish-kumar.serving.adt-alto01-ingress.us.rdlabs.hpecorp.net/v1
  apiKey: '**************'
  title: codellama:13b
  roles:
  • chat
  • edit
  • autocomplete

I am getting following error while connecting to llm...
request to https://ashish-code-llama-1.project-user-ashish-kumar.serving.adt-alto01-ingress.us.rdlabs.hpecorp.net/v1/api/chat failed, reason: unable to verify the first certificate

My question is that how can i configure my ssl certificate since my endpoint is https.

[bdougie]

Hi @ashish-kumar-hpe,

The SSL certificate error you're experiencing ("unable to verify the first certificate") occurs when connecting to HTTPS endpoints with self-signed or custom certificates. Here are solutions for Continue:

Solution 1: Configure Certificate Bundle Path (Recommended)

Add requestOptions.caBundlePath to your model configuration:

name: Config
version: 1.0.0
schema: v1

assistants:
  - name: default
    model: CodeLlama

models:
  - name: CodeLlama
    provider: ollama
    model: codellama/codellama-13b-instruct
    apiBase: https://ashish-code-llama-1.project-user-ashish-kumar.serving.adt-alto01-ingress.us.rdlabs.hpecorp.net/v1
    apiKey: '**************'
    title: codellama:13b
    requestOptions:
      caBundlePath: /path/to/your/certificate.pem  # Add this line
    roles:
      - chat
      - edit
      - autocomplete

Replace /path/to/your/certificate.pem with the actual path to your SSL certificate file.

Solution 2: Client Certificate Authentication

If your endpoint requires client certificate authentication:

models:
  - name: CodeLlama
    provider: ollama
    model: codellama/codellama-13b-instruct
    apiBase: https://ashish-code-llama-1.project-user-ashish-kumar.serving.adt-alto01-ingress.us.rdlabs.hpecorp.net/v1
    apiKey: '**************'
    requestOptions:
      clientCertificate:
        cert: /path/to/client.pem
        key: /path/to/client.key
        passphrase: your_passphrase  # Optional if key is encrypted

Solution 3: Windows VS Code Users

If you're on Windows, installing the win-ca extension in VS Code may automatically resolve certificate issues.

Troubleshooting Steps:

1. Verify the certificate path: Ensure the certificate file exists and is readable
2. Test the endpoint: Try accessing your endpoint with curl to verify connectivity
3. Check certificate format: Ensure your certificate is in PEM format
4. Check logs: Look at ~/.continue/logs/core.log for more detailed error messages

You can find more details in our FAQ documentation about configuring certificates.

Let me know if this resolves your issue or if you need further assistance!