Deployment troubleshooting

---

Installation and environment

SELinux blocking container access (Linux only)

SELinux is a default deployment on many Linux installations and may block Docker containers from accessing configuration files.

Run this command to grant access:

chcon -Rt svirt_sandbox_file_t prometheus grafana trustgraph

Also apply this to any LLM-specific configuration e.g. the vertexai directory.

Processors not running

Check processor state with:

tg-show-processor-state

This shows the state of running processors in Prometheus. If the list is empty or you get an error, Prometheus may not be running or accessible.

Configuration service not responding

Check the configuration service with:

tg-show-config

If you get an error, Pulsar or the configuration service may not be running.

If the output says version 0 and shows an empty configuration, initialisation may have failed.

---

CLI commands not working

Unsupported Python version

TrustGraph requires Python 3.12 or later.

Missing trustgraph-cli package

If CLI commands aren’t found, revisit the installation instructions: installation

Version mismatch

Your trustgraph-cli package version should match the deployed TrustGraph version. At minimum, ensure the major/minor numbers match and both versions are declared stable.

---

LLM invocation not working with LMStudio / Ollama

Incorrect host URL configuration

When accessing Ollama or LMStudio running on the host from inside a container, you must use a service address that’s accessible from within containers.

Linux with Podman:

host.containers.internal

Docker on Linux or Windows (not WSL):

host.docker.internal

WSL with Docker Desktop:

WSL is more complex due to the host, virtualised WSL environment, and Docker containers.

Find WSL2’s IP from Windows command prompt or PowerShell:

wsl hostname -I

Then use that IP in your docker-compose file:

OLLAMA_HOST=http://[WSL2_IP]:11434