Collects system information from test endpoints for reproducibility. Gathers hardware, OS, kernel, and configuration details that may affect benchmark results.
Bash — single script packrat (~715 lines)
| File | Purpose |
|---|---|
packrat |
Main collection script — runs on each endpoint during a test |
rickshaw.json |
Declares integration with rickshaw (roles, parameters) |
workshop.json |
Engine image build requirements (packrat runs on engines/endpoints, not the controller) |
14 parallel collectors: distro, package_manager, env, cpu, device, system, kernel, block, memory, net, cgroup, libvirt, firewall, container
- Primary branch is
master - Creates a
packrat-archive/directory containing collected data, compressed with xz - Logs activity to
packrat.log - Runs as part of the engine script execution phase during a benchmark
- Uses standard Bash modelines and 4-space indentation