virthttp

The purpose of this project is to provide a web HTTP based frontend to libvirt. We use Boost.Beast, a minimal C++ HTTP server that is used to communicate between the web and the libvirt daemon. We use JSON to transmit data between the server and the client.

We also created a C++ binding for libvirt, and use it to interface with libvirt.

Few usage examples

A cURL request to list all domains

curl -X GET "http://localhost:8081/libvirt/domains" \
     -H "Content-Type:application/json" \
     -H "X-Auth-Key:1234567893feefc5f0q5000bfo0c38d90bbeb" \
     -H "X-Auth-Username:smith"

A JSon return listing all domains

{
  "results": [
    {
      "name": "vm1",
      "uuid": "f6ah2js8-dfgv-3f3f-fgs1-d2s09dhjej83",
      "id": -1,
      "status": "Shutoff",
      "os": "hvm",
      "ram": 4194304,
      "ram_max": 4194304,
      "cpu": 4
    },
    {
      "name": "vm2",
      "uuid": "8shfjaif-f8s9-12vg-ger5-e5hbt78n4bhe",
      "id": 2,
      "status": "Running",
      "os": "hvm",
      "ram": 4194304,
      "ram_max": 4194304,
      "cpu": 4
    }
  ],
  "success": true,
  "errors": [],
  "messages": []
}

For more examples and a usage reference, see the API documentation

Building

!! IMPORTANT !!
While compiling for any other target than Linux or BSD based OSes, there is no platform-dependent code in this repository. Also, we currently only test with Clang [7,8,9] and GCC [8,9]

Dependencies (debian packages)

• libvirt-dev
• libboost-system-dev
• libdeflate-dev

Build steps

Getting sources

$ git clone https://github.com/HeavyEyelid/virthttp.git
$ git submodule update --init --recursive

Compiling

$ cd virthttp/
$ mkdir build
$ cd build/
$ cmake ..
$ make -j $(nproc) # or simply 'ninja' if you use ninja-build 

Generating developer documentation

This project uses Doxygen for its developer documentation.
To generate the HTML version from the code, simply run doxygen Doxyfile at the root directory of this repository; you will find the generated HTML in a new doc/ folder.