run-it

Run commands on host machine via http request.

Features

• Core feature is, to run a command on the host and command is submitted via a http request
• Application records the commands its running
• Couple of APIs are implemented
  • /api/submitcmd -> to submit a command via POST method
  • /api/getcmdstatus/:cmd_id -> to get status of the command, cmd_id would be the response of the POST request
• Logging/Tracing is implemented to track the command flow
• Live commands outputs are also captured, a configurable timeout setting for long running commands

Configurable options

• Below are the environment variables which can be set to overwrite default values
  • RUST_LOG: determine log level, Default is INFO.
  • R_CMD_TIMEOUT: Time out value in sec, only numerics, example: 100 for 100sec timeout, Default is 50
  • RUN_IT_PORT: port number to listen on, Default is 48786

DEMO

quick-demo.mov

How to

Build locally

• To build locally it is required to have rust and cargo installed
• Clone the repo and run: cargo build -r
• This would leave a executable in <clone_repo_loc>/target/release/run-it path
• If above environment variables are set those values are used otherwise default values are used.

Run it

• To run this on a linux environment as systemd service, please follow the steps here

Submit commands

• Service has an api exposed, /api/submitcmd, send a POST request with data in json format, hence need the header 'Content-Type: application/json'

  • data has can have these keys

  {
    "cmd": string<required>,
    "args": string<optional>,
    "is_shell": bool<optional>
  }
  

  • cmd - is the command need to run, string type

  • args - if the command accepts any args, string type

  • is_shell - if the cmd passed is a shell like bash or sh and args have the actual task, then this need to be set true, yes its bool type

  • examples:

    • curl http://<IP_ADDRESS>:48786/api/submitcmd -X POST -H 'Content-Type: application/json' -d '{"cmd": "whoami"}'
      • this is a basic example runs cmd whoami on the host and request responds with a unique id which can be used to lookup the results
    • curl http://<IP_ADDRESS>:48786/api/submitcmd -X POST -H 'Content-Type: application/json' -d '{"cmd": "python3", "args": "--version"}'
      • this is running a pyton command and with an arg, if host doesn't have python installed command would fail and its corresponding output is available on the /api/getcmdstatus api
    • curl http://<IP_ADDRESS>:48786/api/submitcmd -X POST -H 'Content-Type: application/json' -d '{"cmd": "bash", "args": "c=0; while true; do let c=c+1; sleep 10; echo $c; done", "is_shell": true}'
      • this is a final variant, cmd is shell, and args have actual single line script, in this scenario set the is_shell to true. and if you notice the script is set to run forever, in these long running session cmd will timeout after the configured setting.

Get command status

• Service also provides an API to get the command status, to get the status request a GET call on /api/getcmdstatus/<CMD_ID>, yes path should have the command ID that POST call responded with.
  • example:
    • curl http://<IP_ADDRESS>:48786/api/getcmdstatus/13995226725389956087 -X GET | jq as the response is json it can be piped to jq

    {
      "state": "",
      "output": ""
    }
    

    If the CMD_ID is not found an empty output is returned otherwise it will have the state and output at that point of time, i mean if the cmd is still active it would return that point in time output and state would be in-progress

TODO

• Valid unit tests
• Ability to upload a script and run it
• external datastore
• server/clinet arch
  • server will have an UI
  • server can not accept commands via api
  • client may or may not an ui module