How do we create a plugin.so file with python code for a server in the Shadow TOR simulation ?

[Shubhi-73]

Function of the plug-in: to execute bash scripts and append the output to logs.

[jtracey]

Shadow doesn't need plugins any more, you can just configure processes directly. Check out the current documentation for some examples on how to do that. Note that executing bash scripts instead of doing everything in a python (or whatever) process will rely heavily on fork/exec. That won't work in Shadow 3.0. It might in the latest git version of Shadow, but it isn't fully implemented (see the relevant milestone for some hints on what does and doesn't work).

[robgjansen]

We have an example of running a python process on a server here:
https://shadow.github.io/docs/guide/getting_started_basic.html#configuring-the-simulation

[Shubhi-73]

Thank you for the resources @robgjansen @jtracey, I was able to execute a few scripts.
I’m planning to run the python code (involves scanning for open ports) directly as a process in the simple TOR network example. I’m repeatedly getting errors. Are there any changes in configurations that are required?
Note- the cURL example with python3 http.server is working just fine.

Shadow.yaml

relay4:
    network_node_id: 0
    processes:
    - path: tor
      args: --Address relay4 --Nickname relay4
            --defaults-torrc torrc-defaults -f torrc
      start_time: 60
      expected_final_state: running
    - path: python3
      args: ~/shadow/examples/docs/tor/test.py
      start_time: 60
      expected_final_state: running

Console

** Starting Shadow v3.0.0-898-g25cb58fc1 2023-12-07--20:00:41 with GLib v2.72.4
00:00:00.687688 [160790:shadow-worker] 00:01:00.000464030 [ERROR] [relay4:11.0.0.8] [process.rs:1527] [shadow_rs::host::process] process 'relay4.python3.10.1001' exited with status Normal(2); expected e
nd state was running but was exited: 2
00:00:05.034148 [160768:shadow] n/a [ERROR] [n/a] [main.rs:403] [shadow_rs::core::main::export] Failed to run the simulation
00:00:05.034335 [160768:shadow] n/a [ERROR] [n/a] [main.rs:403] [shadow_rs::core::main::export] 
00:00:05.034470 [160768:shadow] n/a [ERROR] [n/a] [main.rs:403] [shadow_rs::core::main::export] Caused by:
00:00:05.034533 [160768:shadow] n/a [ERROR] [n/a] [main.rs:403] [shadow_rs::core::main::export]     1 managed processes in unexpected final state
** Shadow did not complete successfully: Failed to run the simulation

[jtracey]

That means the process exited with an error value (2), you should check the process's logs in the shadow.data/hosts/relay4/ directory to see what went wrong.

[robgjansen]

It looks to me like you told shadow that your python3 process should be running at the end of the simulation (expected_final_state: running) but your test.py script actually ended before the simulated ended. If you actually intended your test script to end, then you can just change your python3 process to have expected_final_state: {exited: 0}.

See here: https://shadow.github.io/docs/guide/shadow_config_spec.html#hostshostnameprocessesexpected_final_state

[stevenengler]

Your problem is the line args: ~/shadow/examples/docs/tor/test.py. Shadow doesn't do shell expansion (such as ~ to /home/me/) of program arguments. Python returns an exit code of 2 when the given file isn't found. Like mentioned above, you should see Python's error message in the stderr log file. And you'll also still want the expected_final_state fix.