Facility to bypass cli start to enable concurrency with other actors #1243

casibbald · 2025-02-04T08:08:41Z

casibbald
Feb 4, 2025

I would like to use Loco in my Kubernetes controller to present a UI and aspects that are easier to consume visually rather than kubectl.

I have a concept working where Loco runs and the Kubecontroller runs, however Loco is blocking and the Kubecontroller only starts up when Loco receives its Shutdown signal.

I would like to be able to bypass the cli start that is traditionally setup as follows:

use loco_rs::cli;
use yapp::{
    app::App,
    controllers::{kubecontroller::run, telemetry},
    core::kubecontroller::State,
};

#[tokio::main]
async fn main() -> loco_rs::Result<()> {
    cli::main::<App>().await.expect("TODO: panic message");

    // Initialize Kubernetes controller state
    let state = State::default();
    run(state.clone()).await; // Ensure `run` is awaited
    Ok(())
}

and I was thinking of something like the following which I have compiling but still have issues getting it to work.:

use std::str::FromStr;
use loco_rs::{app::Hooks, boot::StartMode, environment::Environment, prelude::*};
use std::time::Duration;

use tokio::signal;
use yapp::{
    app::App,
    controllers::kubecontroller::run,
    core::{environment::EnvironmentExt, kubecontroller::State},
};

#[warn(dead_code)]
async fn run_kubecontroller() {
    let state = State::default();
    run(state.clone()).await;
}

struct ServeParams {
    binding: String,
    port: i32,
}

#[allow(clippy::redundant_pub_crate)]
#[tokio::main(flavor = "multi_thread")]
async fn main() -> Result<()> {
    // 1. Environment setup ---------------------------------------------------
    let environment = std::env::var("ENVIRONMENT")
        .map_or(Environment::Development, |env| <Environment as FromStr>::from_str(&env).unwrap_or(Environment::Development));
    println!("🚀 Starting in {environment} environment");

    // 2. Config loading ------------------------------------------------------
    let config = App::load_config(&environment)
        .await
        .map_err(|e| {
            eprintln!("Failed to load config: {e:?}");
            e
        })?;

    // 3. Logger initialization -----------------------------------------------
    if !App::init_logger(&config, &environment).expect("Failed to initialize logger") {
        // Default Loco logger if not overridden
        tracing_subscriber::fmt()
            .with_target(false)
            .with_max_level(environment.log_level())
            .init();
    }

    // 4. Application boot ----------------------------------------------------
    let boot = App::boot(StartMode::ServerAndWorker, &environment)
        .await
        .expect("Failed to boot application");

    // 5. Server parameters ---------------------------------------------------
    let serve_params = ServeParams {
        binding: config.server.host.clone(),
        port: config.server.port,
    };

    // 6. Concurrent tasks setup ----------------------------------------------
    let (_tx, rx) = tokio::sync::oneshot::channel();

    // Server task
    let server_task = tokio::spawn({
        let ctx = boot.app_context.clone();
        let (tx, _rx) = tokio::sync::oneshot::channel();
        let router = boot.router.clone();
        async move {
            tracing::info!(
                "Starting server on {}:{}",
                serve_params.binding,
                serve_params.port
            );
            App::serve(router.expect("REASON"), &ctx)
                .await
                .expect("Server failed");

            tx.send(()).expect("Failed to send shutdown signal");
        }
    });

    // Background task
    let background_task = tokio::spawn(async move {
        tracing::info!("Background task started");
        let mut rx: tokio::sync::oneshot::Receiver<()> = rx;
        tokio::spawn(run_kubecontroller());
        loop {
            tokio::select! {
                _ = &mut rx => {
                    tracing::info!("Background task received shutdown signal");
                    break;
                }
                () = tokio::time::sleep(Duration::from_secs(5)) => {
                    tracing::info!("Background task heartbeat");
                }
            }
        }
    });

    // 7. Graceful shutdown ---------------------------------------------------
    tokio::select! {
        _ = signal::ctrl_c() => {
            tracing::info!("Received CTRL+C, shutting down");
        }
        _ = server_task => {
            tracing::info!("Server task completed");
        }
    }

    // 8. Cleanup -------------------------------------------------------------
    background_task.abort();
    App::on_shutdown(&boot.app_context).await;

    Ok(())
}

Log of currently observed blocking behaviour from the first code snippet:

      Building image
  Beginning garbage collecting Kubernetes objects
  Deleting kubernetes objects:
  → Service/yapp-controller
  → Deployment/yapp-controller
  **2025-02-01T19:15:41.919127Z  INFO loco_rs::app: shutting down...** <<< ------ Loco shuts down, releasing the controller to start working. However They should run concurrently, and this is what I am looking to address.
  2025-02-01T19:15:41.919638Z  INFO yapp::core::telemetry: Global default subscriber already set!
  2025-02-01T19:15:41.919936Z DEBUG tower::buffer::worker: service.ready=true processing request
  2025-02-01T19:15:41.920014Z DEBUG HTTP: kube_client::client::builder: requesting http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.920122Z DEBUG HTTP: hyper_util::client::legacy::connect::http: connecting to 10.96.0.1:443 http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.920332Z DEBUG HTTP: hyper_util::client::legacy::connect::http: connected to 10.96.0.1:443 http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.920368Z DEBUG HTTP: rustls::client::hs: No cached session for IpAddress(V4(Ipv4Addr([10, 96, 0, 1])))     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.920437Z DEBUG HTTP: rustls::client::hs: Not resuming any session     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.921841Z DEBUG HTTP: rustls::client::hs: Using ciphersuite TLS13_AES_128_GCM_SHA256     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.921854Z DEBUG HTTP: rustls::client::tls13: Not resuming     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.921932Z DEBUG HTTP: rustls::client::tls13: TLS1.3 encrypted extensions: []     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.921942Z DEBUG HTTP: rustls::client::hs: ALPN protocol is None     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.921953Z DEBUG HTTP: rustls::client::tls13: Got CertificateRequest CertificateRequestPayloadTls13 { context: , extensions: [Unknown(UnknownExtension { typ: StatusRequest, payload:  }), Unknown(UnknownExtension { typ: SCT, payload:  }), SignatureAlgorithms([RSA_PSS_SHA256, ECDSA_NISTP256_SHA256, ED25519, RSA_PSS_SHA384, RSA_PSS_SHA512, RSA_PKCS1_SHA256, RSA_PKCS1_SHA384, RSA_PKCS1_SHA512, ECDSA_NISTP384_SHA384, ECDSA_NISTP521_SHA512, RSA_PKCS1_SHA1, ECDSA_SHA1_Legacy]), AuthorityNames([DistinguishedName(3015311330110603550403130a6b756265726e65746573), DistinguishedName(3019311730150603550403130e66726f6e742d70726f78792d6361)])] }     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.921965Z DEBUG HTTP: rustls::client::common: Client auth requested but no cert/sigscheme available     http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.925026Z DEBUG HTTP: hyper_util::client::legacy::pool: pooling idle connection for ("https", 10.96.0.1) http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=1 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.925219Z  INFO kube_runtime::controller: press ctrl+c to shut down gracefully
  2025-02-01T19:15:41.925238Z DEBUG kube_runtime::controller: applier runner held until store is ready
  2025-02-01T19:15:41.925380Z DEBUG tower::buffer::worker: service.ready=true processing request
  2025-02-01T19:15:41.925438Z DEBUG HTTP: kube_client::client::builder: requesting http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=500 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.925454Z DEBUG HTTP: hyper_util::client::legacy::pool: reuse idle connection for ("https", 10.96.0.1) http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&limit=500 otel.name="list" otel.kind="client"
  2025-02-01T19:15:41.927285Z DEBUG hyper_util::client::legacy::pool: pooling idle connection for ("https", 10.96.0.1)
  2025-02-01T19:15:41.927671Z DEBUG kube_runtime::controller: store is ready, starting runner
  2025-02-01T19:15:41.927743Z DEBUG tower::buffer::worker: service.ready=true processing request
  2025-02-01T19:15:41.927821Z DEBUG HTTP: kube_client::client::builder: requesting http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&watch=true&timeoutSeconds=290&allowWatchBookmarks=true&resourceVersion=1143 otel.name="watch" otel.kind="client"
  2025-02-01T19:15:41.927861Z DEBUG HTTP: hyper_util::client::legacy::pool: reuse idle connection for ("https", 10.96.0.1) http.method=GET http.url=https://10.96.0.1/apis/kube.rs/v1/documents?&watch=true&timeoutSeconds=290&allowWatchBookmarks=true&resourceVersion=1143 otel.name="watch" otel.kind="client"
  2025-02-01T19:15:41.933259Z  INFO reconciling object:reconcile: yapp::core::kubecontroller: Reconciling Document document_name=samuel namespace=default object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel"
  2025-02-01T19:15:41.933342Z DEBUG reconciling object:reconcile: tower::buffer::worker: service.ready=true processing request object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel"
  2025-02-01T19:15:41.933377Z DEBUG reconciling object:reconcile:HTTP: kube_client::client::builder: requesting object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.933452Z DEBUG reconciling object:reconcile:HTTP: hyper_util::client::legacy::connect::http: connecting to 10.96.0.1:443 object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.934859Z  INFO reconciling object:reconcile: yapp::core::kubecontroller: Reconciling Document document_name=lorem namespace=default object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem"
  2025-02-01T19:15:41.935294Z DEBUG reconciling object:reconcile: tower::buffer::worker: service.ready=true processing request object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem"
  2025-02-01T19:15:41.935313Z DEBUG reconciling object:reconcile:HTTP: kube_client::client::builder: requesting object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.935379Z DEBUG reconciling object:reconcile:HTTP: hyper_util::client::legacy::connect::http: connecting to 10.96.0.1:443 object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937269Z  INFO reconciling object:reconcile: yapp::core::kubecontroller: Reconciling Document document_name=illegal namespace=default object.ref=Document.v1.kube.rs/illegal.default object.reason=object updated document="illegal"
  2025-02-01T19:15:41.937313Z DEBUG reconciling object:reconcile:HTTP: hyper_util::client::legacy::connect::http: connected to 10.96.0.1:443 object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937374Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: Resuming session     object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937486Z DEBUG reconciling object:reconcile:HTTP: hyper_util::client::legacy::connect::http: connected to 10.96.0.1:443 object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937490Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: No cached session for IpAddress(V4(Ipv4Addr([10, 96, 0, 1])))     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937503Z  WARN reconciling object: yapp::core::kubecontroller: reconcile failed: Any(ApplyFailed(Any(Custom { kind: Other, error: "IllegalDocument" }))) object.ref=Document.v1.kube.rs/illegal.default object.reason=object updated
  2025-02-01T19:15:41.937531Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: Not resuming any session     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937832Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: Using ciphersuite TLS13_AES_128_GCM_SHA256     object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937848Z DEBUG reconciling object:reconcile:HTTP: rustls::client::tls13: Resuming using PSK     object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937894Z DEBUG reconciling object:reconcile:HTTP: rustls::client::tls13: TLS1.3 encrypted extensions: []     object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.937897Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: ALPN protocol is None     object.ref=Document.v1.kube.rs/lorem.default object.reason=object updated document="lorem" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/lorem/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.939045Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: Using ciphersuite TLS13_AES_128_GCM_SHA256     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.939061Z DEBUG reconciling object:reconcile:HTTP: rustls::client::tls13: Not resuming     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.939105Z DEBUG reconciling object:reconcile:HTTP: rustls::client::tls13: TLS1.3 encrypted extensions: []     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.939108Z DEBUG reconciling object:reconcile:HTTP: rustls::client::hs: ALPN protocol is None     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.939128Z DEBUG reconciling object:reconcile:HTTP: rustls::client::tls13: Got CertificateRequest CertificateRequestPayloadTls13 { context: , extensions: [Unknown(UnknownExtension { typ: StatusRequest, payload:  }), Unknown(UnknownExtension { typ: SCT, payload:  }), SignatureAlgorithms([RSA_PSS_SHA256, ECDSA_NISTP256_SHA256, ED25519, RSA_PSS_SHA384, RSA_PSS_SHA512, RSA_PKCS1_SHA256, RSA_PKCS1_SHA384, RSA_PKCS1_SHA512, ECDSA_NISTP384_SHA384, ECDSA_NISTP521_SHA512, RSA_PKCS1_SHA1, ECDSA_SHA1_Legacy]), AuthorityNames([DistinguishedName(3015311330110603550403130a6b756265726e65746573), DistinguishedName(3019311730150603550403130e66726f6e742d70726f78792d6361)])] }     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.939140Z DEBUG reconciling object:reconcile:HTTP: rustls::client::common: Client auth requested but no cert/sigscheme available     object.ref=Document.v1.kube.rs/samuel.default object.reason=object updated document="samuel" http.method=PATCH http.url=https://10.96.0.1/apis/kube.rs/v1/namespaces/default/documents/samuel/status?&force=true&fieldManager=cntrlr otel.name="patch_status" otel.kind="client"
  2025-02-01T19:15:41.942417Z DEBUG hyper_util::client::legacy::pool: pooling idle connection for ("https", 10.96.0.1)
  2025-02-01T19:15:41.943350Z DEBUG hyper_util::client::legacy::pool: pooling idle connection for ("https", 10.96.0.1)

Answered by kaplanelad

Feb 16, 2025

Steps to Run Loco Without the CLI

Remove the CLI Feature
Open Cargo.toml and remove the cli feature from the loco-rs dependency.
Replace:
```
loco-rs = { workspace = true, features = ["cli"] }
```
With:
```
loco-rs = { workspace = true }
```
Add a Custom Initializer
Navigate to src/bin/main.rs and replace its contents with the following code:

use yair::{
    app::App,
    controllers::{kubecontroller::run, telemetry},
    core::kubecontroller::State,
};

#[tokio::main]
async fn main() -> loco_rs::Result<()> {
    let loco_rs_handle = tokio::spawn(async {
        if let Err(e) = run_loco_rs().await {
            eprintln!("Error in loco_rs: {:?}", e);
        }
    });

    // Await all tasks to comp…

View full answer

kaplanelad · 2025-02-06T10:21:30Z

kaplanelad
Feb 6, 2025
Maintainer

You can turn off the cli feature and boot the server or other component by yourself

6 replies

kaplanelad Feb 6, 2025
Maintainer

you can follow the cli.rs code.

casibbald Feb 15, 2025
Author

@kaplanelad Do you know anyone by chance who would be able to assist with this? even for a small bounty?
I have failed to get the bypass working, even following cli.rs, I still don't have any success

kaplanelad Feb 15, 2025
Maintainer

do you have a repo that i can take a look?

casibbald Feb 15, 2025
Author

casibbald/yair-controller#49

casibbald Feb 15, 2025
Author

casibbald/yair-controller#50

casibbald · 2025-02-15T18:05:47Z

casibbald
Feb 15, 2025
Author

The main branch starts with tilt up, and loco runs, and on restarting when loco terminates the controller reconciles its custom resources and exits.

So if the cli is bypassed and I can have the controller and loco running concurrently I think I will be in a good place to continue.

2 replies

kaplanelad Feb 16, 2025
Maintainer

following the prs I'm not sure what you're trying to do.

Are you trying to run Loco directly from the main function without using the CLI?

casibbald Feb 16, 2025
Author

Yes.

So the loco cli is not async and it blocks and therefore I have been unable to run it concurrently with the controller code.

I have tried things like this:

#[tokio::main]
async fn main() -> loco_rs::Result<()> {
    // Initialize telemetry once
    telemetry::init().await;

    // Spawn the loco_rs task
    let loco_rs_handle = tokio::spawn(async {
        if let Err(e) = run_loco_rs().await {
            eprintln!("Error in loco_rs: {:?}", e);
        }
    });

    // Spawn the kubecontroller task
    let kubecontroller_handle = tokio::spawn(async {
        run_kubecontroller().await;
    });

    // Await all tasks to complete
    let _ = tokio::try_join!(loco_rs_handle, kubecontroller_handle)?;

    Ok(())
}

async fn run_loco_rs() -> loco_rs::Result<()> {
    println!("Starting loco_rs...");
    let result = cli::main::<App>().await;
    if let Err(ref e) = result {
        eprintln!("Error in loco_rs: {:?}", e);
    }
    result
}

async fn run_kubecontroller() {
    let state = State::default();
    run(state.clone()).await;
}

as well, but none of these work.

I tried the example in the original post to try bypass the CLI but have not been able to get any of these scenarios to work.

kaplanelad · 2025-02-16T08:51:55Z

kaplanelad
Feb 16, 2025
Maintainer

Steps to Run Loco Without the CLI

Remove the CLI Feature
Open Cargo.toml and remove the cli feature from the loco-rs dependency.
Replace:
```
loco-rs = { workspace = true, features = ["cli"] }
```
With:
```
loco-rs = { workspace = true }
```
Add a Custom Initializer
Navigate to src/bin/main.rs and replace its contents with the following code:

use yair::{
    app::App,
    controllers::{kubecontroller::run, telemetry},
    core::kubecontroller::State,
};

#[tokio::main]
async fn main() -> loco_rs::Result<()> {
    let loco_rs_handle = tokio::spawn(async {
        if let Err(e) = run_loco_rs().await {
            eprintln!("Error in loco_rs: {:?}", e);
        }
    });

    // Await all tasks to complete
    let _ = tokio::try_join!(loco_rs_handle)?;

    Ok(())
}


async fn run_loco_rs() -> loco_rs::Result<()> {
    println!("Starting loco_rs...");

    let start_mode = loco_rs::boot::StartMode::ServerOnly;
    let environment = loco_rs::environment::Environment::Development;
    let boot_result = loco_rs::boot::create_app::<App>(start_mode, &environment).await?;


    let server_params = loco_rs::boot::ServeParams {
        port: 5150,
        binding: "localhost".to_string(),
    };
    loco_rs::boot::start::<App>(boot_result, server_params, false).await?;
    Ok(())
}

With this setup, you can run Loco without the CLI.

6 replies

kaplanelad Feb 17, 2025
Maintainer

The CLI initializes the logger itself, so you have two options:

Initialize your preferred logger, such as tracing or env_logger in your main.rs file.
Follow the same approach as Loco does here.

casibbald Feb 17, 2025
Author

I went with option one tracing. Thank you so much!

Can I 'buy you a coffee' for your help? I was in quite a rut!

kaplanelad Feb 17, 2025
Maintainer

You're more than welcome! 😊 We have a sponsors page if you’d like to support: https://github.com/sponsors/loco-rs.

casibbald Feb 17, 2025
Author

Perfect, I intend to use loco quite a bit, will add myself

kaplanelad Feb 17, 2025
Maintainer

thanks

kaplanelad · 2025-02-17T08:37:14Z

kaplanelad
Feb 17, 2025
Maintainer

You're more than welcome! 😊 We have a sponsors page if you’d like to support: https://github.com/sponsors/loco-rs.

0 replies

Uh oh!

Facility to bypass cli start to enable concurrency with other actors #1243

Uh oh!

Uh oh!

casibbald Feb 4, 2025

Replies: 4 comments · 14 replies

Uh oh!

kaplanelad Feb 6, 2025 Maintainer

Uh oh!

kaplanelad Feb 6, 2025 Maintainer

Uh oh!

casibbald Feb 15, 2025 Author

Uh oh!

kaplanelad Feb 15, 2025 Maintainer

Uh oh!

casibbald Feb 15, 2025 Author

Uh oh!

casibbald Feb 15, 2025 Author

Uh oh!

casibbald Feb 15, 2025 Author

Uh oh!

kaplanelad Feb 16, 2025 Maintainer

Uh oh!

Uh oh!

casibbald Feb 16, 2025 Author

Uh oh!

kaplanelad Feb 16, 2025 Maintainer

Uh oh!

kaplanelad Feb 17, 2025 Maintainer

Uh oh!

casibbald Feb 17, 2025 Author

Uh oh!

kaplanelad Feb 17, 2025 Maintainer

Uh oh!

casibbald Feb 17, 2025 Author

Uh oh!

kaplanelad Feb 17, 2025 Maintainer

Uh oh!

kaplanelad Feb 17, 2025 Maintainer

casibbald
Feb 4, 2025

Replies: 4 comments 14 replies

kaplanelad
Feb 6, 2025
Maintainer

kaplanelad Feb 6, 2025
Maintainer

casibbald Feb 15, 2025
Author

kaplanelad Feb 15, 2025
Maintainer

casibbald Feb 15, 2025
Author

casibbald Feb 15, 2025
Author

casibbald
Feb 15, 2025
Author

kaplanelad Feb 16, 2025
Maintainer

casibbald Feb 16, 2025
Author

kaplanelad
Feb 16, 2025
Maintainer

kaplanelad Feb 17, 2025
Maintainer

casibbald Feb 17, 2025
Author

kaplanelad Feb 17, 2025
Maintainer

casibbald Feb 17, 2025
Author

kaplanelad Feb 17, 2025
Maintainer

kaplanelad
Feb 17, 2025
Maintainer