1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
#![warn(missing_docs)]
//! A crate to help build custom gitlab runner implementations.
//!
//! ## Implementing a custom runner
//!
//! The overall idea is that this crate handles all interaction with the gitlab
//! server and drives the executions while the the runner implementation focus on
//! how to handle jobs from gitlab.
//!
//! As the focus for an implementer of a custom runner is to implement the async
//! JobHandler trait, which gets calle during job executation. An absolute minimal
//! runner can be implement as such:
//!
//! ```rust,no_run
//! use gitlab_runner::{outputln, Runner, JobHandler, JobResult, Phase};
//! use std::path::PathBuf;
//!
//! #[derive(Debug)]
//! struct Run {}
//!
//! #[async_trait::async_trait]
//! impl JobHandler for Run {
//! async fn step(&mut self, script: &[String], phase: Phase) -> JobResult {
//! outputln!("Running script for phase {:?}", phase);
//! for s in script {
//! outputln!("Step: {}", s);
//! }
//! Ok(())
//! }
//! }
//!
//! #[tokio::main]
//! async fn main() {
//! let mut runner = Runner::new(
//! "https://gitlab.example.com".try_into().unwrap(),
//! "runner token".to_owned(),
//! PathBuf::from("/tmp"));
//! runner.run(move | _job | async move { Ok(Run{}) }, 16).await.unwrap();
//! }
//! ```
//!
//! ## Gitlab runner registration
//!
//! This crate does not support registering new runners with the gitlab server, so this has to be
//! done by hand using the gitlab
//! [runner registration API](https://docs.gitlab.com/ee/api/runners.html#register-a-new-runner).
//!
//! The registration token can be retrieved from the runners section in the Gitlab
//! administration area. With that token the runner can be register using a curl
//! command like:
//! ```shell
//! curl --request POST "https://GITLAB_URL/api/v4/runners" \
//! --form "description=My custom runner" \
//! --form "run_untagged=false" \
//! --form "tag_list=custom-gitlab-runner" \
//! --form "token=REGISTRATION_TOKEN"
//! ```
//!
//! As a response to this command a new token for the registered runner will be
//! provided, this token should be provided to the runner for it's gitlab
//! connection.
//!
//! One thing to key parameter provided here is `run_untagged=false`, which will
//! make the runner *only* pickup jobs which matches its tag. This is important to
//! prevent the runner from picking up "normal" jobs which it will not be able to
//! process.
pub mod artifact;
mod client;
mod logging;
use crate::client::Client;
mod run;
use crate::run::Run;
pub mod job;
use job::{Job, JobData};
pub mod uploader;
pub use logging::GitlabLayer;
use tracing::instrument::WithSubscriber;
mod runlist;
use crate::runlist::RunList;
use uploader::Uploader;
use futures::prelude::*;
use std::path::PathBuf;
use tokio::time::{sleep, Duration};
use tracing::warn;
use tracing_subscriber::{prelude::*, Registry};
use url::Url;
#[doc(hidden)]
pub use ::tracing;
/// Output a line to the gitlab log
#[macro_export]
macro_rules! outputln {
($f: expr) => {
$crate::tracing::trace!(gitlab.output = true, $f)
};
($f: expr, $($arg: tt) *) => {
$crate::tracing::trace!(gitlab.output = true, $f, $($arg)*)
};
}
/// Result type for various stagings of a jobs
pub type JobResult = Result<(), ()>;
pub use client::Phase;
/// Async trait for handling a single Job
///
/// Note that this is an asynchronous trait which should be implemented by using the [`async_trait`]
/// crate. However this also means the rustdoc documentation is interesting...
#[async_trait::async_trait]
pub trait JobHandler: Send {
/// Do a single step of a job
///
/// This gets called for each phase of the job (e.g. script and after_script). The passed
/// string array is the same array as was passed for a given step in the job definition.cold
///
/// Note that gitlab concatinates the `before_script` and `script` arrays into a single
/// [Phase::Script] step
async fn step(&mut self, script: &[String], phase: Phase) -> JobResult;
/// Upload artifacts to gitlab
///
/// This gets called depending on whether the job definition calls for artifacts to be uploaded
/// based on the result of the script run
async fn upload_artifacts(&mut self, _uploader: &mut Uploader) -> JobResult {
Ok(())
}
/// Cleanup after the job is finished
///
/// This method always get called whether or not the job succeeded, allowing the job handler to
/// clean up as necessary.
async fn cleanup(&mut self) {}
}
/// Runner for gitlab
///
/// The runner is responsible for communicating with gitlab to request new job and spawn them.
#[derive(Debug)]
pub struct Runner {
client: Client,
build_dir: PathBuf,
run_list: RunList<u64, JobData>,
}
impl Runner {
/// Create a new Runner for the given server url and runner token, storing (temporary job
/// files) in build_dir
///
/// The build_dir is used to store temporary files during a job run. This will also configure a
/// default tracing subscriber if that's not wanted use [`Runner::new_with_layer`] instead.
///
/// ```
/// # use gitlab_runner::Runner;
/// # use url::Url;
/// #
/// let dir = tempfile::tempdir().unwrap();
/// let runner = Runner::new(Url::parse("https://gitlab.com/").unwrap(),
/// "RunnerToken".to_string(),
/// dir.path().to_path_buf());
/// ```
///
/// # Panics
///
/// Panics if a default subscriber is already setup
pub fn new(server: Url, token: String, build_dir: PathBuf) -> Self {
let (runner, layer) = Self::new_with_layer(server, token, build_dir);
Registry::default().with(layer).init();
runner
}
/// Creates a new runner as per [`Runner::new`] and logging layer
///
/// The logging layer should attached to the current tracing subscriber while further runner
/// calls are made otherwise job logging to gitlab will not work
///
/// ```
/// # use gitlab_runner::Runner;
/// # use url::Url;
/// # use tracing_subscriber::{prelude::*, Registry};
/// #
/// let dir = tempfile::tempdir().unwrap();
/// let (runner, layer) = Runner::new_with_layer(Url::parse("https://gitlab.com/").unwrap(),
/// "RunnerToken".to_string(),
/// dir.path().to_path_buf());
/// let subscriber = Registry::default().with(layer).init();
/// ```
pub fn new_with_layer(server: Url, token: String, build_dir: PathBuf) -> (Self, GitlabLayer) {
let client = Client::new(server, token);
let run_list = RunList::new();
(
Self {
client,
build_dir,
run_list: run_list.clone(),
},
GitlabLayer::new(run_list),
)
}
/// The number of jobs currently running
pub fn running(&self) -> usize {
self.run_list.size()
}
/// Try to request a single job from gitlab
///
/// This does a single poll of gitlab for a new job. If a new job received a new asynchronous
/// task is spawned for processing the job. The passed `process` function is called to create a
/// the actual job handler. Returns whether or not a job was received or an error if polling
/// gitlab failed.
///
/// Note that this function is not cancel save. If the future gets cancelled gitlab might have
/// provided a job for which processing didn't start yet.
pub async fn request_job<F, J, Ret>(&mut self, process: F) -> Result<bool, client::Error>
where
F: FnOnce(Job) -> Ret + Sync + Send + 'static,
J: JobHandler + Send + 'static,
Ret: Future<Output = Result<J, ()>> + Send + 'static,
{
let response = self.client.request_job().await?;
if let Some(response) = response {
let mut build_dir = self.build_dir.clone();
build_dir.push(format!("{}", response.id));
let mut run = Run::new(self.client.clone(), response, &mut self.run_list);
tokio::spawn(
async move { run.run(process, build_dir).await }.with_current_subscriber(),
);
Ok(true)
} else {
Ok(false)
}
}
/// Wait untill there are less then max jobs running
pub async fn wait_for_space(&mut self, max: usize) {
self.run_list.wait_for_space(max).await;
}
/// Wait untill there are no more jobs running
pub async fn drain(&mut self) {
self.run_list.wait_for_space(1).await;
}
/// Run continously, processing at most `maximum` jobs concurrently
///
/// This essentially calls [`Runner::request_job`] requesting jobs until at most `maximum` jobs are
/// running in parallel.
pub async fn run<F, J, Ret>(&mut self, process: F, maximum: usize) -> Result<(), client::Error>
where
F: Fn(Job) -> Ret + Sync + Send + 'static + Clone,
J: JobHandler + Send + 'static,
Ret: Future<Output = Result<J, ()>> + Send + 'static,
{
loop {
self.wait_for_space(maximum).await;
match self.request_job(process.clone()).await {
/* continue to pick up a new job straight away */
Ok(true) => continue,
Ok(false) => (),
Err(e) => warn!("Couldn't get a job from gitlab: {:?}", e),
}
sleep(Duration::from_secs(5)).await;
}
}
}