What causes "environment is already up" in vectordotdev/vector?

Stop the environment first with `cargo vdev int stop `

environment is already up

Error message

environment is already up

What it means

`ComposeTest::start` refuses to start an environment whose containers are already running, detected via `is_running()`. Starting twice would conflict with the existing docker compose project, so it bails early with this clear message instead of a docker error.

Solutions

Stop the environment first with `cargo vdev int stop <integration>`
Or tear down manually: `docker compose -p <project> down` in the test directory
Check `docker ps` for containers from a prior run and remove them
Ensure your test harness calls stop in a cleanup/finally step

Example fix

// before
cargo vdev int start aws   # already running from previous run
// after
cargo vdev int stop aws && cargo vdev int start aws

Defensive patterns

Strategy: try-catch

Validate before calling

use std::process::Command;
fn compose_up(project: &str) -> bool {
    Command::new("docker")
        .args(["compose", "ls", "--format", "json"])
        .output()
        .map(|o| String::from_utf8_lossy(&o.stdout).contains(project))
        .unwrap_or(false)
}
if compose_up("vector-integration-tests-aws") { /* stop first */ }

Try / catch

match result {
    Err(e) if e.to_string().contains("environment is already up") => {
        // treat as success: environment ready for use
    }
    other => other?,
}

Prevention

Always pair start with stop in a cleanup/finally block
Run `cargo vdev int stop <integration>` before start in scripts for idempotency
Avoid concurrent integration test runs using the same docker project name
Check `docker ps` when a previous test run crashed

When it happens

Trigger: Calling `start()` (e.g. `cargo vdev int start <integration>`) when a previous start left the compose project up, or two invocations run concurrently.

Common situations: A previous test run crashed before calling stop; forgetting to run `cargo vdev int stop`; running start in two terminals at once; re-running a test script that assumes a clean state.

Understand the failure class

Background: "Invalid state transition" errors: "status must be X, actually Y", "already rejected/charging/uninstalled", "cannot ... while running" — what they mean when a library rejects your call — this error's family across 31 libraries.

Related errors

AI-assisted analysis of vectordotdev/vector@bdb87aeaa4 (2026-09-16). Data as JSON: /api/errors/d869cc4bd9e51db9. Report an issue: GitHub.

Appendix: source

Thrown at vdev/src/testing/integration.rs:276

    pub(crate) fn start(&self) -> Result<()> {
        // For end-to-end tests, we want to run vector as a service, leveraging the
        // image for the runner. So we must build that image before starting the
        // compose so that it is available.
        if self.local_config.kind == ComposeTestKind::E2E {
            self.runner.build(
                Some(&self.config.features),
                &self.env_config,
                true, // E2E tests build Vector in the image
            )?;
        }

        self.config.check_required()?;
        if let Some(compose) = &self.compose {
            self.runner.ensure_network()?;
            self.runner.ensure_external_volumes()?;

            if self.is_running()? {
                bail!("environment is already up");
            }

            let project_name = self.project_name();
            compose.start(&self.env_config, &project_name)?;
        }
        Ok(())
    }

    pub(crate) fn stop(&self) -> Result<()> {
        if let Some(compose) = &self.compose {
            if !self.is_running()? {
                bail!("No environment for {} is up.", self.test_name);
            }

            let project_name = self.project_name();
            compose.stop(&self.env_config, &project_name)?;
        }

View on GitHub (pinned to bdb87aeaa4)