GoogleContainerTools/skaffold · error

inspecting init container

Error message

inspecting init container

What it means

Immediately after creating a debug init container, Skaffold inspects it with d.client.ContainerInspect to discover its mounts (it expects exactly one, the shared debug volume). If the inspect call itself fails — container removed, daemon error, connection lost — the error is wrapped with this message.

Source

Thrown at pkg/skaffold/deploy/docker/deploy.go:286

		})
		if err != nil {
			return nil, err
		}

		m := d.createMount(v, labels)

		// create the init container
		c.Labels = labels
		_, _, id, err := d.client.Run(ctx, out, dockerutil.ContainerCreateOpts{
			ContainerConfig: c,
			Mounts:          []mount.Mount{m},
		})
		if err != nil {
			return nil, errors.Wrap(err, "creating container in local docker")
		}
		r, err := d.client.ContainerInspect(ctx, id)
		if err != nil {
			return nil, errors.Wrap(err, "inspecting init container")
		}
		if len(r.Mounts) != 1 {
			olog.Entry(ctx).Warnf("unable to retrieve mount from debug init container: debugging may not work correctly!")
		}
		// we know there is only one mount point, since we generated the init container config ourselves
		d.debugger.AddSupportMount(c.Image, m)
	}

	bindings := make(network.PortMap)
	config := d.debugger.ConfigurationForImage(containerCfg.Image)
	for _, port := range config.Ports {
		p, err := network.ParsePort(fmt.Sprint(port) + "/tcp")
		if err != nil {
			return nil, err
		}
		bindings[p] = []network.PortBinding{
			{HostIP: netip.MustParseAddr("127.0.0.1"), HostPort: fmt.Sprint(port)},
		}

View on GitHub (pinned to a1189de023)

Solutions

1. Retry `skaffold debug` — often transient (daemon restart or racing cleanup)
2. Check `docker events`/daemon logs around the failure time to see why the container vanished
3. Ensure only one skaffold process manages the deployment (concurrent runs delete each other's containers)
4. Verify daemon stability: `docker info` and Docker Desktop/colima health
5. If reproducible, capture the wrapped inspect error and file it with `skaffold diagnose` output

Example fix

// before
error: inspecting init container: Error: No such container: <id>
// after
$ skaffold delete
$ skaffold debug   # single clean run, no concurrent skaffold processes

null

null

if err := deploy(ctx); err != nil && strings.Contains(err.Error(), "inspecting init container") {
  log.Printf("transient inspect failure (%v); retrying once", err)
  err = deploy(ctx)
}

When it happens

Trigger: setupDebugging calls d.client.ContainerInspect(ctx, id) right after a successful container create; inspect fails because the container exited/was auto-removed, the daemon restarted between create and inspect, or the Docker API connection dropped.

Common situations: An `--rm`-like auto-removal or a competing cleanup deleting the container between create and inspect; Docker daemon crash/restart mid-deploy; very short-lived init container removed by another skaffold process.

Related errors

• transforming image for debugging
• pulling init container image
• %q running container image %q errored during run with status
• docker deployment not supported alongside cluster deployment
• INIT_DOCKER_NETWORK_INVALID_MODE

AI-assisted analysis of GoogleContainerTools/skaffold@a1189de023 (2026-09-05). Data as JSON: /api/errors/49c804ee30c09c6d. Report an issue: GitHub.