GoogleContainerTools/skaffold · error

hook execution failed for pod %q container %q: %w

Error message

hook execution failed for pod %q container %q: %w

What it means

Once a matching pod/container is found, the hook command runs via util.RunCmd in an errgroup goroutine. If the command exits non-zero or fails to run, the error is wrapped as 'hook execution failed for pod %q container %q: %w', attributing the failure to the specific pod and container.

Source

Thrown at pkg/skaffold/hooks/container.go:125

			for _, c := range p.Spec.Containers {
				if matched, err := h.selector(p, c); err != nil {
					return err
				} else if !matched {
					continue
				}
				args := []string{p.Name, "--namespace", p.Namespace, "-c", c.Name, "--"}
				args = append(args, h.cfg.Command...)
				cmd := h.cli.Command(ctx, "exec", args...)
				tr, tw := io.Pipe()
				cmd.Stderr = tw
				cmd.Stdout = tw
				podName := p.Name
				containerName := c.Name
				errs.Go(func() error {
					defer tw.Close()
					err := util.RunCmd(ctx, cmd)
					if err != nil {
						return fmt.Errorf("hook execution failed for pod %q container %q: %w", podName, containerName, err)
					}
					return nil
				})
				pod := p
				var containerStatus v1.ContainerStatus
				for _, status := range pod.Status.ContainerStatuses {
					if status.Name == c.Name {
						containerStatus = status
					}
				}
				errs.Go(func() error {
					return stream.StreamRequest(ctx, out, h.formatter(pod, containerStatus, func() bool { return false }), tr)
				})
			}
		}
	}
	return errs.Wait()
}

View on GitHub (pinned to a1189de023)

Solutions

1. Check the wrapped stderr/stdout in the error to see the actual command failure
2. Verify the command exists in the container image (kubectl exec <pod> -c <ctr> -- which <cmd>)
3. Fix the command/args in the skaffold.yaml hook definition
4. Ensure the container stays Running during the hook and the app exits 0

Example fix

// before
containerHooks:
  postBuild:
    - command: ["./scripts/notify.sh"]   # not present in image
// after
containerHooks:
  postBuild:
    - command: ["/bin/sh", "-c", "[ -f /scripts/notify.sh ] && /scripts/notify.sh || true"]

// ensure the command exists in the image before running the hook
if err := exec.Command("kubectl", "exec", pod, "-c", container, "--", "which", cmdName).Run(); err != nil {
    return fmt.Errorf("command %q not present in %s/%s", cmdName, pod, container)
}

err := hook.Run(ctx, out)
var exitErr *exec.ExitError
if err != nil {
    if errors.As(err, &exitErr) {
        log.Printf("hook failed in pod/container (exit %d): %v", exitErr.ExitCode(), err)
    }
    return err
}

When it happens

Trigger: run dispatching errs.Go for a matched container where util.RunCmd returns an error — the hook command binary is missing in the container, exits non-zero, or exec into the pod fails.

Common situations: Hook command not installed in the target container image; script exits with a non-zero status due to app logic; insufficient permissions to exec into the pod; container restarted/terminated mid-hook; wrong command name in skaffold.yaml.

Related errors

• failed to evaluate pod name pattern %q due to error %w
• failed to evaluate container name pattern %q due to error %w
• getting Kubernetes client: %w
• getting pods for namespace %q: %w
• starting cmd: %w

AI-assisted analysis of GoogleContainerTools/skaffold@a1189de023 (2026-09-05). Data as JSON: /api/errors/ff7e0519246296e2. Report an issue: GitHub.