What causes "QEMU driver can't execute commands" in hashicorp/nomad?

Use exec-based tooling inside the guest VM (SSH, cloud-init, QEMU guest agent) instead

QEMU driver can't execute commands

Error message

QEMU driver can't execute commands

What it means

Unconditional error from the QEMU driver's ExecTask: commands cannot be executed inside a QEMU-started virtual machine, so any exec request against a QEMU task always fails.

Source

Thrown at drivers/qemu/driver.go:814

func (d *Driver) TaskStats(ctx context.Context, taskID string, interval time.Duration) (<-chan *drivers.TaskResourceUsage, error) {
	handle, ok := d.tasks.Get(taskID)
	if !ok {
		return nil, drivers.ErrTaskNotFound
	}

	return handle.exec.Stats(ctx, interval)
}

func (d *Driver) TaskEvents(ctx context.Context) (<-chan *drivers.TaskEvent, error) {
	return d.eventer.TaskEvents(ctx)
}

func (d *Driver) SignalTask(_ string, _ string) error {
	return fmt.Errorf("QEMU driver can't signal commands")
}

func (d *Driver) ExecTask(_ string, _ []string, _ time.Duration) (*drivers.ExecTaskResult, error) {
	return nil, fmt.Errorf("QEMU driver can't execute commands")

}

// GetAbsolutePath returns the absolute path of the passed binary by resolving
// it in the path and following symlinks.
func GetAbsolutePath(bin string) (string, error) {
	lp, err := exec.LookPath(bin)
	if err != nil {
		return "", fmt.Errorf("failed to resolve path to %q executable: %v", bin, err)
	}

	return filepath.EvalSymlinks(lp)
}

func (d *Driver) handleWait(ctx context.Context, handle *taskHandle, ch chan *drivers.ExitResult) {
	defer close(ch)
	var result *drivers.ExitResult
	ps, err := handle.exec.Wait(ctx)

View on GitHub (pinned to 482b49bf1a)

Solutions

Use exec-based tooling inside the guest VM (SSH, cloud-init, QEMU guest agent) instead
Switch exec-based health checks to TCP/HTTP checks against the VM
Use the raw_exec or exec driver if in-VM exec via Nomad is essential

Example fix

// before
res, err := driver.ExecTask(taskID, []string{"ls"}, time.Second) // unsupported
// after
// check: nomad alloc exec only works for exec/java/raw_exec drivers; use ssh into the VM

Defensive patterns

Strategy: validation

Validate before calling

// Go: gate exec calls by driver kind
if driverName == "qemu" {
    return errors.New("exec unsupported; use ssh/guest agent")
}

Try / catch

if _, err := drv.ExecTask(id, cmd, timeout); err != nil && strings.Contains(err.Error(), "can't execute commands") {
    // fall back to ssh into the VM
}

Prevention

Avoid exec-based health checks on QEMU tasks; use TCP/HTTP checks
Do not use `nomad alloc exec` on QEMU allocations
Plan VM debugging via SSH or cloud-init

When it happens

Trigger: Calling Driver.ExecTask(taskID, cmd, timeout) on a QEMU task, e.g. nomad alloc exec or exec-based health checks against a QEMU workload.

Common situations: Running `nomad alloc exec <alloc> qemu-task -- ls` ; exec-based check blocks in job specs; debugging scripts assuming exec support.

Related errors

AI-assisted analysis of hashicorp/nomad@482b49bf1a (2026-09-04). Data as JSON: /api/errors/9353ac725f455e9e. Report an issue: GitHub.