Mistral workflow (with variations) implements the compute node maintanance procedure, as described here: http://docs.openstack.org/trunk/openstack-ops/content/maintenance.html#planned_maintenance_compute_node. It raises some assumptions and questiosn for good discussion.

nova-node-maintenance

Workflow:
  name: Compute node planned maintainance
  input: [ host, migrage_target_host]
  output: ''
  on-task-failure: std.continue #catch-all, just like exception handling...

  tasks:
    list-vms:
      action: ListVMs
      parameters: 
        host: $host
      on-success: migrate-vms

    migrage-vms:
      action: MigrateVMs
      parameters:
        id: $list_vms.vm-id
        host: $migrage_target_host
      on-success: stop-nova
      on-failure: [notifyFailure, std.fail]

    stop-nova:
      action: SSH-CMD
      parameters:
        hosts: $host
        cmd: stop nova-compute
      on-success: disable-auto-start
      on-error: [notifyFailure, std.fail]

    disable-auto-start:
      doc: > If you use a configuration management system, 
        such as Puppet, that ensures the nova-compute service is 
        always running, you can temporarily move the init files
      action: SSH-CMD
      parameters: 
        hosts: $host
        cmd: 
          - mkdir /root/tmp
          - mv /etc/init/nova-compute.conf /root/tmp
          - mv /etc/init.d/nova-compute /root/tmp
      on-finish: shut-down-host

    shut-down-host:
      action: SHUTDOWN-HOST
      parameters: 
        hosts: $host
      on-finish: waitForMaintananceToComplete

    waitForMaintananceToComplete:
      doc: > check point, wait till sys admin notifies that 
        maintainance is complete and the system is up.
        We can be smarter, and auto configure a Sensu check 
      action: WAIT_FOR_USER_INPUT
      parameters:
        prompt: "Reply when system is ready"
        on-success: re-enable-auto-start

    re-enable-auto-start:
      action: SSH-CMD
      parameters:
        host: $host
        cmd:
          - mv /root/tmp/nova-compute.conf /etc/init
          - mv /root/tmp/nova-compute /etc/init.d/
      on-success: start-nova

    start-nova:
      action: SSH-CMD
      parameters:
        hosts: $host
        cmd: start nova-compute 
      on-success: check-nova-status

    check-nova-status:
      doc: Validate that Nova has successfully started up
      action: NovaVerify
      parameters:
        host: $host
      on-error: [notifyFailure, std.fail]
      on-success: migrate-vms-back

    migrate-vms-back: 
      doc: | Returns the previously migrated vms back to the host.
             Notice the power of keeping the context! And an elegance 
             of action reuse within the same workflow.
      action: MigrateVMs
      parameters:
        id: $list_vms.vm-id
        host: $host
      on-success: # Oops... I can't do std.end here, because this will be
                  # called multiple times!!!
                  # Should I flag "multi-value" here? 

    notifyFailure:
      # how can we get the task info here? 
      action: SEND_EMAIL_TASK
      parameters:
        to: admin@example.com
        subject: Workflow $__workflow_name__  $__execution_id__ failed
        body: | The worklfow has failed. Please verify.