Bug 1087937

Within the launch instance window, there is an instance count input box. Type in a number larger than 1, and nova will spin up more than one instance of that type.

To avoid (4) (from comment 0), one would migrate those instances from the node to be rebooted. That's also possible via Horizon, it has been implemented during Icehouse release.

As the request is already implemented, and there is a IMHO more elegant way to even keep the instances runing, I see this bug closed. Please re-open, if anything is unclear or isn't solved properly.

I apologize and did not see that this bug was closed and I think my wording was wrong. The customer requested a way to click all the instances and start them when they are all shut off. This happens during an upgrade and it shuts off all the instances.

Created attachment 917630 [details]
Instance Actions

Note that this can be achieved in a deployment by using the customisation module ( http://docs.openstack.org/developer/horizon/topics/customizing.html#modifying-existing-dashboards-and-panels ), it would look something like this:

  class MyStartInstance(tables.StartInstance):

      def allowed(self, request, instance):
          # We need to tweak this because the original doesn't account
          # for 'None' at the moment, which is what happens when this is
          # called as a table action.
          if instance is None:
              return True
          else:
              return super(MyStartInstance, self).allowed(request, instance)


  class MyInstancesTable(tables.InstancesTable):

      class Meta(tables.InstancesTable.Meta):
          table_actions = (tables.LaunchLink,
                           MyStartInstance,
                           tables.SoftRebootInstance,
                           tables.TerminateInstance,
                           tables.InstancesFilterAction)

  views.IndexView.table_class = MyInstancesTable

Considering there is a workaround and the lack of answer to the needinfo, I will close this for now. This is also being tracked upstream at https://bugs.launchpad.net/bugs/1354395

Julie, should this be a 'Known Issue' with your workaround in the doc text? So that it shows up in the Release Notes? thanks, Summer

Hi Summer, I'm not sure if I would qualify this as a known issue as it is a low-severity RFE, rather than a bug. Looking at the upstream tracker, this has been implemented in Kilo.