Open Stack

You appear to be missing the following line from each of your
alarm resource definitions in the Heat template:

  repeat_actions: True

(this is required for the Heat autoscaling cooldown logic to
work)

On the second compute node, do you actually have any nova
instances running there?

Cheers,
Eoghan

----- Original Message -----
> Hi all,
> 
> I want to autoscale with ceilometer and heat in devstack. The
> Single-Node-Configuration works. But now I want to configure for Multi-Node.
> 
> Here is the localrc of the controller-node:
> https://gist.github.com/anonymous/8006733
> 
> And here is the localrc of the compute-node:
> https://gist.github.com/anonymous/8006723
> 
> At the Hypervisor-tab in horizon I can see both servers. With the "ceilometer
> statistics -m cpu_util -p 30" I am getting only the values of the VM at the
> controller-node, I guess. I tried it with different "DesiredCapacity"
> (3,5,7) in my template and saw, that in the "count"-column the value is
> everytime the same as the amount of used Virtual CPUs at the
> controller-node.
> Furthermore did the scaling not work. I used the cpu_util-metric and
> stress-program at the VMs. The initial capacity were created, but the
> scaling did not work. Probably the two problems stick together, don't they?
> :)
> 
> Here is the template:
> https://gist.github.com/anonymous/8009540
> 
> Did I forget a service to enable at the compute-node? I used the
> devstack-multinode-guide from devstack.org and the manual of openstack
> (install the orchestration/telemetry service).
> 
> Many thanks in advance!
> 
> Johnny
> 
> 
> 
> _______________________________________________
> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
> Post to     : openstack at lists.openstack.org
> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>