[Openstack] Fw: Aw: Re: [Heat] Ceilometer with Multi-Node

Eoghan Glynn eglynn at redhat.com
Wed Dec 18 15:37:01 UTC 2013



----- Original Message -----
> Thanks for your help.
> Now I added to the template:
>   repeat_actions: True
> 
> I omitted it, because I read in the guide:
>   Optional property, defaults to “true”.

Can you provide a link to that docco, because it seems wrong
and potentially very misleading.

So we need to open an LP docco bug and get that fixed ASAP.
 
> With repeat_actions it scals, but only at the base of the instances running
> at the controller (probably).

OK.
 
> Did you mean the controller node, because you wrote the second compute node?
> If I create f. e. 5 instances, these are in running-state.
> At the Hypervisor-Tab I can see f. e. 3 VCPUs(used) at the controller and 2
> at the compute-node.
> But in the statistics the count value is exact the same of the VCPUs at the
> controller.

Sounds like the ceilometer compute agent on the compute-node is not
publishing cpu_util metrics for the local instances.

Can you set debug=True in /etc/ceilometer/ceilometer.conf, restart the
compute agent and check the compute agent log for errors etc?

> Cheers,
> 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
>




More information about the Openstack mailing list