openstack team mailing list archive

[Julien Danjou <julien@xxxxxxxxxxx>]

On Thu, Jun 28 2012, Doug Hellmann wrote:

> I propose that we move the code that connects to libvirt and gets the list
> of instances into the class that calls the pollsters (AgentManager) so we
> can support both calling patterns. That will make the AgentManager the
> ComputeAgentManager (since network pollsters won't necessarily need to talk
> to libvirt) and we will need other managers for calling pollsters that do
> not use compute resources. I don't see that as a problem, since those
> pollsters probably won't run on the compute nodes anyway so we will want
> them to be loaded in another agent process.
>
> What do the rest of you think about this change?

That recalls me my first idea of having one nova.service.Service (but
only one agent daemon) by OS component. Except this time we would put
them in only one daemon. So that sounds fine to me.

How do you propose we configure the enabled manager then? Using the same
kind of system that we have for plugins?
That'd basically makes us have a 2 levels plugin system in the end,
which doesn't bother me anyway since for now everything is automagically
configured and running.

-- 
           Julien

Attachment: pgpJgiuldRCCG.pgp
Description: PGP signature