[Openstack] a problem with nova scheduler

Mike Wilson geekinutah at gmail.com
Thu Oct 17 01:29:03 UTC 2013


There is a scheduler hint called 'force_hosts' for this sort of thing. I believe the command line looks like this:

Nova boot <server_name> --hints force_hosts=host1 <everything else>

Does that work for you?

This does not bypass the scheduler, but it is built into the scheduler to allow you to explicitly place new instances on a subset of hardware.

-Mike Wilson

Sent from my iPhone

> On Oct 16, 2013, at 6:35 PM, Li Ma <skywalker.nick at gmail.com> wrote:
> 
> Hi all,
> 
> I encountered a problem with nova-scheduler. Given that I have 20 different compute nodes, I need to run several instances on a certain compute node for testing or verification. I go through all the scheduling algorithms and find out I can achieve this by availability zone.
> 
> But this is a little bit weird to create more than 20 availability zone with only one compute node temporarily. Is there any other solutions for this?
> 
> In my opinion, there should be a certain parameter with nova CLI to specify which compute node this instance should be placed directly without any scheduling. Using this method, I can bypass the scheduling options defined in the nova.conf and do what we want to meet the specific requirements. 
> 
> Thanks,
> -- 
> Li Ma (Nick)
> 
> Email: skywalker.nick at gmail.com
> 
> _______________________________________________
> 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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20131016/d6d34960/attachment.html>


More information about the Openstack mailing list