How to ensure nova-compute nodes work before they get added to the scheduler?

asked 2014-01-08 15:04:49 -0500

mfischer gravatar image

If I bring up a new nova-compute node or bring one back-up that's been having some maintenance done, I'd like to be able to prove to myself that the node could successfully run guests before adding it back into the pool for scheduling. As I understand it, as soon as nova-compute starts, when properly configured, the nova-scheduler (on another node) will see it and it will be available for the scheduler to use. Ideally, I'd boot my node and run some nova commands to prove that I could, for example, start a VM with functional and correct network settings before people start using it. In the case of maintenance, I could simply disable the node with nova-manage and re-enable it later, but for a new node, what's the solution?

As a follow-up, can I run nova commands on a compute node without a scheduler? One idea was to bring up the compute node in an isolated network and test nova, but perhaps this is not possible. Could I temporarily run a local nova-schedule?

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2014-01-09 00:22:46 -0500

unmesh-gurjar gravatar image

A request to start a VM will always be acted upon by Scheduler. Follow [1] to configure the Compute so that only admin user can spawn instance on the specific Compute and later revert the policy.json to ensure that the Compute is available for all users!

[1] -

edit flag offensive delete link more


Perfect! I didn't think of that (now obvious) option.

mfischer gravatar imagemfischer ( 2014-01-09 21:11:42 -0500 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2014-01-08 15:04:49 -0500

Seen: 469 times

Last updated: Jan 09 '14