Comment 5 for bug 1422775
Revision history for this message
| Nikola Đipanov (ndipanov) wrote Re: nova libvirt driver assumes qemu support for NUMA pinning | #5 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
1b1ed1a
(Get the code!)
FWIW - I think we should just be failing the guest and provide a non-criptic message. That's the only way to ensure that there is a clear unambiguos semantics of the extra specs we have introduced.
Allowing partial support will lead to even more confusion from the users.
We skip the missing capabilities already (and that is fine because we don't want to be crashing the compute service) but when we detect that a capability was requested that cannot be unambiguously executed.
We may even want to do this at the scheduler level.