vcpu_weight in flavor-create API is not validated

[Matt Riedemann]

Bug 1243195 has some background for context, and there is some discussion in this review:

https://review.openstack.org/#/c/54026/

But essentially it comes down to the fact that the vcpu_weight value for the flavor model isn't validated. The reason it's probably not validated is due to the fact that no one seems to know how the user can set it, because it's not set via the flavor-create API.

The only code using it is in the xenapi virt driver, so I'm wondering if they have something working out of tree at Rackspace that uses it?

If the user can't set it via the code that's in community, maybe it should be removed?

[John Garbutt (johngarbutt)]

It is certainly used by the XenAPI driver here:
https://github.com/openstack/nova/blob/master/nova/virt/xenapi/vm_utils.py#L211

[Matt Riedemann (mriedem)]

John, yeah, it's used by the xenapi virt driver code, just some bad wording on my part. I cleaned up the description.

[Christopher Yeoh (cyeoh-0)]

I think this bug comes down to if we want to be able to set the vcpu_weight via the API we need to validate it. No public api currently lets you set vcpu_weight.

[Blair Bethwaite (blair-bethwaite)]

The vcpu_weight field is Xen specific and quite narrow (e.g., it's cpu specific, but there are lots of other shared hardware resources that can be limited). It should probably be cleaned up in favor of an approach that uses extra_specs, such as https://wiki.openstack.org/wiki/InstanceResourceQuota.

[Markus Zoeller (markus_z) (mzoeller)]

This still looks odd. Change [1] introduced "vcpu_weight" due to bug 861776. There is no reasoning at all *why* this got introduced. With current Newton master () I'm not able to see where a user can set this value, neither as an API parameter (when creating a flavor) nor a config option.

References:
[1] https://github.com/openstack/nova/commit/d6b6cc5

[Markus Zoeller (markus_z) (mzoeller)]

I tend to believe that this is dead code which can be removed. Waiting for XEN folks to confirm this.

[Matt Riedemann (mriedem)]

Markus, apparently rackspace sets the value in the DB directly.

[Markus Zoeller (markus_z) (mzoeller)]

Shortly discussed in IRC: http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2016-05-17.log.html#t2016-05-17T15:05:26

[Markus Zoeller (markus_z) (mzoeller)]

This is an automated cleanup. This bug report has been closed because it
is older than 18 months and there is no open code change to fix this.
After this time it is unlikely that the circumstances which lead to
the observed issue can be reproduced.

If you can reproduce the bug, please:
* reopen the bug report (set to status "New")
* AND add the detailed steps to reproduce the issue (if applicable)
* AND leave a comment "CONFIRMED FOR: <RELEASE_NAME>"
  Only still supported release names are valid (LIBERTY, MITAKA, OCATA, NEWTON).
  Valid example: CONFIRMED FOR: LIBERTY