multiqueue vhostuser interface should not have kernel limits applied
Bug #1858102 reported by
sean mooney
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| OpenStack Compute (nova) |
In Progress
|
Low
|
sean mooney | ||
| Queens |
Triaged
|
Low
|
Unassigned | ||
| Rocky |
Triaged
|
Low
|
Unassigned | ||
| Stein |
Triaged
|
Low
|
Unassigned | ||
| Train |
Triaged
|
Low
|
Unassigned | ||
| Ussuri |
In Progress
|
Low
|
sean mooney | ||
Bug Description
When multiqueue support was introduced for vhost-user it reuesed the same code to calculate the queue count as kernel vhost.
In https:/
this change did not take account of the fact that this was shared code and incorrectly applied
the hard coded kernel limits to vhost-user interface not just tap devices.
Revision history for this message
| sean mooney (sean-k-mooney) wrote | #1 |
Revision history for this message
| OpenStack Infra (hudson-openstack) wrote Fix proposed to nova (master) | #2 |
| Changed in nova: | |
| status: | Triaged → In Progress |
Revision history for this message
| OpenStack Infra (hudson-openstack) wrote Change abandoned on nova (master) | #3 |
To post a comment you must log in.
note i have triaged this as low as this is a latent behaviour we have know about for years
however as a bug was filed downstream for this https:/
i have tagged it against all release back to queens where it was reported.