Bug 1309274 - [RFE] set vnic queue# intelligently
[RFE] set vnic queue# intelligently
Status: NEW
Product: ovirt-engine
Classification: oVirt
Component: RFEs (Show other bugs)
3.6.0
x86_64 Linux
medium Severity medium (vote)
: ovirt-4.3.0
: ---
Assigned To: Yaniv Lavi (Dary)
meital avital
: FutureFeature, Performance
Depends On: 1258206 1293233
Blocks:
  Show dependency treegraph
 
Reported: 2016-02-17 06:07 EST by Dan Kenigsberg
Modified: 2017-06-23 04:17 EDT (History)
12 users (show)

See Also:
Fixed In Version:
Doc Type: Enhancement
Doc Text:
Story Points: ---
Clone Of: 1258206
Environment:
Last Closed:
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: Network
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---
ylavi: ovirt‑4.3?
danken: planning_ack?
danken: devel_ack?
danken: testing_ack?


Attachments (Terms of Use)

  None (edit)
Description Dan Kenigsberg 2016-02-17 06:07:22 EST
There are use cases where setting multiple queues per vnic makes sense (e.g running two multi-vcpu heavily-intercommunicating VMs on the same host).

Setting multiple queues has a (minor) side effect of wasting host memory.

This RFE tracks a request to have a smarter policy of setting the number of queues: we'd like to make it equal to the number of vcpu, unless another number is requested by the VM admin.

Queues should be set only on virtio vnics; in particular, they should not set on passed-through vnics.


http://libvirt.org/formatdomain.html#elementsDriverBackendOptions

queues
    The optional queues attribute controls the number of queues to be used for either Multiqueue virtio-net or vhost-user network interfaces. Use of multiple packet processing queues requires the interface having the <model type='virtio'/> element. Each queue will potentially be handled by a different processor, resulting in much higher throughput. virtio-net since 1.0.6 (QEMU and KVM only) vhost-user since 1.2.17 (QEMU and KVM only)

Note You need to log in before you can comment on or make changes to this bug.