openstack
/
charm-deployment-guide
Code Issues Proposed changes

Add nova-compute configuration for hwoffload 

Ensure that the nova-compute charm is configured for PCI
passthrough of offloaded network devices via the
pci-passthrough-whitelist configuration option.

Change-Id: I23fea95b1f6646f420746fddf4fb4b63ba9c3c86
This commit is contained in:
James Page 2020-05-07 12:12:16 +01:00
parent 2c9811d298
commit 8e17db7e6e
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
deploy-guide/source/app-hardware-offload.rst
View File

@ -101,6 +101,9 @@ The following overlay may be used with the OpenStack base deployment bundle:
charm: cs:neutron-api
options:
enable-hardware-offload: true
nova-compute:
options:
pci-passthrough-whitelist: '{"address": "*:03:*", "physical_network": null}'
In this overlay ``enp3s0f0`` and ``enp3s0f1`` are two ports on the same
Mellanox ConnectX-5 card and are configured as a Linux bond ``bond1`` to enable
@ -108,6 +111,10 @@ VF-LAG for resilience and performance. ``bond1`` is also configured with the
network interface used for VXLAN overlay traffic to allow full offloading of
networks of this type.
The nova-compute charm is configured to use the VF functions provided by the
network cards using the ``pci-passthrough-whitelist`` option. The above example
demonstrates configuration for VXLAN overlay networking.
Creating hardware offloaded ports
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~