starlingx/docs
Code Issues Proposed changes

Merge "Update with additional info (dsr8)"

Browse Source
This commit is contained in:
Zuul 2023-11-08 21:29:28 +00:00 committed by Gerrit Code Review
commit 1737692e0d
1 changed files with 76 additions and 36 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

112
doc/source/node_management/kubernetes/hardware_acceleration_devices/configure-sriov-fec-operator-to-enable-hw-accelerators-for-hosted-vran-containarized-workloads.rst
View File

@ -104,7 +104,6 @@ following |vRAN| |FEC| accelerators:
accelerator-discovery-svh87 1/1 Running 0 3m26s
sriov-device-plugin-j54hh 1/1 Running 0 3m26s
sriov-fec-controller-manager-77bb5b778b-bjmr8 2/2 Running 0 3m28s
sriov-fec-controller-manager-77bb5b778b-wpfld 2/2 Running 1 (3m26s ago) 3m28s
sriov-fec-daemonset-stnjh 1/1 Running 0 3m26s
#. List all the nodes in the cluster with |FEC| accelerators installed.
@ -252,7 +251,7 @@ following |vRAN| |FEC| accelerators:
acceleratorSelector:
pciAddress: 0000:17:00.0
physicalFunction:
pfDriver: "pci-pf-stub"
pfDriver: "vfio-pci"
vfDriver: "vfio-pci"
vfAmount: 1
bbDevConfig:
@ -297,7 +296,7 @@ following |vRAN| |FEC| accelerators:
acceleratorSelector:
pciAddress: 0000:17:00.0
physicalFunction:
pfDriver: "pci-pf-stub"
pfDriver: "vfio-pci"
vfDriver: "vfio-pci"
vfAmount: 2
bbDevConfig:
@ -469,7 +468,7 @@ following |vRAN| |FEC| accelerators:
acceleratorSelector:
pciAddress: 0000:f7:00.0
physicalFunction:
pfDriver: pci-pf-stub
pfDriver: vfio-pci
vfDriver: vfio-pci
vfAmount: 1
bbDevConfig:
@ -518,7 +517,7 @@ following |vRAN| |FEC| accelerators:
acceleratorSelector:
pciAddress: 0000:f7:00.0
physicalFunction:
pfDriver: pci-pf-stub
pfDriver: vfio-pci
vfDriver: vfio-pci
vfAmount: 2
bbDevConfig:
@ -551,10 +550,8 @@ following |vRAN| |FEC| accelerators:
#. If you need to run the operator on a |prod-long| (|AIO-SX|), then you
should provide ``SriovFecClusterConfig`` with ``spec.drainSkip: True``
to avoid node draining, because it is impossible to drain a node if
there is only one node.
#. The ``SriovFecClusterConfig`` must be provided with
``spec.drainSkip: True`` to avoid node draining.
#. Create and apply a ``SriovFecClusterConfig`` custom resource using
the above examples as templates, setting the parameters
@ -613,7 +610,7 @@ following |vRAN| |FEC| accelerators:
numAqsPerGroups: 16
numQueueGroups: 4
pciAddress: "0000:17:00.0"
pfDriver: pci-pf-stub
pfDriver: vfio-pci
vfAmount: 1
vfDriver: vfio-pci
status:
@ -627,7 +624,7 @@ following |vRAN| |FEC| accelerators:
inventory:
sriovAccelerators:
- deviceID: 0d5c
driver: pci-pf-stub
driver: vfio-pci
maxVirtualFunctions: 16
pciAddress: "0000:17:00.0"
vendorID: "8086"
@ -752,7 +749,7 @@ following |vRAN| |FEC| accelerators:
numAqsPerGroups: 16
numQueueGroups: 4
pciAddress: 0000:f7:00.0
pfDriver: pci-pf-stub
pfDriver: vfio-pci
vfAmount: 1
vfDriver: vfio-pci
status:
@ -766,7 +763,7 @@ following |vRAN| |FEC| accelerators:
inventory:
sriovAccelerators:
- deviceID: 57c0
driver: pci-pf-stub
driver: vfio-pci
maxVirtualFunctions: 16
pciAddress: 0000:f7:00.0
vendorID: "8086"
@ -805,9 +802,13 @@ following |vRAN| |FEC| accelerators:
|FEC| operator and has the default value of
``02bddbbf-bbb0-4d79-886b-91bad3fbb510``
- It is highly recommended to change the default vfio-token when
configuring the accelerator in vfio mode (ie., vfio-pci driver for PF
interface).
- The ``VFIO_TOKEN`` could be changed by setting
``SRIOV_FEC_VFIO_TOKEN`` before application Apply with
:command: `system helm-override-update`.
:command:`system helm-override-update`.
- This example sets the ``SRIOV_FEC_VFIO_TOKEN`` using ``uuidgen``.
@ -816,7 +817,8 @@ following |vRAN| |FEC| accelerators:
~(keystone_admin)$ system helm-override-update sriov-fec-operator sriov-fec-operator sriov-fec-system --set env.SRIOV_FEC_VFIO_TOKEN=`uuidgen`
- For the |VF| interface, the same ``VFIO_TOKEN`` must be configured by
the application.
the application. You can get the token using the command
:command:`system helm-override-show sriov-fec-operator sriov-fec-operator sriov-fec-system`.
- To configure ACC100, N3000 and ACC200 in vfio mode, you should provide
``sriovFecClusterConfig`` with
@ -837,25 +839,23 @@ following |vRAN| |FEC| accelerators:
- See :ref:`Set Up Pods to Use SRIOV to Access ACC100/ACC200 HW Accelerators
<set-up-pods-to-use-sriov>`.
- The resource name for |FEC| |VFs| configured with |SRIOV| |FEC| operator
must be ``intel.com/intel_fec_acc100`` for ACC100,
``intel.com/intel_fec_5g`` for N3000 and ``intel.com/intel_fec_acc200``
for ACC200 when requested in a pod spec unless the resource name was
modified using the `system helm-override-update` command.
- Resource Request: The resource name for |FEC| |VFs| configured with |SRIOV|
|FEC| operator must be ``intel.com/intel_fec_acc100`` for ACC100,
``intel.com/intel_fec_5g`` for N3000 and ``intel.com/intel_fec_acc200`` for
ACC200 when requested in a pod spec unless the resource name was modified
using the `system helm-override-update` command.
- If the ACC100 resource name was modified to ``intel_acc100_fec``, then
the resource requests and limits must match the same name as shown
below.
- Resource request for ACC100.
.. code-block:: none
resources:
requests:
intel.com/intel_acc100_fec: '16'
intel.com/intel_fec_acc100: '16'
limits:
intel.com/intel_acc100_fec: '16'
intel.com/intel_fec_acc100: '16'
- Use the default resource name for N3000 in a pod spec.
- Resource request for N3000.
.. code-block:: none
@ -865,7 +865,7 @@ following |vRAN| |FEC| accelerators:
limits:
intel.com/intel_fec_5g: '2'
- Use the default resource name for ACC200 in a pod spec.
- Resource request for ACC200.
.. code-block:: none
@ -875,15 +875,9 @@ following |vRAN| |FEC| accelerators:
limits:
intel.com/intel_fec_acc200: '16'
- Applications that are using |FEC| |VFs| when the |PF| interface is bound
with the ``vfio-pci`` driver, should provide the ``vfio-token`` to |VF|
interface.
- vfio-token: (in case of vfio mode)
For example, a sample |DPDK| application can provide ``vfio-vf-token`` via
Environment Abstraction Layer (EAL) parameters.
:command:`./test-bbdev.py -e="--vfio-vf-token=02bddbbf-bbb0-4d79-886b-91bad3fbb510 -a0000:f7:00.1"`
- An application pod can get the |VFIO| token through a pod environment
An application pod can get the |VFIO| token through a pod environment
variable.
For example, reference the pod spec section for vfio token injection.
@ -894,4 +888,50 @@ following |vRAN| |FEC| accelerators:
- name: SRIOV_FEC_VFIO_TOKEN
value: "02bddbbf-bbb0-4d79-886b-91bad3fbb510"
:command:`./test-bbdev.py -e="--vfio-vf-token=$SRIOV_FEC_VFIO_TOKEN -a0000:f7:00.1"`
.. note::
The application can get the existing vfio-token using the command below,
if the user updates the custom vfio-token.
:command:`system helm-override-show sriov-fec-operator sriov-fec-operator sriov-fec-system`
If the vfio-token is available by default, it will not be displayed in the
output file.
.. note::
Use the default vfio-token for testing purposes only.
- Run the following command once the application pod is ready to get the |PCI|
address of the allocated |FEC| device along with the |VFIO| token when
applicable.
- ACC100
.. code-block:: none
sysadmin@controller-0:~$ kubectl exec -ti app-pod -- env | grep FEC
PCIDEVICE_INTEL_COM_INTEL_FEC_ACC100=0000:32:00.0
SRIOV_FEC_VFIO_TOKEN=02bddbbf-bbb0-4d79-886b-91bad3fbb510
- ACC200
.. code-block:: none
sysadmin@controller-0:~$ kubectl exec -ti app-pod -- env | grep FEC
PCIDEVICE_INTEL_COM_INTEL_FEC_ACC200=0000:f7:00.0
SRIOV_FEC_VFIO_TOKEN=02bddbbf-bbb0-4d79-886b-91bad3fbb510
- N3000
.. code-block:: none
sysadmin@controller-0:~$ kubectl exec -ti app-pod -- env | grep FEC
PCIDEVICE_INTEL_COM_INTEL_FEC_5G=0000:1f:00.0
- Applications that are using |FEC| |VFs| when the |PF| interface is bound
with the ``vfio-pci`` driver, should provide the ``vfio-token`` to the |VF|
interface.
For example, a sample |DPDK| application can provide ``vfio-vf-token`` via
Environment Abstraction Layer (EAL) parameters.
:command:`./test-bbdev.py -e="--vfio-vf-token=$SRIOV_FEC_VFIO_TOKEN -a$PCIDEVICE_INTEL_COM_INTEL_FEC_ACC200"`