OpenStack: add CI coverage for SR-IOV #25388
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: EmilienM The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/retest |
3 similar comments
|
/retest |
|
/retest |
|
/retest |
Until 4.10 (and future versions under development), we need to install the PAO from source because it's not present in the Marketplace.
The default will remain 3 but we can now change the number of workers that we want to deploy.
This step registry is meant to be used for the workers so they have /var/config mounted with metadata infos from OpenStack. This step will be useful when deploying OpenShift on OpenStack with SR-IOV enabled, the SR-IOV network operator needs config-drive on the workers.
This new step registry will do the following: * Scale down the workers to 0 * Download the current worker MachineSet * Apply a patch to the MachineSet which includes the necessary bits for SR-IOV: configDrive enabled, direct type port connected to the SR-IOV network. * Deploy the new MachineSet and wait for the worker to be up and running. This step will be useful later when testing SR-IOV.
* Disable the webhook on 4.9, until a PR has landed * Create a SriovNetworkNodePolicy on the SR-IOV workers using the SR-IOV provider network. This step will be required later when we'll test SR-IOV.
* Rename OPENSTACK_PERFORMANCE_NETWORK to OPENSTACK_DPDK_NETWORK * Rename NFV tests to DPDK, so later we can extend it to add SR-IOV
When a deployment runs behind a proxy, we need this snippet, like it's already used by many step registries. This will help the OpenStack team to deploy Operators in our CI that deploys disconnected OpenShift clusters on OpenStack platform.
Like we did for DPDK, this step will create a Pod that will use a VF and consume the SR-IOV network operator. If we run testpmd successfuly and see the NIC, the test will be considered as passing.
* Add necessary steps to the openstack-nfv workflow for testing SR-IOV, and its parameters. * Remove the openstack-nfv job from shiftstack-ci, as it will move under openshift-sriov-network-operator.
Add the OpenStack NFV CI job that test the SR-IOV network operator in the context of OpenShift deployed on top of OpenStack (virtualized worker nodes), for both 4.10 and 4.9. The jobs are non-voting, so won't block anything to merge in the operator.
The operator has a webhook that can take time to start and be ready, the sleep will avoid errors like we saw here: https://prow.ci.openshift.org/view/gs/origin-ci-test/logs/periodic-ci-shiftstack-shiftstack-ci-main-periodic-4.9-e2e-openstack-nfv/1483816862163668992 ``` Error from server (InternalError): error when creating "STDIN": Internal error occurred: conversion webhook for performance.openshift.io/v1, Kind=PerformanceProfile failed: Post "https://performance-operator-service.openshift-performance-addon-operator.svc:443/convert?timeout=30s": dial tcp 10.130.0.42:4343: connect: connection refused ```
| # Remove this workaround when https://github.com/openshift/sriov-network-operator/pull/600 | ||
| # is merged and released in the operator | ||
| if [[ "${oc_version}" == *"4.9"* ]]; then | ||
| oc patch sriovoperatorconfig default --type=merge -n openshift-sriov-network-operator --patch '{ "spec": { "enableOperatorWebhook": false } }' |
There was a problem hiding this comment.
It seems this patch can be removed since the PR got merged.
There was a problem hiding this comment.
not yet, because we need a new zstream release that will contain the fix
| oc create -f - -o jsonpath='{.metadata.name}' <<EOF | ||
| oc_version=$(oc version -o json | jq -r '.openshiftVersion') | ||
| # Once 4.10 is GA, we need to bump it to 4.11 and so on. | ||
| if [[ "${oc_version}" == *"4.10"* ]]; then |
There was a problem hiding this comment.
doesn't the same apply for previous versions?
There was a problem hiding this comment.
no, in previous versions we want to install the operator from the Marketplace to test what customers have (I think, for now). We can probably change it later if we want to catch problems earlier than that, by installing it from source even on 4.9 etc. WDYT?
| echo "${SHARED_DIR}/sriov-worker-node file not found, no worker node for SR-IOV was deployed" | ||
| exit 1 | ||
| fi | ||
| #WORKER_NODE=$(cat "${SHARED_DIR}/sriov-worker-node") |
| @@ -7,10 +7,6 @@ base_images: | |||
| name: builder | |||
| namespace: ocp | |||
| tag: rhel-8-golang-1.16-openshift-4.9 | |||
| openstack-installer: | |||
There was a problem hiding this comment.
Is it being disabled just to reduce the load?
There was a problem hiding this comment.
yeah i was doing it for testing only. Ignore this change.
|
FYI I've (again) started a new PR where we focus on OpenStack related changes: #25608 |
|
@EmilienM: PR needs rebase. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
openshift-ci
bot
added
the
needs-rebase
|
@EmilienM: The following tests failed, say
Full PR test history. Your PR dashboard. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
/close |
|
@EmilienM: Closed this PR. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
This PR is combining the following commits:
openstack/performanceprofile: deploy PAO from source on 4.10
Until 4.10 (and future versions under development), we need to install
the PAO from source because it's not present in the Marketplace.
openstack/generateconfig: allow to change # of workers
The default will remain 3 but we can now change the number of workers
that we want to deploy.
openstack: add new step for config-drive
This step registry is meant to be used for the workers so they have
/var/config mounted with metadata infos from OpenStack.
openstack: add new step for sriov-worker
This new step registry will do the following:
SR-IOV: configDrive enabled, direct type port connected to the SR-IOV
network
openstack: add new step for sriov-networknodepolicy
the SR-IOV provider network.
openstack: make DPDK tests its own
optional-operators/subscribe: add support for proxy
When a deployment runs behind a proxy, we need this snippet, like it's
already used by many step registries.
openstack: add new step for testing SR-IOV
Like we did for DPDK, this step will create a Pod that will use a VF and
consume the SR-IOV network operator.
If we run testpmd successfuly and see the NIC, the test will be considered as passing.
openstack: update openstack-nfv workflow to add SR-IOV
and its parameters.
openshift-sriov-network-operator.
sriov-network-operator: add CI jobs for OpenStack
Add the OpenStack NFV CI job that test the SR-IOV network operator in
the context of OpenShift deployed on top of OpenStack (virtualized
worker nodes), for both 4.10 and 4.9. The jobs are non-voting, so won't
block anything to merge in the operator.