Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Pullbot #1

Merged
merged 1,121 commits into from
Dec 21, 2023
Merged

Pullbot #1

merged 1,121 commits into from
Dec 21, 2023

Conversation

zeeke
Copy link
Owner

@zeeke zeeke commented Dec 21, 2023

.

SchSeba and others added 30 commits June 27, 2023 15:52
…Nics

Add support for Intel X710 Backplane and Base T
The sriovoperatorconfigs resource was not present in the
validating webhook configuration. The logic to validate this object
was not triggered.

Signed-off-by: Marcelo Guerrero <[email protected]>
The operatorconfig controller tests were broken due to
the validating webhook.

Signed-off-by: Marcelo Guerrero <[email protected]>
…ebhook

Add sriovoperatorconfigs resource to webhook
Early exit from the function to make the validation
logic more evident and allow further improvements.

Signed-off-by: Andrea Panattoni <[email protected]>
Disable the drain while a node is updating its configuration
prevents the config-daemon to uncordon the node, leading
to a node that needs to be manually uncordoned.

Signed-off-by: Andrea Panattoni <[email protected]>
Remove imagePullPolicy alltogether and use k8s defaults
Implementation for new systemd configuration method
Currently for the kind CI, Kind deploys Kubernetes version
v1.27.3, in this version the `ImagePullpolicy` of containers
follow the procedures in [1], this breaks the CI, since the
CI uses the `latest` tag when building the images, and Kubernetes
tries to pull the images always instead of using the built image.

This patch changes the tag of the image built by CI to `e2e-test`,
so that K8S will set the image policy to be `IfNotPresent`.

[1] https://kubernetes.io/docs/concepts/containers/images/#imagepullpolicy-defaulting

Signed-off-by: Abdallah Yasin <[email protected]>
copy the BF3 to the right manifest folder
Signed-off-by: Sebastian Sch <[email protected]>
Add support for Intel X710 Backplane and Base T
…cy-on-ci

Use `e2e-test` tag for images built in e2e CI
now that the issue with kubelet with fixed all the way to k8s 1.25
we can add again the check if the number of existing vfs is 0
there is no need to drain the node because there are no workloads
running with vfs from that nic

Signed-off-by: Sebastian Sch <[email protected]>
Re-introduce the check if the number of VFs is 0
Problem: if you apply a policy with vdpa devices, no VFs are created
Solution: make sure all the references to the file sriov_config.json
are correct, since its location has changed

k8snetworkplumbingwg/sriov-network-operator#471

Signed-off-by: Leonardo Milleri <[email protected]>
Set a TimeEncoder for zap logging system
to have more speaking timestamp.

from `1.6893201844680622e+09` to
`2023-07-14T09:52:29.539392293Z`

Signed-off-by: Andrea Panattoni <[email protected]>
cgoncalves and others added 28 commits November 8, 2023 11:53
a recent refactor to logging introduced these issues[1].
fix them to avoid panic during runtime

[1] c155c82

Signed-off-by: adrianc <[email protected]>
fix odd number of key-value in error logs
Increase end-to-end test coverage
…oller_runtime

Update controller-runtime to v0.16.3
Container `sriov-network-config-daemon` in the daemonset
is the most interesting in the list. Make `kubectl logs ...` target
that by default.

Signed-off-by: Andrea Panattoni <[email protected]>
…penshift-4.15-sriov-network-config-daemon

OCPBUGS-19245: Updating sriov-network-config-daemon images to be consistent with ART
Previously, the code was using the resource version as a short-hand
to check if the contents of the cm has changed. However, the
resource version can change for any reason without any changes to
the actual data inside the cm. For example, when etcd encryption is
enabled, and key rotation occurs, the version will change.

Instead, use a unique hash that depends solely on the contents of
the cm to track the "version" of the DpConfig.

Signed-off-by: Balazs Nemeth <[email protected]>
Signed-off-by: Balazs Nemeth <[email protected]>
Logs `sriov-network-config-daemon` by default
Hash contents of cm instead of using resource version
Previously, the code was using the resource version as a short-hand
to check if the contents of the cm has changed. However, the
resource version can change for any reason without any changes to
the actual data inside the cm. For example, when etcd encryption is
enabled, and key rotation occurs, the version will change.

Instead, use a unique hash that depends solely on the contents of
the cm to track the "version" of the DpConfig.

Signed-off-by: Balazs Nemeth <[email protected]>
Signed-off-by: Balazs Nemeth <[email protected]>
OCPBUGS-23548: Hash contents of cm instead of using resource version
Skip vf configuration when not found in VF group
with command `make -f Makefile.bundle generate bundle`

Synced fields
- `SriovNetwork.Spec.LogLevel`
- `SriovNetwork.Spec.LogFile`
- `SriovOperatorConfigs.Spec.UseCDI`

Signed-off-by: Andrea Panattoni <[email protected]>
Downstream changes for bundles related to
k8snetworkplumbingwg/sriov-network-operator@4d153f8

run `make -f Makefile.bundle bundle` to align manifests.

Signed-off-by: Andrea Panattoni <[email protected]>
dsonly: Allow config-daemon to send k8s events
OCPBUGS-23950: Add support for ConnectX-7
A downstream merge [1][2] overrode the annotations in the config
daemonset. This resulted in he sriov-network-config-daemon pod missing
the `target.workload.openshift.io/management` annotation which led to
the pod not being pinned to the management cores when workload
partitioning is enabled.

This is a downstream-only patch.

[1] openshift#868
[2] commit 0f17753

Signed-off-by: Carlos Goncalves <[email protected]>
OCPBUGS-24231: Fix config daemon annotation override
Copy link

Thanks for your PR,
To run vendors CIs use one of:

  • /test-all: To run all tests for all vendors.
  • /test-e2e-all: To run all E2E tests for all vendors.
  • /test-e2e-nvidia-all: To run all E2E tests for NVIDIA vendor.

To skip the vendors CIs use one of:

  • /skip-all: To skip all tests for all vendors.
  • /skip-e2e-all: To skip all E2E tests for all vendors.
  • /skip-e2e-nvidia-all: To skip all E2E tests for NVIDIA vendor.
    Best regards.

@zeeke zeeke merged commit 0475f94 into master Dec 21, 2023
12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.