e2e: RPS: fix expected cpu set #703
Conversation
Because of a logic oversight, the expected cpu set was overwritten with the value expected for physical devices after the first loop iteration. This broke the test when running on clusters with more than a single worker node. Co-Authored-By: Andrea Panattoni <apanatto@redhat.com> Signed-off-by: Francesco Romani <fromani@redhat.com>
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: ffromani The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci
bot
added
the
approved
|
/cc @MarSik |
|
LGTM |
|
/lgtm |
|
/lgtm |
|
/retest-required |
|
/override ci/prow/e2e-hypershift none of the failures are obviously related to this lane, which seems to be failing also with other (still unrelated) PRs |
|
@ffromani: Overrode contexts on behalf of ffromani: ci/prow/e2e-hypershift 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. |
|
@ffromani: all tests passed! 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. |
Because of a logic oversight, the expected cpu set was overwritten with the value expected for physical devices after the first loop iteration. This broke the test when running on clusters with more than a single worker node. Signed-off-by: Francesco Romani <fromani@redhat.com> Co-authored-by: Andrea Panattoni <apanatto@redhat.com>
…el API (#760) * rps: machine-config: configure default rps mask https://lore.kernel.org/netdev/cover.1676635317.git.pabeni@redhat.com/ exposes a new API to configure default rps mask for network devices. We're adding a MachineConfig file to configure the desired rps mask We are configuring the rps mask directly from sysctl and not from TuneD because at the boot process sysctl spins up earlier than the veth devices. Signed-off-by: Talor Itzhak <titzhak@redhat.com> * rps: tuned: configure default rps mask Altough rps configured directly from sysctl, we're adding it as part of tuned profile as well to keep the profile up to date with all the system tunings. esentially we want to have a single source of truth which is the tuned profile. Signed-off-by: Talor Itzhak <titzhak@redhat.com> * rps: cleanup: remove oci hook This hook point is not in used and should be removed long time ago Signed-off-by: Talor Itzhak <titzhak@redhat.com> * e2e: rps: enable and reformat tests Enable the tests again and look for the default rps value instead of the systemd unit. In addition verify the rps mask not being applied on physical devices. Signed-off-by: Talor Itzhak <titzhak@redhat.com> * E2E: Use appropriate device path for rpsmask test (#691) When using mco pod to execute find command to search for bridges we use /rootfs/sys/devices Signed-off-by: Niranjan M.R <mrniranjan@redhat.com> Co-authored-by: Niranjan M.R <mrniranjan@redhat.com> * e2e: RPS: fix expected cpu set (#703) Because of a logic oversight, the expected cpu set was overwritten with the value expected for physical devices after the first loop iteration. This broke the test when running on clusters with more than a single worker node. Signed-off-by: Francesco Romani <fromani@redhat.com> Co-authored-by: Andrea Panattoni <apanatto@redhat.com> --------- Signed-off-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Niranjan M.R <mrniranjan@redhat.com> Signed-off-by: Francesco Romani <fromani@redhat.com> Co-authored-by: Niranjan M.R <niranjan@ashoo.in> Co-authored-by: Niranjan M.R <mrniranjan@redhat.com> Co-authored-by: Francesco Romani <fromani@redhat.com> Co-authored-by: Andrea Panattoni <apanatto@redhat.com>
Because of a logic oversight, the expected cpu set was overwritten with the value expected for physical devices after the first loop iteration. This broke the test when running on clusters with more than a single worker node.