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

[release-4.15] OCPBUGS-27503: Use DSR load balancing in kube-proxy #2024

Commits on Jan 22, 2024

  1. Change sourceVIP used for kube-proxy

    Changes the method of choosing a sourceVIP for kube-proxy to be in line
    with Microsoft guidelines. The HNS endpoint for the VIPEndpoint
    interface is no longer needed.
    sebsoto authored and openshift-cherrypick-robot committed Jan 22, 2024
    Configuration menu
    Copy the full SHA
    ab9a169 View commit details
    Browse the repository at this point in the history
  2. Switch to DSR load balancing

    The port reservation logic in kube-proxy's dsr mode is better tested.
    Without using DSR, packet loss was being seen by on Windows Server 2019.
    sebsoto authored and openshift-cherrypick-robot committed Jan 22, 2024
    Configuration menu
    Copy the full SHA
    f242d7c View commit details
    Browse the repository at this point in the history
  3. Add Windows Exporter port to firewall

    Explicity allows incoming traffic for the Windows Exporter metrics
    server running on Windows nodes.
    sebsoto authored and openshift-cherrypick-robot committed Jan 22, 2024
    Configuration menu
    Copy the full SHA
    d055ab7 View commit details
    Browse the repository at this point in the history
  4. Refactor machine approver scaling

    Adds new function to help scale up/down the cluster machine approver
    deployment.
    sebsoto authored and openshift-cherrypick-robot committed Jan 22, 2024
    Configuration menu
    Copy the full SHA
    a7c4ee9 View commit details
    Browse the repository at this point in the history
  5. Fix upgrade test for Machine recreation

    Adjusts the update test ensure it works when Machines are recreated as
    part of an upgrade. This will happen when the userdata is changed
    between two releases. The existing problems are that the Cluster Machine
    Approver needs to be scaled so CSRs can be approved, and that storage
    workloads which are being used to test across upgrades need to not be
    assigned a Node affinity. If there is an affinity on the Node that is
    replaced, the workload will not be able to be rescheduled.
    sebsoto authored and openshift-cherrypick-robot committed Jan 22, 2024
    Configuration menu
    Copy the full SHA
    e07b57f View commit details
    Browse the repository at this point in the history