Fleet Management

  • Fixed

    • The kubectl-gs login command no longer writes to the main kubeconfig file in case there are no changes in access tokens and/or the current context.
    • Using the --context flag will no longer modify the current context setting in kubeconfig.
  • WARNING Please ensure you’re running kiam-app with App Version v2.5.1 or higher.

    Changed

    • Update aws-efs-csi-driver version to v1.4.5.
  • WARNING Please ensure you’re running kiam-app with App Version v2.5.1 or higher. This is necessary to allow efs-csi-driver to get the Identity Document from Instance Metadata Service. You can verify this by describing kiam-agent DaemonSet and checking if --allow-route-regexp="/latest/*" is set.

    Changed

    • Update aws-efs-csi-driver version to v1.4.4.
    • Disable hostNetwork.
  • Changed

    • Use n1-standard-4 as default instance types for CAPG worker nodes.
    • Adjusted behaviour of the login command to prevent updates of the main kubeconfig file in case there are no changes in access tokens and/or the current context, or if the current context is provided via override (e.g. by using the --context flag).
  • Fixed

    • Fixed a bug in CLI output tests that made them fail randomly
  • Changed

    • Disable kubectl gs template nodepool command for Cluster API (CAPI) based workload Clusters.
  • Fixed

    • Avoid panic in get nodepools when node pool is lacking the release version label.
    • When templating Vintage Azure Cluster, use Flatcar version from the Release CR rather than hardcoded one.
  • Fixed

    • Change module name to github.com/giantswarm/kubectl-gs/v2.

    Changed

    • Upgraded dependencies
  • Changed

    • Use the cluster-values configmap when templating the default-apps-aws app.

    Fixed

    • Fixed a bug in modifying existing entries in self-contained config files where new data for the existing entries failed to be written to the file.
  • Fixed

    • Fixed a bug in login command where the issuer URL was used instead of the server address in login retry attempt.

    Added

    • Add flags --gcp-machine-deployment-sa-email and --gcp-machine-deployment-sa-scopes to template cluster that specify a Google Cloud Platform service account and its scope to a cluster’s machine deployments
    • Added read header timeout to http server

    Changed

    • Adjusted kubectl gs login command to ensure that it writes to the main kubeconfig file only in case there are actual changes in the content of the file.