Kubectl Gs

  • As part of our automatic upgrades journey, we have learnt that cluster chart should structure in such a way that allows overwriting all sections in different levels Warning: This results in a breaking change in the template output of a capa clusters machine pools.

    Changed

    • Breaking: Update capa machine pools to template usings an object instead of arrays as of cluster-aws v0.24.0
  • Fixed

    • Fixed creating client certificates for workload clusters in capvcd installations.
  • Changed

    • Updated CAPA template output to support new subnet layout as of cluster-aws v0.21.0
    • Change default folder for cluster base templates generated by gitops add base from bases/clusters/<PROVIDER> to bases/clusters/<PROVIDER>/template

    Added

    • Add default machine pool name for gitops add base generated templates to capa and gcp (CAPG) providers with value: machine-pool0
  • Changed

    • Adjusted communication with Dex in the login command to provide an option to choose from multiple connectors
    • Modified detection of installation providers - downloading the provider information from Athena with a fallback to the old detection from API URLs
    • Added a new provider - cloud-director

    Added

    • Introduced a new --connector-id flag in the login command to specify a Dex connector to use and skip the selection step
    • Ensured that ID tokens needed for OIDC authentication are renewed only when they expire
  • Added

    • Add flags --cluster-type, --https-proxy, --http-proxy, --no-proxy, --api-mode, --dns-mode, --vpc-mode and `–topol
    • ogy-modetotemplate clusterthat specifycapa` as provider.
    • Add gitops add base command to generate CAP[A,G,O] bases. The values for --provider flag is compatible with the template cluster command (A: capa, G: gcp, O: openstack).
  • Changed

    • Extended detection of providers in the login command to take the provider value primarily from Athena with fallback to the original way of inspecting the API URL
  • Fixed

    • Respect --control-plane-instance-type for AWS cluster templating. Previously, the default value m5.xlarge was always used.
  • Fixed

    • Fixed logging in to clusters running on custom domains by removing domain restriction from API URL validation
  • Fixed

    • Add missing namespace in SOPS related secrets created by GitOps subcommands.
  • Fixed

    • Fix update cluster command when scheduling an upgrade for a Cluster when the Cluster CR had no previous annotations.

This part of our documentation refers to our vintage product. The content may be not valid anymore for our current product. Please check our new documentation hub for the latest state of our docs.