Skip to content

Remove unused yaml in cvo to speed up reconciliation #1207

@lance5890

Description

@lance5890
  1. Extract the release image from ocp-release:4.20.0-ec.3, found that there are a few unused yaml in the 4.20 version, as follows:
grep -rn "release.openshift.io/delete" ./release-manifests/*
./0000_26_cloud-controller-manager-operator_10_deployment_tombstone.yaml:10:    release.openshift.io/delete: "true"
./0000_30_cluster-api_10_webhooks.yaml:112:    release.openshift.io/delete: "true"
./0000_50_cloud-credential-operator_01-service-delete.yaml:7:    release.openshift.io/delete: "true"
./0000_50_cluster-autoscaler-operator_99_tombstones.yaml:8:    release.openshift.io/delete: "true"
./0000_50_cluster-autoscaler-operator_99_tombstones.yaml:18:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:10:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:21:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:32:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:43:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:54:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:65:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:76:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:87:    release.openshift.io/delete: "true"
./0000_50_cluster-storage-operator_ibm-cloud-managed-cleanup.yaml:98:    release.openshift.io/delete: "true"
./0000_50_console-operator_ocs-install-tour-quickstart.yaml:10:    release.openshift.io/delete: "true"
./0000_50_olm_11-olm-operators.configmap.removed.yaml:7:    release.openshift.io/delete: "true"
./0000_50_olm_12-olm-operators.catalogsource.removed.yaml:7:    release.openshift.io/delete: "true"
./0000_50_olm_14-packageserver.subscription.removed.yaml:7:    release.openshift.io/delete: "true"
./0000_70_cluster-network-operator_02_rbac.yaml:33:    release.openshift.io/delete: "true"
./0000_90_cluster-authentication-operator_03_prometheusrule.yaml:10:    release.openshift.io/delete: "true"
./0000_90_kube-apiserver-operator_04_servicemonitor-apiserver.yaml:259:    release.openshift.io/delete: "true"
./0000_90_kube-apiserver-operator_05_api_performance_dashboard.yaml:3340:    release.openshift.io/delete: "true"
./0000_90_machine-config_90_deletion.yaml:11:    release.openshift.io/delete: "true"

  1. we have to analysis every exact yaml to decide which should be reserved and which should be removed
  2. As the cvo reconcile all these yamls in the directory, and too many unused yamls will cost memory consumption, remove these yamls will to do some help to speed ​​up the reconcile too in the 'permute-flatten-by-number-and-component' model

The following are the related PRs:

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions