Skip to:

  1. Skip to Jira Navigation
  2. Skip to Side Navigation
  3. Skip to Main Content

Some fields are ineffective due to the too generic PodSpec struct

Description

In a XtraDB cluster deployment managed by this operator, we can have mainly three kinds of pods: PXC pods, haproxy pods and proxysql pods. The specs for the three kinds of pods are different due to their different functionalities and reconciliation process.

However, these three kinds of pods share the same PodSpec struct definition, which leads to a consequence that certain fields in CRD are ineffective for certain kinds of pods.

For example, the following listed fields are ineffective:

  • spec.haproxy.sslInternalSecretName

  • spec.logcollector.runtimeClassName

  • spec.haproxy.vaultSecretName

  • spec.pxc.replicasExternalTrafficPolicy

  • spec.pxc.replicasServiceType

  • spec.pxc.serviceAnnotations

  • spec.pxc.serviceLabels

  • spec.pxc.serviceType

  • spec.pxc.sidecarResources

Steps To Reproduce

As an example, we will be reproducing this issue using field spec.pxc.serviceLabels.

Apply the following CR in which spec.pxc.serviceLabels is set to key: value.

apiVersion: pxc.percona.com/v1 kind: PerconaXtraDBCluster metadata: finalizers: - delete-pxc-pods-in-order name: test-cluster spec: pxc: serviceLabels: key: value affinity: antiAffinityTopologyKey: kubernetes.io/hostname autoRecovery: true gracePeriod: 600 image: percona/percona-xtradb-cluster:8.0.27-18.1 imagePullPolicy: IfNotPresent podDisruptionBudget: maxUnavailable: 1 resources: requests: cpu: 100m memory: 512Mi size: 3 volumeSpec: persistentVolumeClaim: resources: requests: storage: 512Mi haproxy: labels: key: value affinity: antiAffinityTopologyKey: kubernetes.io/hostname enabled: true gracePeriod: 30 image: percona/percona-xtradb-cluster-operator:1.11.0-haproxy imagePullPolicy: IfNotPresent podDisruptionBudget: maxUnavailable: 1 resources: requests: cpu: 100m memory: 512Mi size: 3 allowUnsafeConfigurations: false backup: image: percona/percona-xtradb-cluster-operator:1.11.0-pxc8.0-backup imagePullPolicy: IfNotPresent pitr: enabled: false schedule: - keep: 3 name: sat-night-backup schedule: 0 0 * * 6 storageName: s3-us-west - keep: 5 name: daily-backup schedule: 0 0 * * * storageName: fs-pvc storages: fs-pvc: type: filesystem volume: persistentVolumeClaim: accessModes: - ReadWriteOnce resources: requests: storage: 512Mi s3-us-west: s3: bucket: S3-BACKUP-BUCKET-NAME-HERE credentialsSecret: my-cluster-name-backup-s3 region: us-west-2 type: s3 verifyTLS: true crVersion: 1.11.0 enableCRValidationWebhook: true logcollector: enabled: true image: percona/percona-xtradb-cluster-operator:1.11.0-logcollector imagePullPolicy: IfNotPresent resources: requests: cpu: 100m memory: 100M pmm: enabled: false proxysql: enabled: false updateStrategy: SmartUpdate upgradeOptions: apply: 8.0-recommended schedule: 0 4 * * * versionServiceEndpoint: https://check.percona.com

Observe that the key: value does not show up in any kubernetes resources.

Suggested Fix

To fix the custom resource definition, an naive approach would be to define different PodSpec for PXC, haproxy and proxySQL.

Environment

Operating System: Ubuntu 20.04 LTS

Software Platform: kind@0.14.0

Client Version: version.Info{Major:"1", Minor:"24", GitVersion:"v1.24.2", GitCommit:"f66044f4361b9f1f96f0053dd46cb7dce5e990a8", GitTreeState:"clean", BuildDate:"2022-06-15T14:22:29Z", GoVersion:"go1.18.3", Compiler:"gc", Platform:"linux/amd64"}

Server Version: version.Info{Major:"1", Minor:"22", GitVersion:"v1.22.9", GitCommit:"6df4433e288edc9c40c2e344eb336f63fad45cd2", GitTreeState:"clean", BuildDate:"2022-05-19T19:53:08Z", GoVersion:"go1.16.15", Compiler:"gc", Platform:"linux/amd64"}

Smart Checklist

Activity

Details

Assignee

Reporter

Found by Automation

Yes

Needs Review

Yes

Needs QA

Yes

Affects versions

Priority

Smart Checklist

Created July 23, 2022 at 7:30 AM
Updated March 5, 2024 at 5:35 PM

Flag notifications