Upd Velero v1.17.0 - #1484

Walkthrough

Helm Velero chart bumped to appVersion 1.17.0 / chart 11.0.0. CRDs updated (new fields: cancel, progress, acceptedTimestamp, node, volumeGroupSnapshotLabelKey, csiSnapshot.driver; adjusted required fields). Helm templates/values/schema/docs add repo maintenance ConfigMap wiring, resizePolicy, dual-stack service options, and template/variable tweaks.

Changes

Sequence Diagram(s)

sequenceDiagram
  autonumber
  actor User
  participant Helm as Helm (install/upgrade)
  participant Chart as Velero Chart
  participant K8s as Kubernetes API

  User->>Helm: helm upgrade --install
  Helm->>Chart: render templates with values
  alt repositoryConfigData present
    Chart->>K8s: create ConfigMap repo-maintenance-config
    Chart->>K8s: Deployment velero w/ --repo-maintenance-job-configmap=<name>
  else
    Note right of Chart: omit ConfigMap & flag
  end
  opt metrics.ipFamily settings
    Chart->>K8s: Service with ipFamilyPolicy/ipFamilies
  end
  opt resizePolicy provided
    Chart->>K8s: Deployment/DaemonSet include resizePolicy
  end
  Chart->>K8s: apply CRDs (updated schemas)

sequenceDiagram
  autonumber
  participant Controller as Velero Controller
  participant PVB as PodVolumeBackup CR
  participant Worker as Node Agent
  participant Store as Backup Storage

  Controller->>PVB: create/update (phase -> Accepted)
  PVB-->>Controller: status.acceptedTimestamp set
  Controller->>Worker: assign backup (status.node)
  Worker->>PVB: phase -> Prepared -> InProgress
  Worker->>PVB: update status.progress(bytesDone/totalBytes)
  alt user requests cancel
    User->>PVB: set spec.cancel = true
    Worker->>PVB: phase -> Canceling -> Canceled
  else complete
    Worker->>Store: upload data
    Worker->>PVB: phase -> Completed
  end

sequenceDiagram
  autonumber
  participant Controller as Velero Controller
  participant PVR as PodVolumeRestore CR
  participant Worker as Node Agent
  participant Store as Backup Storage

  Controller->>PVR: create/update (phase -> Accepted)
  PVR-->>Controller: status.acceptedTimestamp set
  Controller->>Worker: assign restore (status.node)
  Worker->>PVR: phase -> Prepared -> InProgress
  Worker->>Store: read data
  alt user requests cancel
    User->>PVR: set spec.cancel = true
    Worker->>PVR: phase -> Canceling -> Canceled
  else complete
    Worker->>PVR: phase -> Completed
  end

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~60 minutes

Suggested labels

release

Suggested reviewers

• lllamnyp
• klinch0

Poem

A hop, a bump—to one-one-seven-oh,
New fields sprout where CRDs grow.
ConfigMaps hum for tidy care,
Canceled jobs stop mid-air.
Dual-stack winds and resize cheer—🥕✨

Pre-merge checks and finishing touches

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

Summary of Changes

Hello Andrei Kvapil (@kvaps), I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request performs a significant upgrade of the Velero application and its Helm chart to version 1.17.0. The primary goal is to incorporate the latest features and improvements from Velero, including a more robust and configurable system for managing repository maintenance jobs. It also updates several Custom Resource Definitions to support new functionalities and provides various Helm chart enhancements for better deployment and operational flexibility.

Highlights

• Velero Version Upgrade: The Velero application version has been updated from v1.16.1 to v1.17.0, and the Helm chart version has been incremented from 10.0.5 to 11.0.0.
• Enhanced Repository Maintenance Configuration: A new, more flexible system for configuring repository maintenance jobs has been introduced. This now uses a ConfigMap-driven approach, allowing for global and per-repository resource limits, node affinity, and priority class settings, replacing the previous direct requests, limits, and latestJobsCount parameters.
• CRD Updates for Backup and Restore: Several Custom Resource Definitions (CRDs) have been updated. Notable changes include the addition of volumeGroupSnapshotLabelKey to Backup and Schedule specs, a driver field for CSI snapshots in DataUploads, and new cancel fields, acceptedTimestamp, and expanded phase enums (e.g., Accepted, Prepared, Canceling, Canceled) for PodVolumeBackup and PodVolumeRestore resources. The resticIdentifier is also no longer a required field in BackupRepository.
• Helm Chart and Deployment Refinements: The Helm chart includes various refinements such as updated conditional logic for revisionHistoryLimit and hostAliases, support for resizePolicy in both the main Velero deployment and the node-agent daemonset, and new checks in NOTES.txt to warn about deprecated configuration parameters. The metrics service now supports ipFamilyPolicy and ipFamilies.
• Default Node Agent Deployment: The deployNodeAgent setting in the main values.yaml has been uncommented and set to true, indicating that the node agent will now be deployed by default.