K8SPG-859 [POC] Percona Server MySQL Hibernation Feature

[hors]

CHANGE DESCRIPTION

Problem:
This PR implements a hibernation feature for Percona Server MySQL clusters that allows automatic pausing and unpausing based on cron schedules. This is particularly useful for development environments, test clusters, or any scenario where you want to automatically stop MySQL clusters during off-hours to save resources.

🎯 Key Features

✅ Core Hibernation Functionality

• Automatic Pause/Unpause: Schedule-based hibernation using cron expressions
• Manual Override: Manual pause/unpause via spec.pause field
• State Synchronization: Hibernation state automatically syncs with cluster state
• Health Checks: Only allows hibernation when cluster is in Ready state
• Backup/Restore Awareness: Prevents hibernation during active backups or restores

✅ Smart Scheduling Logic

• Next Window Scheduling: If cluster is unhealthy during scheduled time, automatically schedules for next window
• Schedule Change Detection: Automatically updates next pause/unpause times when schedules change
• First-time Evaluation: Handles initial hibernation setup correctly
• Proactive Scheduling: Prevents immediate pausing when cluster becomes ready after being unready

✅ Robust Error Handling

• Invalid Schedule Handling: Gracefully handles invalid cron expressions
• Cluster State Management: Proper handling of Initializing, Error, Stopping, Paused, and Ready states
• Race Condition Prevention: Prevents state flipping during cluster startup/recovery

🏗️ Architecture

New Controller: PerconaServerMySQLHibernationReconciler

• Dedicated controller for hibernation logic
• Registered in cmd/manager/main.go
• RBAC permissions for PS objects and backup/restore resources

Enhanced CRD Fields

spec:
  hibernation:
    enabled: true
    schedule:
      pause: "0 18 * * 1-5"    # 6 PM Mon-Fri
      unpause: "0 8 * * 1-5"   # 8 AM Mon-Fri
  pause: false  # Manual override

Status Fields

status:
  hibernation:
    state: "Active"  # Active, Paused, Scheduled, Blocked, Disabled
    nextPauseTime: "2025-09-24T18:00:00Z"
    nextUnpauseTime: "2025-09-25T08:00:00Z"
    lastPauseTime: "2025-09-23T18:00:00Z"
    lastUnpauseTime: "2025-09-24T08:00:00Z"
    reason: "Cluster not ready during scheduled time"

CHECKLIST

Jira

• Is the Jira ticket created and referenced properly?
• Does the Jira ticket have the proper statuses for documentation (Needs Doc) and QA (Needs QA)?
• Does the Jira ticket link to the proper milestone (Fix Version field)?

Tests

• Is an E2E test/test case added for the new feature/change?
• Are unit tests added where appropriate?

Config/Logging/Testability

• Are all needed new/changed options added to default YAML files?
• Are all needed new/changed options added to the Helm Chart?
• Did we add proper logging messages for operator actions?
• Did we ensure compatibility with the previous version or cluster upgrade process?
• Does the change support oldest and newest supported PS version?
• Does the change support oldest and newest supported Kubernetes version?

[JNKPercona]

Test Name	Result	Duration
version-service	passed	00:00:00
async-ignore-annotations	passed	00:00:00
async-global-metadata	passed	00:00:00
auto-config	passed	00:00:00
config	passed	00:00:00
config-router	passed	00:00:00
demand-backup-minio	passed	00:00:00
demand-backup-cloud	passed	00:00:00
async-data-at-rest-encryption	passed	00:00:00
gr-global-metadata	failure	00:15:24
gr-data-at-rest-encryption	failure	00:17:33
gr-demand-backup-minio	failure	00:13:37
gr-demand-backup-cloud	failure	00:13:06
gr-demand-backup-haproxy	passed	00:00:00
gr-finalizer	passed	00:00:00
gr-haproxy	passed	00:00:00
gr-ignore-annotations	passed	00:00:00
gr-init-deploy	passed	00:00:00
gr-one-pod	failure	00:10:08
gr-recreate	failure	00:06:22
gr-scaling	passed	00:00:00
gr-scheduled-backup	passed	00:00:00
gr-security-context	passed	00:00:00
gr-self-healing	passed	00:00:00
gr-tls-cert-manager	passed	00:00:00
gr-users	passed	00:00:00
haproxy	passed	00:00:00
init-deploy	passed	00:00:00
limits	passed	00:00:00
monitoring	passed	00:00:00
one-pod	passed	00:00:00
operator-self-healing	passed	00:00:00
recreate	passed	00:00:00
scaling	passed	00:00:00
scheduled-backup	passed	00:00:00
service-per-pod	passed	00:00:00
sidecars	passed	00:00:00
smart-update	passed	00:00:00
storage	passed	00:00:00
telemetry	passed	00:00:00
tls-cert-manager	passed	00:00:00
users	passed	00:00:00
pvc-resize	passed	00:00:00
We run 43 out of 43		01:16:12

commit: 78dcbca
image: perconalab/percona-server-mysql-operator:PR-1092-78dcbca6