[Baylem]

Recover an owned, failed game pod when its rolling StatefulSet has observed a different target revision. The controller checks active lifecycle operations and stop fences, then requests a namespace-scoped eviction with UID and resource-version preconditions. Disruption budgets can defer recovery.

The controller suite exercises image-pull and crash-loop failures, unrelated objects, recovery fences and disruption-budget retries. Adds an operator e2e scenario that deploys an invalid image, restores the original image, and waits for a healthy replacement ordinal.

Closes #934.

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 24 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

Run configuration

• Source mode: Auto
• Repository: GameplanePanel/Gameplane
• Branch: fix/recover-corrected-game-rollout
• File: .coderabbit.yaml
• Commit: fc8ded3
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Advanced
• Run ID: 35119a70-96d2-46ab-92b0-598653563503

Commits

Reviewing files that changed from the base of the PR and between e3826d8 and fc8ded3.

Files selected for processing: 10

• changelog/unreleased/934-statefulset-rollout-recovery.md
• charts/gameplane/templates/operator.yaml
• operator/config/rbac/role_namespace.yaml
• operator/internal/controller/gameserver_controller.go
• operator/internal/controller/gameserver_rollout_recovery.go
• operator/internal/controller/gameserver_rollout_recovery_test.go
• operator/internal/controller/lifecycle_operation.go
• operator/specs.md
• test/e2e/buckets.sh
• test/e2e/template_rollout_e2e_test.go

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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.

Share

• X
• Mastodon
• Reddit
• LinkedIn

[github-actions]

CI Report - 66ea4bb

run 43161320932 - attempt 1

0 failed - 13 passed - 0 skipped - 20 cancelled - 18m44s elapsed

Test selection: full (selection-rule change) - per-suite table

Coverage

Module	Measured	Gate	Margin
web	L94.03 S92.14 B86.51 F90.19	L92 S92 B82 F76	+0.14

Changed

Area	Files	+/-
operator/	6	+401/-4
test/	2	+52/-0
charts/	1	+4/-0
changelog/	1	+3/-0

Slowest jobs

Job	Duration
go (gp-module / amd64)	15m6s
go (svcutil / amd64)	15m6s
go (tunnel / amd64)	15m6s
go (tunnel / arm64)	15m6s
lint (api)	15m5s

e2e buckets run

api-agent, api-auth, api-mods, api-rbac, api-roles, bot-fast, multicluster, operator, optional-components, restricted-workloads, telemetry, upgrade

[Baylem]

Verified recovery against a disposable Kubernetes cluster. A healthy server received an invalid image, entered image pull backoff, and returned to Ready on a new pod UID after restoring the original image. The scenario passed in 25.21 seconds.

Reviewed the shared lifecycle claim, pod UID and resource-version preconditions, and cleanup after a lost eviction response. Tests exercise a competing backup claim, disruption-budget refusal, a changed server generation, and controller restart. The operator retains ownership through termination of the old pod.