Upgrade/longhorn 1.9.2

[m-brando]

This PR upgrades longhron from v1.8.1 to v1.9.2.

Before upgrading

• All volumes must be in the Detached state: kubectl get volumes -n longhorn-system
• For additional details, see the pre-upgrade checks in the Longhorn documentation: https://longhorn.io/docs/1.9.2/deploy/upgrade/#manual-checks-before-upgrade

Notes:
To force the upgrade in a Claudie-managed cluster, changes are required in the input manifest (for example, adding a label to the node pool).

Close #1910

Summary by CodeRabbit

• New Features

  • Added a version-aware pre-upgrade step that applies a node-drain-policy for older Longhorn installations; non-fatal errors emit warnings and installation continues.

• Chores

  • Upgraded Longhorn baseline to v1.9.2 (images and metadata).
  • Migrated CRDs from v1beta1 to v1beta2 and extended schemas with new fields.
  • Updated default node-drain-policy and set default-data-locality to best-effort.
  • Bumped Longhorn setting templates and two deployment image tags.

[coderabbitai]

Walkthrough

Adds Longhorn v1.9.2 manifests and CRD upgrades, updates default Longhorn settings, and adds installer logic to detect older Longhorn versions and conditionally patch the node-drain-policy before applying manifests.

Changes

Cohort / File(s)	Summary
Longhorn manifests & CRDs services/kuber/server/manifests/longhorn.yaml	Bumps Longhorn baseline to v1.9.2: updates images and app.kubernetes.io/version labels, promotes many CRDs from v1beta1→v1beta2 (deprecates v1beta1), and extends CRD schemas/status with new fields (e.g., starting, ublkID, uuid).
Defaults & templates services/kuber/server/manifests/claudie-defaults.yaml, services/kuber/templates/enable-ca.goyaml	Updates Setting resources to longhorn.io/v1beta2; changes node-drain-policy value to block-for-eviction-if-contains-last-replica; sets default-data-locality to best-effort.
Installer longhorn utils services/kuber/server/domain/utils/longhorn/longhorn.go	Adds semver and logger imports, constants (minVersionForDrainPolicy, nodeDrainPolicyValue), and helper methods: getCurrentLonghornVersion, isVersionBelow, patchNodeDrainPolicy. Implements a version-aware pre-install patch of node-drain-policy with non-fatal warnings on errors.
Kustomization image tags manifests/claudie/kustomization.yaml	Updates two Claudie image newTag values for autoscaler-adapter and kuber.

Sequence Diagram(s)

sequenceDiagram
  participant Installer as "Installer"
  participant Kubectl as "kubectl"
  participant KubeAPI as "Kubernetes API"

  Installer->>KubeAPI: Read longhorn-system Deployment labels (getCurrentLonghornVersion)
  alt version found and isVersionBelow(minVersionForDrainPolicy)
    Installer->>Kubectl: patchNodeDrainPolicy(node-drain-policy)
    Kubectl->>KubeAPI: PATCH Setting longhorn-system/node-drain-policy
    KubeAPI-->>Kubectl: patch response
    Kubectl-->>Installer: success/failure (logged, non-fatal)
  else version not found or compare error
    KubeAPI-->>Installer: warn and continue
  end
  Installer->>KubeAPI: apply Longhorn manifests (v1.9.2)
  KubeAPI-->>Installer: apply response

Loading

Possibly related PRs

• Update/longhorn data locality #1933: Modifies services/kuber/server/manifests/claudie-defaults.yaml, touching the same default-data-locality and node-drain-policy settings.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the main change as upgrading Longhorn to version 1.9.2, which aligns with the primary objective of the changeset.
Linked Issues check	✅ Passed	The PR addresses the core requirement to deploy Longhorn v1.9.2 [#1910] through manifest updates, version-aware patching logic, and image upgrades; compatibility appears maintained through CRD version migrations and pre-upgrade safety checks.
Out of Scope Changes check	✅ Passed	All changes directly support the Longhorn upgrade objective: manifest versions updated to 1.9.2, CRD schemas migrated to v1beta2, pre-upgrade node-drain-policy patching added, and container images bumped appropriately.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch upgrade/longhorn-1.9.2

---

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

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

[jakubhlavacka]

Left some comments. Mostly questions.

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@services/kuber/server/domain/utils/longhorn/longhorn.go`:
- Around line 227-231: The current handler around the kc.KubectlGet call
swallows all kubectl errors and returns nil, treating any failure as "not
installed"; update the error handling so that when kc.KubectlGet(...) returns an
error you propagate that error (e.g., return the empty version string and the
err) instead of returning nil, allowing SetUp to receive and log
connectivity/auth errors; locate the kc.KubectlGet call in the Longhorn
version-read function and change the error return to propagate err to the
caller.