Longhorn - v1.3.2

Release Note

v1.3.2 released! 🎆

This release introduces improvements and bug fixes as described below including stability, data correction, performance and so on.

There are two important fixes that will prevent rarely potential data corruption during replica rebuilding, and also improve write performance via the specific filesystem block size.
- 4354: Introduce Data Alignment Correction for existing volumes if the filesystem block size is less than 4096.
- 4594: Use the specific block size for the filesystem to avoid unnecessary Ready-Modify-Write operations between volume head and snapshots.

Please try it and feedback. Thanks for all the contributions!

Installation

Longhorn supports 3 installation ways including Rancher catalog, Kubectl, and Helm. Follow the installation instructions here.

Upgrade

Note Please ensure the Kubernetes cluster is >= v1.18 and <= v1.24 before upgrading to Longhorn v1.3.2 from v1.2.x or v1.3.x. Only support upgrading from v1.2.x and v1.3.x.

Follow the upgrade instructions here.

Deprecation & Incompatibilities

No deprecations and incompatibilities in this version.

Known Issues after Release

Please follow up on here about any outstanding issues found after this release.

Highlights

• [BUG] data corruption due to COW and block size not being aligned during rebuilding replicas (4354) - @PhanLe1010 @chriscchien
• [Improvement] Using specific block size in Longhorn volume's filesystem (4594) - @derekbit @Lichiehyao

Improvement

• [IMPROVEMENT] Support failed/obsolete orphaned backup cleanup (3898) - @mantissahz @chriscchien
• [IMPROVEMENT] Check if node schedulable condition is set before trying to read it (4581) - @weizhe0422 @Lichiehyao
• [IMPROVEMENT] Review/consolidate the sectorSize in replica server, replica volume, and engine (4599) - @yangchiu @derekbit
• [IMPROVEMENT] Reorganize longhorn-manager/k8s/patches and auto-generate preserveUnknownFields field (4600) - @yangchiu @derekbit

Stability

• [BUG] Somehow the Rebuilding field inside volume.meta is set to true causing the volume to stuck in attaching/detaching loop (4212) - @yangchiu @derekbit

Bugs

• [BUG] Restore from backup sometimes failed if having high frequent recurring backup job w/ retention (3055) - @mantissahz @Lichiehyao
• [BUG] The default access mode of a restored RWX volume is RWO (3444) - @weizhe0422 @Lichiehyao
• [BUG] Backend sizes do not match 5368709120 != 10737418240 in the engine initiation phase (3601) - @derekbit @chriscchien
• [BUG] Error/warning during uninstallation of Longhorn v1.3.1 via manifest (4405) - @PhanLe1010 @Lichiehyao
• [BUG] can't upgrade engine if a volume was created in Longhorn v1.0 and the volume.spec.dataLocality is "" (4412) - @derekbit @chriscchien
• [BUG] Confusing description the label for replica delition (4430) - @yangchiu @smallteeths
• [BUG] Misleading message "The volume is in expansion progress from size 20Gi to 10Gi" if the expansion is invalid (4475) - @yangchiu @smallteeths
• [BUG] Continuously rebuild when auto-balance==least-effort and existing node becomes unschedulable (4502) - @yangchiu @c3y1huang
• [BUG] Longhorn accidentally schedule all replicas onto a worker node eventhough the setting Replica Node Level Soft Anti-Affinity is currently disabled (4546) - @yangchiu @mantissahz
• [BUG] the values.yaml in the longhorn helm chart contains values not used. (4601) - @weizhe0422 @Lichiehyao
• [BUG] Can not pull a backup created by another Longhorn system from the remote backup target (4637) - @yangchiu @mantissahz @Lichiehyao
• [BUG] Updating volume-scheduling-error failure for RWX volumes and expanding volumes (4654) - @derekbit @chriscchien
• [BUG] charts/longhorn/questions.yaml include oudated csi-image tags (4669) - @PhanLe1010 @Lichiehyao
• [BUG] rebuilding the replica failed after upgrading from 1.2.4 to 1.3.2-rc2 (4705) - @derekbit @chriscchien

Misc

• [TASK] Update preserveUnknownFields fields in longhorn-manager CRD manifest (4505) - @derekbit @Lichiehyao
• [TASK] Add K8s version upper limitation (< 1.25) to < 1.4 Longhorn chart (4525) - @weizhe0422 @chriscchien
• [TASK] Update longhorn components with the latest backupstore (4552) - @derekbit

Contributors

• @Lichiehyao
• @PhanLe1010
• @c3y1huang
• @chriscchien
• @derekbit
• @innobead
• @mantissahz
• @smallteeths
• @weizhe0422
• @yangchiu