Longhorn v1.3.3-rc2 release

Release Note

DON'T UPGRADE from/to any RC/Preview releases, because the operation is not supported.

v1.3.3-rc2 released! 🎆

This release introduces improvements and bug fixes as described below about stability, performance, space efficiency, resilience, and so on. Please try it and feedback. Thanks for all the contributions!

Installation

Please ensure your Kubernetes cluster is >= v1.18 and <= v1.24 before installing Longhorn v1.3.3.

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

Upgrade

Please ensure your Kubernetes cluster is >= v1.18 and <= v1.24 before upgrading to Longhorn v1.3.3 from v1.2.x or v1.3.x. Only support upgrading from v1.2.x and v1.3.x.

The live upgrade from v1.2.x/v1.3.x is not supported until the GA release. When available, follow the upgrade instructions here. here.

Deprecation & Incompatibilities

N/A

Known Issues after Release

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

Highlights

• [IMPROVEMENT] Use PDB to protect Longhorn components from unexpected drains (3304) - @PhanLe1010
• [IMPROVEMENT] Periodically clean up volume snapshots (3836) - @c3y1huang @chriscchien
• [IMPROVEMENT] Recurring jobs create new snapshots while being not able to clean up old ones (4898) - @mantissahz @chriscchien

Improvement

• [IMPROVEMENT] Change the script into a docker run command mentioned in 'recovery from longhorn backup without system installed' doc (1521) - @weizhe0422 @chriscchien
• [IMPROVEMENT] liveness and readiness probes with longhorn csi plugin daemonset (3907) - @c3y1huang @roger-ryao
• [IMPROVEMENT] Too many debug-level log messages in engine instance-manager (4427) - @derekbit @chriscchien
• [IMPROVEMENT] share-manager pod bypasses the kubernetes scheduler (4789) - @joshimoo @chriscchien
• [IMPROVEMENT] Unify the format of returned error messages in longhorn-engine (4828) - @derekbit
• [IMPROVEMENT] Affinity in the longhorn-ui deployment within the helm chart (4987) - @mantissahz @chriscchien
• [IMPROVEMENT] Upgrade tcmalloc in longhorn-engine (5050) - @derekbit
• [IMPROVEMENT] Fix Guaranteed Engine Manager CPU recommendation forumula in UI (5338) - @c3y1huang @smallteeths @roger-ryao
• [IMPROVEMENT] Set write-cache of longhorn block device to off explicitly (5382) - @derekbit @chriscchien
• [DOC] Update Kubernetes version info to have consistent description from the longhorn documentation in chart (5399) - @ChanYiLin @roger-ryao
• [IMPROVEMENT] Fix BackingImage uploading/downloading flow to prevent client timeout (5443) - @ChanYiLin @chriscchien
• [IMPROVEMENT] Create a new setting so that Longhorn removes PDB for instance-manager-r that doesn't have any running instance inside it (5549) - @PhanLe1010 @khushboo-rancher
• [IMPROVEMENT][UI] Recurring jobs create new snapshots while being not able to clean up old one (5610) - @mantissahz @smallteeths @roger-ryao
• [IMPROVEMENT] Only active replica if it doesn't have deletion timestamp (5632) - @PhanLe1010

Performance

• [TASK] Disable tcmalloc in data path because newer tcmalloc version leads to performance drop (5096) - @derekbit @chriscchien

Stability

• [BUG] Longhorn won't fail all replicas if there is no valid backend during the engine starting stage (1330) - @derekbit @roger-ryao
• [BUG] Engine binary cannot be recovered after being removed accidentally (4380) - @yangchiu @c3y1huang
• [BUG] volume is stuck in attaching/detaching loop with error Failed to init frontend: device... (4959) - @derekbit @PhanLe1010 @chriscchien
• [BUG] Memory leak in CSI plugin caused by stuck umount processes if the RWX volume is already gone (5296) - @derekbit @roger-ryao
• [BUG] share-manager pod failed to restart after kubelet restart (5507) - @yangchiu @derekbit
• [BUG] RWX volume is stuck at detaching when the attached node is down (5558) - @derekbit @roger-ryao

Bugs

• [BUG] Restoring volume stuck forever if the backup is already deleted. (1867) - @mantissahz @chriscchien
• [BUG] Duplicated default instance manager leads to engine/replica cannot be started (3000) - @PhanLe1010 @roger-ryao
• [BUG] Delete a uploading backing image, the corresponding LH temp file is not deleted (3682) - @ChanYiLin @chriscchien
• [BUG] Replica rebuilding failure with error "Replica must be closed, Can not add in state: open" (3828) - @mantissahz @roger-ryao
• [BUG] Max length of volume name not consist between frontend and backend (3917) - @weizhe0422 @roger-ryao
• [BUG] Can't delete volumesnapshot if backup removed first (4107) - @weizhe0422 @chriscchien
• [BUG] LH continuously reports invalid customized default setting taint-toleration (4554) - @weizhe0422 @roger-ryao
• [BUG] longhorn-engine integration test test_restore_to_file_with_backing_file failed after upgrade to sles 15.4 (4632) - @mantissahz
• [BUG] The old instance-manager-r Pods are not deleted after upgrade (4726) - @mantissahz @chriscchien
• [BUG] Replica Auto Balance repeatedly delete the local replica and trigger rebuilding (4761) - @c3y1huang @roger-ryao
• [BUG] Unable to reuse existing failed replica causes test case test_allow_volume_creation_with_degraded_availability_restore failed (4791) - @yangchiu @mantissahz
• [BUG] Volume metafile getting deleted or empty results in a detach-attach loop (4846) - @mantissahz @chriscchien
• [BUG] Backing image is stuck at in-progress status if the provided checksum is incorrect (4852) - @FrankYang0529 @chriscchien
• [BUG] Duplicate channel close error in the backing image manage related components (4865) - @weizhe0422 @roger-ryao
• [BUG] The node ID of backing image data source somehow get changed then lead to file handling failed (4887) - @shuo-wu @chriscchien
• [BUG] Cannot upload a backing image larger than 10G (4902) - @smallteeths @shuo-wu @chriscchien
• [BUG] System backup showing wrong age (5047) - @smallteeths @khushboo-rancher
• [BUG] Longhorn 1.3.2 fails to backup & restore volumes behind Internet proxy (5054) - @mantissahz @chriscchien
• [BUG] Sync up with backup target during DR volume activation (5292) - @yangchiu @weizhe0422
• [BUG] environment_check.sh does not handle differnt kernel versions in cluster correctly (5304) - @achims311 @roger-ryao
• [BUG] Replica rebuilding caused by rke2/kubelet restart (5340) - @derekbit @chriscchien
• [BUG] Error message not consistent between create/update recurring job when retain number greater than 50 (5434) - @c3y1huang @chriscchien
• [BUG] Do not copy Host header to API requests forwarded to Longhorn Manager (5438) - @yangchiu @smallteeths
• [BUG] test case test_backup_lock_deletion_during_restoration failed (5458) - @yangchiu @derekbit
• [BUG] Volume restoration will never complete if attached node is down (5464) - @derekbit @weizhe0422 @chriscchien
• [BUG] Physical node down test failed (5477) - @derekbit @chriscchien
• [BUG] Backing image with sync failure (5481) - @ChanYiLin @roger-ryao
• [BUG] Example of data migration doesn't work for hidden/./dot-files) (5484) - @hedefalk @shuo-wu @chriscchien
• [BUG] test case test_dr_volume_with_backup_block_deletion failed (5489) - @yangchiu @derekbit
• [BUG] RWX volume attachment failed if tried more enough times (5537) - @yangchiu @derekbit

Misc

• [DOC] RWX support for NVIDIA JETSON Ubuntu 18.4LTS kernel requires enabling NFSV4.1 (3157) - @yangchiu @derekbit
• [TASK][UI] add new recurring job tasks (5272) - @smallteeths @chriscchien

Contributors

• @ChanYiLin
• @FrankYang0529
• @PhanLe1010
• @achims311
• @c3y1huang
• @chriscchien
• @derekbit
• @hedefalk
• @innobead
• @joshimoo
• @khushboo-rancher
• @mantissahz
• @roger-ryao
• @shuo-wu
• @smallteeths
• @weizhe0422
• @yangchiu